<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Prateeksha Web Design on Medium]]></title>
        <description><![CDATA[Stories by Prateeksha Web Design on Medium]]></description>
        <link>https://medium.com/@prateekshawebdesign?source=rss-c4c4cda9f964------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*f4x5t4NZ4gNAi1aj.</url>
            <title>Stories by Prateeksha Web Design on Medium</title>
            <link>https://medium.com/@prateekshawebdesign?source=rss-c4c4cda9f964------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 07:46:56 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@prateekshawebdesign/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[A robots.txt line can say no to AI training. Here is what it does, and what it doesn’t.]]></title>
            <link>https://medium.com/@prateekshawebdesign/a-robots-txt-line-can-say-no-to-ai-training-here-is-what-it-does-and-what-it-doesnt-710774031752?source=rss-c4c4cda9f964------2</link>
            <guid isPermaLink="false">https://medium.com/p/710774031752</guid>
            <category><![CDATA[web-development]]></category>
            <category><![CDATA[ai-training]]></category>
            <category><![CDATA[robotstxt]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[seo]]></category>
            <dc:creator><![CDATA[Prateeksha Web Design]]></dc:creator>
            <pubDate>Tue, 10 Nov 2026 11:03:47 GMT</pubDate>
            <atom:updated>2026-11-10T11:03:47.609Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Google and Microsoft say they will honour a shared preference token in 2027. It states a preference about use. It does not block crawling, and today it enforces nothing.</em></p><p>A robots.txt file has always answered one question: which crawlers may fetch which URLs. The new AI-training token answers a different one: how a site would like its content used once it has been fetched. The two look the same on the page, which is why the new line is easy to misread.</p><p>This is what the token is, who has said they will honour it and when, and what a business can sensibly do with it now. Everything below was checked against the primary pages on 6 November 2026 and is dated, because most of it is a promise about 2027.</p><h3>The line itself</h3><p>On 30 October 2026 an Internet-Draft titled “Training Preference Signal” was published by K. Kelley of Anthropic. It adds no new directive and no new grammar. It defines a policy token, __usage_ai-training, that a site names in an ordinary robots.txt group. The draft&#39;s own example:</p><pre>User-Agent: __usage_ai-training<br>Disallow: /<br>Allow: /press-kit/</pre><p>The draft says this expresses a preference against AI training for everything except /press-kit/, where it expresses a preference in favour. Two more rules from the draft matter in practice. A wildcard group (User-Agent: *) must not be read as matching the token, so your existing rules for all crawlers say nothing about training. And if no group names the token, &quot;then no preference has been expressed.&quot;</p><p>It is a working document. The draft itself says it can be updated, replaced or obsoleted at any time, and it expires on 3 May 2027 unless it is renewed. It is not an IETF standard.</p><h3>What it is not</h3><p>The draft states: “This policy token is not a directive for crawling behavior.” Bing’s announcement says the same thing from the implementer’s side: “These settings govern use of content, not crawl access,” and an Allow preference does not override a robots.txt rule that blocks Bingbot.</p><p>So the line does not stop a bot from fetching a page, and it does not remove anything from search. If you want to limit crawling, the draft points you to groups for crawler product tokens, which is what robots.txt has always done.</p><h3>Who has said they will honour it</h3><ul><li><strong>Microsoft (Bing).</strong> “Starting February 1, 2027, Microsoft will honor the new preferences as it crawls.” AI Training can be set in robots.txt for a site or path, and a page-level setting overrides robots.txt. (Bing blog, 30 October 2026.)</li><li><strong>Google.</strong> Aims to support the control “along with other AI providers in 2027”; its crawler documentation says “early 2027”. If a site uses the token to opt out and does not mention Google-Extended, Google plans to treat that as disallowing Google-Extended. (Google Search Central blog, 30 October 2026, and Google’s crawler documentation.)</li><li><strong>OpenAI.</strong> Reported by TheNextWeb as aiming to implement it “in the coming months”, with no date. I did not find a primary statement. (Secondary report, 31 October 2026.)</li><li><strong>Anthropic.</strong> Authored the draft. TheNextWeb reports it has started work but gives no date. (Secondary report, 31 October 2026.)</li></ul><p>Until those dates, nothing listed above is enforced by the token. Publishing the line now records your preference; it does not make anyone follow it.</p><h3>The catch with Google</h3><p>Google-Extended is the existing control for Gemini. Google’s crawler documentation says it governs “whether content Google crawls from their sites may be used for training future generations of Gemini models,” and also covers grounding, which supplies Search index content to models at prompt time. It also says Google-Extended “does not impact a site’s inclusion in Google Search.”</p><p>Put the two statements side by side. A site-wide opt-out through the new token will be treated as a Google-Extended disallow, and Google-Extended covers grounding as well as training. My reading, which is inference and not Google’s wording, is that a training opt-out for Google could also remove your pages from Gemini’s grounding. Google’s own post says sites that explicitly allow Google-Extended are eligible to appear in Gemini Apps responses and receive clicks through grounding. If appearing in those answers matters to you, make that decision on purpose before 2027.</p><p>Bing separates the two. Its grounding preference is page-level only: “There is no site-wide robots.txt grounding preference.” Both preferences default to Allow, and neither changes eligibility for ordinary Bing results. The preferences replace NoArchive and NoCache for training and grounding, and Microsoft asks sites to complete the steps by 1 March 2027. Microsoft also says “Neither choice guarantees inclusion, traffic, or payment.”</p><h3>What a business can reasonably do now</h3><ol><li><strong>Decide the policy first.</strong> Three separate questions get mixed together: may anyone train on this content, may AI products ground answers in it, and should it appear in search. The token addresses the first. Google ties the first two together; Bing keeps them apart.</li><li><strong>Find out who controls your robots.txt.</strong> On many sites it is generated by a CMS, plugin, framework or CDN rather than written by hand. Editing the wrong copy changes nothing.</li><li><strong>Add the group, with a comment recording the date and reason.</strong> Keep any existing rules for named crawlers, including Google-Extended if you have set one, because Google says company-specific controls keep working alongside the new signal.</li><li><strong>Do not use it to protect anything private.</strong> A preference is not access control. Confidential material belongs behind authentication, whatever robots.txt says.</li><li><strong>Put review dates in the calendar:</strong> 1 February 2027 for Microsoft, and the release of Google’s support in 2027.</li><li><strong>Write down what you chose.</strong> In a year, someone will ask whether the site said no, and when.</li></ol><h3>Where this is still unsettled</h3><p>The vocabulary is not final: Google calls it a draft proposal based on the IETF AI Preferences working group, with the wider vocabulary still under discussion. Support from the big providers is stated intent, not shipped behaviour. And the question of whether any individual crawler will actually comply is, as with all of robots.txt, a matter of the crawler’s good faith.</p><p>For the wider picture of how search and AI answers choose sources, there is a <a href="https://prateeksha.com/blog/seo-aeo-geo-business-visibility-guide">practical visibility guide to SEO, AEO and GEO</a> on the Prateeksha blog, and a <a href="https://prateeksha.com/blog/generate-sitemap-xml-robots-txt-nextjs-google-indexing">walkthrough of generating sitemap.xml and robots.txt in Next.js</a> for the mechanics of the file.</p><p>Sources (read 6 November 2026):<br> — Internet-Draft, Training Preference Signal: <a href="https://www.ietf.org/archive/id/draft-kelley-aipref-training-preference-signal-00.html">https://www.ietf.org/archive/id/draft-kelley-aipref-training-preference-signal-00.html</a><br> — Google Search Central, AI training control (30 Oct 2026): <a href="https://developers.google.com/search/blog/2026/10/ai-training-control">https://developers.google.com/search/blog/2026/10/ai-training-control</a><br> — Google crawler documentation, Google-Extended: <a href="https://developers.google.com/crawling/docs/crawlers-fetchers/google-common-crawlers">https://developers.google.com/crawling/docs/crawlers-fetchers/google-common-crawlers</a><br> — Bing Search blog, clearer preferences for training and grounding (30 Oct 2026): <a href="https://blogs.bing.com/search/2026/10/your-content-your-choice-clearer-preferences-for-training-and-grounding/">https://blogs.bing.com/search/2026/10/your-content-your-choice-clearer-preferences-for-training-and-grounding/</a><br> — TheNextWeb, shared AI-training opt-out (31 Oct 2026, secondary): <a href="https://thenextweb.com/news/ai-training-opt-out-robots-txt-google-openai-microsoft-anthropic">https://thenextweb.com/news/ai-training-opt-out-robots-txt-google-openai-microsoft-anthropic</a></p><p>Disclosure: the two prateeksha.com links in this article are to articles on my agency’s blog.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=710774031752" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Next.js Cache Components: What Existing Sites Should Do Now]]></title>
            <link>https://medium.com/@prateekshawebdesign/next-js-cache-components-what-existing-sites-should-do-now-c35ab0c40cf5?source=rss-c4c4cda9f964------2</link>
            <guid isPermaLink="false">https://medium.com/p/c35ab0c40cf5</guid>
            <category><![CDATA[web-development]]></category>
            <category><![CDATA[react]]></category>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[nextjs]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <dc:creator><![CDATA[Prateeksha Web Design]]></dc:creator>
            <pubDate>Wed, 04 Nov 2026 10:20:42 GMT</pubDate>
            <atom:updated>2026-11-04T10:24:46.228Z</atom:updated>
            <content:encoded><![CDATA[<h4>Next.js 16.4 now recommends Cache Components for every app. For a site already in production, the sensible order is to patch first, take inventory second and migrate in stages.</h4><figure><img alt="Illustration of a modular office-style building made of reusable content blocks, with separate teal, amber and blue ribbons flowing alongside it." src="https://cdn-images-1.medium.com/max/1024/1*8nJmfL-uZ0MMQW0ZzQOA9Q.png" /><figcaption>Cover illustration generated with ChatGPT image generation.</figcaption></figure><p>On 6 October the Next.js team <a href="https://nextjs.org/blog/next-16-4">recommended Cache Components</a> as the best choice for every Next.js app. New projects created with create-next-app now have it switched on, and the team says it will become the default in Next.js 17.</p><p>A recommendation can change overnight. A production site’s caching boundaries cannot. They decide which visitor sees which content, how fresh a price or headline is, and how much work the server does per request.</p><p>So the question for an existing site is not “should we turn it on?” It is “what do we know about our caching, and in what order should we act?”</p><h3>What changed in 16.4, and what did not</h3><p>Cache Components is a programming model for the App Router. You mark parts of the component tree as cacheable with the use cache directive. Everything else stays dynamic by default. Next.js builds a static shell for each route and streams request-specific parts in behind it.</p><p>The 16.4 post does three things: it recommends the model universally, makes it the default for newly scaffolded apps, and announces it as the Next.js 17 default. The team says it held back from a universal recommendation earlier because the model could not always match the cost and performance of the old one, and that 16.4 closes those gaps.</p><p>It does not change your existing app. The <a href="https://nextjs.org/docs/app/getting-started/caching">caching documentation</a> says an existing app enables the feature itself, with cacheComponents: true and partialPrefetching: true in its config. Those flags apply to the whole application. A greenfield project adopts the model with no history; an existing site carries years of implicit caching behaviour that the new model replaces.Where it sits among your existing layers</p><p><strong>React Server Components</strong> decide where code runs. <strong>Client components</strong> handle interaction in the browser. <strong>A CDN</strong> stores responses at the network edge.</p><p><strong>Cache Components</strong> decide what the server may reuse, and for how long, at function or component level. The documentation compares use cache to a component-level Cache-Control header.</p><p>A CDN cannot know that a cached component inside a page was built for one tenant. Cache Components add a layer whose correctness depends on your code.</p><h3>The semantics that matter for review</h3><ul><li><strong>Data or UI can be cached.</strong> Cached functions and components must be async.</li><li><strong>Give every cache a lifetime.</strong> The documentation recommends pairing each directive with cacheLife. Without one, a default profile applies: five minutes stale on the client, fifteen minutes before server revalidation, no time-based expiry.</li><li><strong>Tag what you will invalidate.</strong> cacheTag labels an entry. After a Server Action, updateTag expires it so users see their own change. In a webhook or Route Handler, revalidateTag takes a cache profile as its second argument.</li><li><strong>Request data stays outside.</strong> A cached scope cannot read cookies(), headers() or searchParams. Read them outside and pass values in as arguments, or use the use cache: private variant, which keeps results in the browser.</li><li><strong>Dynamic parts sit behind Suspense.</strong> The fallback joins the static shell; the real content streams in.</li><li><strong>Keys come from inputs.</strong> The key includes the build ID, a function ID, the arguments and any variables captured from the surrounding scope. That is where review effort belongs.</li></ul><h3>Why cache keys are security boundaries</h3><p>Caching never replaces authorisation. A cached function returns what it computed for a key. If the key does not capture who the data is for, someone else can receive it.</p><p>Consider a <strong>hypothetical</strong> CMS-backed site. Articles, product pages and navigation come from the CMS and are identical for every visitor. A banner and an account menu depend on the signed-in user. A regional variant depends on the locale in the URL.</p><p>A reasonable split: public CMS content goes into cached functions with a lifetime and a tag, invalidated by a webhook when an editor publishes. The personalised strip reads the session outside any cache, sits behind Suspense, and passes only what it needs inward. Authorisation checks run on every request, whatever the cache holds.</p><p>The mistakes to hunt for: a cached function that quietly depends on a tenant, locale or user the key does not capture; one that is handed a user identifier but returns data meant for a wider audience; a tag that is too broad or never invalidated.</p><p>The November security release shows this is not theoretical.</p><h3>The November release: only the relevant parts</h3><p>On 3 November the team <a href="https://nextjs.org/blog/november-2026-security-release">published fixes</a> in 16.4.5 and 15.5.29: one high and six medium issues. This is not a full patch review. Four items touch the territory of this article.</p><ul><li><strong>unstable_cache combined with </strong><strong>use cache.</strong> The release post rates it high; the GitHub advisory page I read lists it as moderate, so check which rating your process uses. It needs Cache Components and a use cache function that reads a root parameter such as a tenant identifier, called both directly and through unstable_cache. During a cold cache fill, a concurrent call can receive another request&#39;s result. The advisory says avoiding that nesting is a complete workaround. Its range begins at 16.3.0.</li><li><strong>Nested </strong><strong>use cache.</strong> An inner function reads a root parameter and returns a value with pending promises; the outer entry can be stored without that parameter in its key. The advisory range begins at 16.2.0.</li><li><strong>Binary arguments.</strong> Distinct Uint8Array or Blob arguments can share a cache identity, which matters if they distinguish users. The advisory page shows a range from 16.0.0.</li><li><strong>Shared shell poisoning, self-hosted only.</strong> It affects next start with Cache Components on and Partial Prefetching disabled or unset, for a particular route shape: a prerenderable dynamic parameter plus at least two deeper parameters that are never prerendered, with a blocking fallback shell.</li></ul><p>The advisory pages do not fully agree with the release post on fixed versions: one names 16.4.2, two name none. I would not use the advisory ranges to decide you are safe. The post says to upgrade to 16.4.5 or 15.5.29, and npm lists 16.4.5 as latest. (The post&#39;s install example has a stray &quot;for 16.5&quot; comment; 16.5 exists only as a canary build.)</p><p>Three further medium issues concern response framing behind certain proxies, cookie exposure in Edge routes and an SVG check in image optimisation. Check those against your setup too.</p><p>The point: <strong>patching is separate from adopting.</strong> A site that never enables Cache Components still needs this assessment, and a 15.x site has 15.5.29 without moving to 16.</p><h3>A migration inventory</h3><p>Before anyone edits next.config.ts, write down:</p><ul><li><strong>Router and version.</strong> The migration guide requires Next.js 16 and the App Router.</li><li><strong>Rendering configuration.</strong> With the feature on, dynamic, revalidate and fetchCache exports error, and dynamicParams fails the build.</li><li><strong>Runtime.</strong> Cache Components needs Node.js and does not support the runtime segment config, including edge.</li><li><strong>Cookies, headers and search parameters:</strong> where each is read.</li><li><strong>Draft Mode.</strong> Cached functions re-run on every request while it is on; confirm editors still see drafts.</li><li><strong>Content mutations.</strong> Server Actions, webhooks and jobs each need a tag and an invalidation path.</li><li><strong>Existing caches.</strong> fetch options and unstable_cache keep working, and migrating the latter is optional. Note any overlap with new use cache code, given the first advisory above.</li><li><strong>Host adapter and cache storage.</strong> By default use cache is stored in memory: entries typically do not persist across serverless requests, but persist on a self-hosted server. use cache: remote and custom cache handlers exist for shared storage. Nothing carries across deployments.</li></ul><p>Behaviour changes too. Next.js keeps recently visited routes mounted but hidden using React’s Activity, so state in dropdowns and forms can persist across navigation.</p><h3>Adopt now, pilot or defer?</h3><p>My recommendation, not a rule:</p><ul><li><strong>Adopt now</strong> for new projects, or small sites on a Node.js host you control with few mutations and good tests.</li><li><strong>Pilot in staging</strong> for multi-tenant, personalised or content-heavy sites. The flags are global, so evaluate in an isolated staging deployment. The migration guide documents an instant = false opt-out so routes that are not ready can wait while others are converted.</li><li><strong>Defer the feature, not the patch</strong> if you rely on Edge runtime routes, have an unconfirmed host adapter or lack testing capacity. Stay on a patched release.</li></ul><p>Host support needs its own check. OpenNext’s Cloudflare documentation, when I read it on 4 November, listed Partial Prerendering and use cache among features not yet supported. The page is undated and mentions much older Next.js versions, so it may be stale, but it is a reason to test an adapter rather than assume.</p><h3>Freshness, cost and rollback</h3><p>Write acceptance criteria before the pilot.</p><p><strong>Freshness.</strong> State the maximum acceptable staleness per content type and test publish-to-visible, including tag invalidation.</p><p><strong>Cold and warm requests.</strong> Compare the first request after a deploy with later ones.</p><p><strong>Navigation.</strong> Prefetching trades speed for server work. In its <a href="https://nextjs.org/blog/improving-navigation-experience-in-v0-with-nextjs-16-4">v0 write-up</a>, the Next.js team reports that an early rollout raised prefetch requests to one route 2.9 times until prefetching waited for hover, focus or touch. Afterwards they report total daily cost 2.9 percent lower, and caution that traffic and unrelated deployments changed during the period. That is one documented implementation, not a guarantee for yours. The team’s <a href="https://x.com/nextjs/status/2111137399489405166">X post</a> summarised it as “at no added cost”; the blog’s own wording is more careful.</p><p><strong>Server cost.</strong> Compare invocations, CPU and transfer on your own traffic.</p><p><strong>Rollback.</strong> Keep the change in one deployable commit. Reverting the flags will not undo segment-config removals, so plan for both.</p><h3>What to watch for Next.js 17</h3><p>Announced: Cache Components becomes the default, and the documentation says it and Partial Prefetching will be enabled everywhere in the next major release, with the options removed. No release date has been given, so I would not plan around one.</p><p>My inference, not a stated fact: sites that arrive at 17 with explicit lifetimes, tagged content and navigation tests will have less to untangle. Watch your host’s adapter support, the migration guide and each monthly security release.</p><h3>The short version</h3><ol><li>Patch to a current release and assess the November advisories against your actual setup.</li><li>Inventory router, version, config, runtime, request data, mutations and hosting.</li><li>Pilot in isolated staging with written acceptance criteria.</li><li>Treat cache keys as security boundaries; keep authorisation outside caches.</li><li>Decide on evidence, not on the headline.</li></ol><p>Prateeksha Web Design works on Next.js sites and migrations. If you are weighing this for a production site, which part of your inventory looks hardest?</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c35ab0c40cf5" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Next.js Now Has a Patch Calendar. Does Your Business Have a Patch Owner?]]></title>
            <link>https://medium.com/@prateekshawebdesign/next-js-now-has-a-patch-calendar-does-your-business-have-a-patch-owner-5e75f2060391?source=rss-c4c4cda9f964------2</link>
            <guid isPermaLink="false">https://medium.com/p/5e75f2060391</guid>
            <category><![CDATA[nextjs]]></category>
            <category><![CDATA[web-development]]></category>
            <category><![CDATA[website-maintenance]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <dc:creator><![CDATA[Prateeksha Web Design]]></dc:creator>
            <pubDate>Wed, 04 Nov 2026 10:13:24 GMT</pubDate>
            <atom:updated>2026-11-04T10:13:24.765Z</atom:updated>
            <content:encoded><![CDATA[<h4>A predictable security release schedule only helps if someone owns the work it creates. Here is a practical way to assign it.</h4><figure><img alt="Illustration of a row of calendar pages above a ring of arrows, joined by a single amber line to one lit desk and chair among several dim, empty workstations." src="https://cdn-images-1.medium.com/max/1024/1*8DK_RK-RUG5lhiJ9Xp7AsA.png" /><figcaption>Cover illustration generated with ChatGPT image generation.</figcaption></figure><p>On 3 November the Next.js team is scheduled to publish fixes for seven medium-severity vulnerabilities. It said so <a href="https://nextjs.org/blog/upcoming-nextjs-security-release-november-2026">a week ahead</a>, which is the whole point of the security release programme it introduced in July: notice, so that teams can plan.</p><p>Notice only helps if someone is there to receive it. A calendar tells you when work will arrive. It does not tell you whose work it is.</p><p>If a Next.js site generates leads or sales for your business, that gap is the subject of this piece. The framework has become more predictable. Whether your site is kept safe is still a question about people.</p><h3>What the programme promises</h3><p>When the team <a href="https://nextjs.org/blog/next-security-release-program">announced the programme in July</a>, it described a simple change. Historically, security fixes were ad-hoc patches with no advance notice. Now, roughly once a month, the team publishes a notice on its blog with the expected release timeline and the highest severity among the vulnerabilities involved.</p><p>The lead time has a second purpose. It lets the team coordinate with hosting providers and other platform partners, who may be able to deploy mitigations such as firewall rules for applications that have not yet been patched.</p><p>Urgent disclosures, and anything already being exploited, still get out-of-band patches.</p><p>The team’s stated reason is that vulnerability research across the industry is rising, helped by AI-assisted discovery, and that it runs similar tooling against its own code. That is the team’s account of its situation, and I treat it as context rather than as a measured trend.</p><p>Notice what is being promised: timing. The programme does not tell you which fixes touch your site, who on your side reads the announcement, or when your version gets updated.</p><h3>The calendar since July has not been tidy</h3><p>A schedule suggests order. The record is more instructive than that.</p><p>The first scheduled release in July <a href="https://nextjs.org/blog/july-2026-security-release">fixed nine issues</a>, four high and five medium, in 16.2.11 and 15.5.21. In August the team <a href="https://nextjs.org/blog/nextjs-security-release-august-2026-update">brought the release forward by a day</a> after finding a second critical issue, so that users would only need to upgrade once. The <a href="https://nextjs.org/blog/august-2026-security-release">August release</a> shipped in 16.3.3 and 15.5.24.</p><p>On 22 September there was an <a href="https://nextjs.org/blog/nextjs-security-update-september-22-2026">out-of-band update</a> for a critical issue in the Node.js ImageResponse implementation. The scheduled <a href="https://nextjs.org/blog/september-2026-security-release">September release</a> followed on 30 September with seven fixes, after two others, one critical and one high, were postponed because of upstream coordination. Those two <a href="https://nextjs.org/blog/nextjs-security-update-october-14-2026">arrived on 14 October</a>, in another out-of-band update, in 16.4.1 and 15.5.28.</p><p>So the dates moved, the counts changed, fixes were deferred, and the patched version number for any given release line kept changing. None of this is a criticism. Coordinating fixes across upstream libraries is hard. But a plan that assumes one tidy monthly event will not survive contact with this record. It needs slack.</p><h3>“Are we affected?” is the expensive question</h3><p>Several of the recent issues were critical, and some involved remote code execution. That term means an attacker can make your server run commands of their choosing, here without needing to log in. It is the category of flaw that deserves a fast response.</p><p>But look at how the 14 October advisories describe themselves. One, in the App Router with Cache Components, involved use cache functions that return multiple typed arrays; the advisory says there is no known workaround. The other concerned the Image Optimization API handling SVG on glibc-based Linux, with a workaround of making sure the sharp library is at version 0.35.5 or later. A third, high-severity issue, a denial of service through Server Actions, did not affect applications using the Pages Router or without Server Actions.</p><p>Earlier advisories are similarly conditional. The August remote code execution flaw in a Pages Router plus App Router setup applied to servers using a Windows filesystem, not Linux or macOS. September’s image-optimisation request forgery issue did not affect applications without images.remotePatterns configured. One September cache-poisoning issue concerned self-hosted Pages Router applications and not those deployed on Vercel.</p><p>Two things follow, and they pull in opposite directions.</p><p>First, not every advisory applies to every site. Whether one applies depends on the framework version, the router, the bundler, the features switched on, the operating system and the hosting. Someone has to know those facts about your site. Most business owners do not, and an agency that has moved on to other projects may no longer hold them either.</p><p>Second, applicability decides urgency, not whether to patch. The October advisory is blunt about it: upgrade immediately even if your application has no similar functions. A sensible reading is that assessment tells you how fast to move, while the default is still to move.</p><h3>Four jobs the calendar does not do</h3><p>Behind the phrase “keeping the site patched” sit four separate jobs.</p><p><strong>Monitoring.</strong> Someone reads the advance notice and the advisories when they appear.</p><p><strong>Assessment.</strong> Someone compares each advisory with what the site actually runs.</p><p><strong>Deployment.</strong> Someone updates, tests and ships.</p><p><strong>Verification.</strong> Someone confirms the fix is live and nothing else broke.</p><p>A calendar helps with the first job only. It is also easy to assume that managed hosting covers the rest. Hosting can reduce exposure to some issues, and the programme’s coordination with providers exists for that reason. But in my reading, a host does not take ownership of your application’s version, your configuration or your testing. Whether it does for your site is a question for your contract, not an assumption.</p><h3>A patch-ownership framework</h3><p>Nothing here is exotic. It is a short list of decisions, each of which should have an answer written down.</p><p><strong>1. A named owner and a backup.</strong> One accountable person, not a team alias. It might be an in-house developer, your agency or a maintenance provider. Name a deputy too, because advisories do not wait for holidays. If the answer is “the agency”, check the contract actually says so, and for how long.</p><p><strong>2. An inventory.</strong> Record the deployed Next.js and React versions, the router in use, the bundler, whether Server Actions, image optimisation, use cache or Cache Components are enabled, where the site is hosted and on what operating system. Update it whenever the site changes. This is the page that makes the next step possible.</p><p><strong>3. Monitoring and an applicability call.</strong> The owner reads each advance notice and advisory, then writes down a short judgement: affected, not affected or unsure, with the reason. “Unsure” should default to treating it as affected until someone can show otherwise.</p><p><strong>4. A routine window and an urgent route.</strong> Agree a regular slot for ordinary patch releases, and a separate escalation path for critical or actively exploited issues. The time targets are policy choices for you to make, not standards. As one example only: assess a critical advisory the same working day, ship a tested fix within a few days, and put medium issues into the next routine window. A small brochure site and a transactional platform may reasonably choose different numbers.</p><p><strong>5. Staging, rollback and proportionate checks.</strong> Apply the update to a staging copy first. Know how to go back to the previous build before you deploy. Then test in proportion to the change: for a patch release, that is usually the main journeys (enquiry form, checkout, login, key pages) rather than a full regression of everything.</p><p><strong>6. Confirmation and a record.</strong> After deployment, confirm the production site is actually running the patched version, not just that the pull request merged. Then keep a short note: which advisory, the applicability decision, the date shipped, who confirmed. This is a few lines per release. It is also what you will want if a customer, insurer or auditor asks.</p><h3>A hypothetical</h3><p><em>The following scenario is invented for illustration.</em></p><p>A professional services firm launches a Next.js site in the spring. The agency delivers on time and the retainer ends with the final invoice. Nobody is named as patch owner. The website uses Server Actions for its enquiry form and image optimisation for a few remote images.</p><p>Months later the monthly advance notice goes out. The agency’s developer who built the site has left, and the notice reaches no one who reads it. The site keeps running, the form keeps submitting, and nothing looks wrong. Whether any given advisory applies is unknown, because nobody knows which options are switched on.</p><p>Compare a second firm with the same site and a named owner, a deputy and a one-page inventory. The owner reads the notice, checks the inventory, notes that two advisories do not apply and one does, and schedules the update into the next window. Nobody is heroic. The work is dull and finished.</p><p>The technology in the two cases is identical. The difference is ownership.</p><h3>What 16.4 adds, and what it doesn’t</h3><p>Next.js 16.4, <a href="https://nextjs.org/blog/next-16-4">released on 6 October</a>, recommends Cache Components for every app, enables them by default in new projects, and says they will become the default in Next.js 17. My own inference, and nothing more, is that newer programming models deserve closer attention from whoever owns patches, since several recent advisories concerned use cache behaviour. Adopting a feature and maintaining it are one decision.</p><p>The same release introduced experimental.agentUpgrade. Its default policy, security, reminds you or your coding agent during next dev and next build when an upgrade addresses known vulnerabilities affecting your installed version.</p><p>That is a useful nudge, and it is labelled experimental. But a reminder in a terminal reaches only the person who runs that terminal. It does not assess applicability for the business, schedule the window or confirm production. It supports an owner. It does not replace one.</p><h3>One question for this week</h3><p>Ask it plainly: who is responsible for applying the next Next.js security release to our site, and who covers for them?</p><p>If the answer comes back quickly, ask to see the inventory. If the answer is unclear, you have found the first task, and it costs nothing to do. A calendar is a gift only to the people who are looking at it.</p><p>If you maintain Next.js sites, I would be interested to hear what your patch window and escalation targets look like, and what you have found hardest to keep in the inventory.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5e75f2060391" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Next.js vs WordPress in 2026: The Maintenance Cost Nobody Quotes]]></title>
            <link>https://medium.com/@prateekshawebdesign/next-js-vs-wordpress-in-2026-the-maintenance-cost-nobody-quotes-43d0aaf800c5?source=rss-c4c4cda9f964------2</link>
            <guid isPermaLink="false">https://medium.com/p/43d0aaf800c5</guid>
            <dc:creator><![CDATA[Prateeksha Web Design]]></dc:creator>
            <pubDate>Mon, 02 Nov 2026 11:13:35 GMT</pubDate>
            <atom:updated>2026-11-02T11:13:35.918Z</atom:updated>
            <content:encoded><![CDATA[<h4>Both platforms now ship security fixes on a rhythm. The cost difference is who has to act on each one, and what can break when they do.</h4><figure><img alt="Two identical glass office buildings side by side, each cut away to show a different underground pipe network: a few large conduits with valves on the left, a dense web of small pipes on the right." src="https://cdn-images-1.medium.com/max/1024/1*pgN_6g1Uf5RzcQpbSgRuIA.png" /></figure><p>Ask three agencies to quote a new website and you will usually get three build prices and almost no information about the fourth year.</p><p>That is the part I would examine first. When a quote compares Next.js with WordPress, the line that matters is rarely the build. It is the cost of owning change afterwards: watching for updates, applying them, testing that nothing broke, fixing what did, writing down what was done, and recovering when something goes wrong.</p><p>I do not think “which platform is cheaper?” is a useful question. A more useful one is: where does the maintenance work land, and is anyone assigned to it?</p><p>This piece is for founders, business owners and technical decision-makers who are choosing between the two for a marketing, lead-generation or content-led website. It is not a framework showdown for developers.</p><h3>The comparison is not symmetrical</h3><p>Next.js is an application framework. WordPress is a content management system with a large extension ecosystem. They are different kinds of thing, and I will not pretend otherwise.</p><p>Businesses still compare them because an agency may propose either for the same brief. So I compare the maintenance each asks of the owner, which is what the owner pays for.</p><h3>How I evaluate maintenance cost</h3><p>I set out the criteria before looking at either platform, so they cannot be bent towards a conclusion I already hold:</p><ul><li>Initial build or migration effort</li><li>Planned patching</li><li>Dependency or extension surface</li><li>Staging and regression testing</li><li>Hosting responsibility</li><li>Content-team independence</li><li>Specialist availability</li><li>Security monitoring</li><li>Backups and recovery</li><li>Integration and customisation risk</li><li>Incident response</li><li>Exit or migration cost</li><li>Predictability of maintenance</li></ul><p>My working framework, which is a planning aid and not a measured formula, is:</p><blockquote><em>Total cost of ownership = initial build or migration + hosting, licences and infrastructure + planned maintenance labour + monitoring, backup and recovery + change and integration work + an allowance for incidents + eventual migration or exit cost.</em></blockquote><p>It has no numbers, deliberately. I explain why below.</p><h3>What I found for Next.js</h3><p>I read the official Next.js blog and documentation on 2 November 2026. These are the facts I would put in front of an owner.</p><p><strong>Security fixes now follow a calendar.</strong> On 13 July the Next.js team <a href="https://nextjs.org/blog/next-security-release-program">announced a formal security release programme</a> with roughly monthly, pre-announced releases, while keeping ad-hoc patches for urgent problems. Since then there have been scheduled releases in July, August and September, an out-of-band update on 22 September, and another on <a href="https://nextjs.org/blog/nextjs-security-update-october-14-2026">14 October</a> that fixed two critical remote-code-execution issues and one high-severity denial-of-service issue. A <a href="https://nextjs.org/blog/upcoming-nextjs-security-release-november-2026">further scheduled release</a> covering seven medium-severity issues is announced for 3 November.</p><p><strong>Support is limited to two lines.</strong> The <a href="https://nextjs.org/support-policy">support policy</a> gives security patches to Active LTS (16.x) and Maintenance LTS (15.x, with security releases extended to 31 January 2027). Older versions are unsupported and, in the policy’s words, rarely patched.</p><p><strong>Applying a fix is a development task.</strong> Nothing updates itself. Someone changes the dependency, rebuilds, tests and deploys. The framework is trying to reduce that effort: version 16.4 adds next upgrade --agent and an option that reminds you about security upgrades. A reminder is not an owner.</p><p><strong>The runtime is part of the bill.</strong> Next.js’s installation page lists Node.js 20.9 as the minimum. The <a href="https://raw.githubusercontent.com/nodejs/Release/main/schedule.json">Node.js release schedule</a> shows Node 20 reached end-of-life on 30 April 2026. I read that as a prompt to check the runtime, not just the framework. That is my inference from two documents, not a finding about any site.</p><p><strong>Self-hosting adds duties.</strong> The <a href="https://nextjs.org/docs/app/guides/self-hosting">self-hosting guide</a>, updated 15 October, recommends a reverse proxy, notes that the cache is per-instance by default, and describes the extra work for multiple instances: a shared cache handler, a common encryption key for Server Functions, and a deployment identifier to avoid version skew. Some advisories also depend on hosting. A September cache-poisoning issue, for example, <a href="https://nextjs.org/blog/september-2026-security-release">affected self-hosted Pages Router applications but not those deployed on Vercel</a>. I did not evaluate managed-platform contracts, so I cannot say what any particular provider covers.</p><h3>What I found for WordPress</h3><p>I read the WordPress.org news, documentation and security pages the same day.</p><p><strong>It also patches often.</strong> The <a href="https://wordpress.org/news/category/releases/">release archive</a> shows seven security-related releases between 17 July and 28 October (7.0.2 to 7.0.4 and 7.1.1 to 7.1.4), alongside the 7.1 core release on 19 August. WordPress 7.0.2 <a href="https://wordpress.org/news/2026/07/wordpress-7-0-2/">fixed a critical and a high-severity issue</a>, and the team enabled forced updates through the auto-update system for affected sites.</p><p><strong>Core updates itself, mostly.</strong> According to the <a href="https://developer.wordpress.org/advanced-administration/upgrade/upgrading/">updating documentation</a>, minor core releases, which include security fixes, are applied automatically by default on most sites. Plugins and themes are not updated broadly by default. That is a significant gap, because extensions are where a typical site’s custom behaviour lives.</p><p><strong>Only the latest version is officially supported.</strong> The <a href="https://wordpress.org/about/security/">security page</a> says so, adding that the security team backports critical fixes to older versions as a courtesy. It also describes coordinating rollouts with major hosts and developing web application firewall mitigations. It says little about the site owner’s duties, which tells me those duties are left to the owner.</p><p><strong>The stack underneath matters.</strong> The <a href="https://wordpress.org/about/requirements/">requirements page</a> recommends PHP 8.3 or greater, and warns that older PHP and MySQL versions are end-of-life and may expose a site to vulnerabilities.</p><p><strong>Recovery depends on you.</strong> The documentation says to back up before updating and calls verifying the backup essential.</p><h3>Where the work lands</h3><p>My reading, which is interpretation and not measurement:</p><p><strong>How the work lands, criterion by criterion (Next.js first, WordPress second):</strong></p><ul><li><strong>Patch cadence:</strong> Monthly scheduled, plus urgent releases <em>vs</em> Frequent security and maintenance releases</li><li><strong>Core fixes applied:</strong> By a developer, through build and deploy <em>vs</em> Mostly automatic for minor core releases</li><li><strong>Extension surface:</strong> npm dependencies plus your own code <em>vs</em> Plugins and themes, often many, mostly not auto-updated</li><li><strong>Runtime responsibility:</strong> Node.js, build tooling, hosting configuration <em>vs</em> PHP, database, web server, hosting</li><li><strong>Main regression risk:</strong> Build breaks, behaviour changes in caching or rendering <em>vs</em> Plugin or theme conflicts after an update</li><li><strong>Content independence:</strong> Depends on how the content layer is built <em>vs</em> Strong out of the box</li><li><strong>Specialist needed:</strong> Engineers comfortable with React and deployment <em>vs</em> WordPress developers and, often, a host or security specialist</li><li><strong>Predictability:</strong> Planned and lumpy <em>vs</em> Frequent and smaller, with plugin surprises</li></ul><p>I would not read this as a score. Next.js concentrates the work in engineering. WordPress spreads it across core, extensions, runtime and host. Neither is free, and neither is automatically cheaper.</p><h3>Where the answer flips</h3><p><strong>Brochure or lead-generation sites.</strong> With a handful of pages and a form, the extension surface on either platform can be small. The cost then depends on who is assigned to patching.</p><p><strong>Content-heavy publishing.</strong> A busy editorial team benefits from WordPress’s editing tools. A Next.js build can match that, but someone has to build and maintain the content layer.</p><p><strong>Highly customised applications.</strong> The more bespoke logic and integrations, the more either approach becomes ordinary software maintenance. I would be sceptical of identical maintenance quotes for a simple site and a custom application.</p><p><strong>No internal technical team.</strong> Automatic core updates reduce the owner’s burden on WordPress. On Next.js, without a retained developer or agency, patches simply do not get applied.</p><p><strong>Established development capability.</strong> A team that already runs build pipelines absorbs Next.js maintenance more easily.</p><p><strong>Managed versus self-hosted.</strong> Hosting changes the list of duties on both sides. I would ask for the split in writing.</p><h3>What I would choose, and when</h3><p>I would choose WordPress when the content team needs to work independently, the design is conventional, the extension list can be kept short, and nobody on the owner’s side can be accountable for deployments.</p><p>I would choose Next.js when the site’s behaviour is bespoke, performance and front-end control are central to the brief, and there is a named developer or agency committed to monthly maintenance.</p><p>I would consider a hybrid approach only when editorial independence and front-end control are both required. It carries two maintenance surfaces, so I would price both before agreeing.</p><p>I would reconsider the brief when the owner wants the lowest build price and no maintenance commitment. On either platform, that combination tends to postpone the cost, not remove it.</p><h3>A maintenance contract I would want to see</h3><p>Whichever platform you choose, I would look for:</p><ol><li><strong>A named owner</strong> accountable for updates, with a backup person.</li><li><strong>An inventory</strong> of core, plugins or dependencies, runtime versions and hosting components.</li><li><strong>Monitoring</strong> of the vendor’s release announcements and advisories.</li><li><strong>Staging</strong> that mirrors production, and a rule that updates are tested there first.</li><li><strong>Regression testing</strong> of the pages that bring in enquiries or sales: forms, checkout, tracking.</li><li><strong>Backups</strong> that are verified by restoring, not just by existing.</li><li><strong>Rollback</strong>, with an agreed method and an estimated time.</li><li><strong>Responsibilities</strong> split across agency, host and owner, in writing.</li><li><strong>Reporting</strong>: what was applied, when and what was tested.</li><li><strong>Response expectations</strong> for critical and ordinary updates.</li></ol><p>If you want a reference point, Prateeksha Web Design, the agency I run, describes how it approaches <a href="https://prateeksha.com/website-maintenance">website maintenance</a>. The checklist above should hold whoever you hire.</p><h3>What I cannot responsibly price</h3><p>I will not put a figure on any of this without the actual scope. The inputs that move the number are the number of plugins or dependencies, how much custom code exists, the hosting model, the testing needed for revenue-critical flows, the response times required and the team’s existing skills. Any universal monthly figure for “Next.js maintenance” or “WordPress maintenance” would be invented.</p><p>What I can say is what to ask for: a quote that lists the tasks above, the person responsible for each, and what is excluded.</p><h3>What to watch next</h3><p>Verified: both ecosystems patch frequently. Next.js 16.4 recommends Cache Components and says they will be the default in Next.js 17. WordPress 7.2 is in beta.</p><p>Prediction, which is my view and nothing firmer: vulnerability discovery will stay busy and cadences will stay short, so maintenance will matter more to ownership cost, not less. The Next.js team itself cites rising, AI-assisted vulnerability research as a reason for its new programme.</p><h3>My conclusion</h3><p>The honest comparison is not Next.js against WordPress. It is each platform against the maintenance you are actually willing to fund.</p><p>If you are weighing the two, I would start with that question, and then ask what each proposal would do on 3 November. Prateeksha works with both, on <a href="https://prateeksha.com/nextjs">Next.js</a> and through <a href="https://prateeksha.com/wordpress-customization">WordPress customisation</a>, and my advice is the same for each: do not sign off a build until the maintenance owner has a name.</p><p>How are others splitting these duties between agencies, hosts and owners? I would like to hear where it has worked and where it has fallen apart.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=43d0aaf800c5" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Your Shopify Theme Is No Longer a Design Project — It’s a Conversion Experiment]]></title>
            <link>https://medium.com/@prateekshawebdesign/your-shopify-theme-is-no-longer-a-design-project-its-a-conversion-experiment-cfabb480332b?source=rss-c4c4cda9f964------2</link>
            <guid isPermaLink="false">https://medium.com/p/cfabb480332b</guid>
            <category><![CDATA[conversion-optimization]]></category>
            <category><![CDATA[web-development]]></category>
            <category><![CDATA[shopify]]></category>
            <category><![CDATA[user-experience]]></category>
            <category><![CDATA[ecommerce]]></category>
            <dc:creator><![CDATA[Prateeksha Web Design]]></dc:creator>
            <pubDate>Thu, 29 Oct 2026 10:57:29 GMT</pubDate>
            <atom:updated>2026-10-29T10:59:09.244Z</atom:updated>
            <content:encoded><![CDATA[<h4>What Shopify’s 2026 Rollouts, analytics changes, and merchant speed complaints mean for custom themes, product-page UX, development costs, and revenue.</h4><figure><img alt="Split-screen ecommerce storefront illustrating Shopify theme testing, product-page performance, and conversion measurement." src="https://cdn-images-1.medium.com/max/1024/1*BrUKVcHk9SEswXBI1prCww.png" /></figure><p>For years, Shopify merchants approved a theme in a staging link, launched it, and hoped it would sell. In 2026, Shopify turned that guess into an experiment: entire themes and checkout configurations can now be scheduled, gradually released, and A/B tested. But a testing button does not fix weak product UX, slow JavaScript, unreliable measurement, or a development budget spent on the wrong things.</p><p>That matters for anyone planning Shopify and ecommerce work this quarter. The redesign brief is changing from “make it look better” to “make it measurable, then prove it improved.” Here is what that shift means in practice, where the new tooling stops, and how to spend a theme budget sensibly.</p><h3>The 2026 shift: themes can now be treated as measurable experiments</h3><p>On June 5, 2026, Shopify’s changelog announced that <a href="https://changelog.shopify.com/posts/schedule-publish-and-a-b-test-new-themes-and-checkout-and-customer-account-configurations">Rollouts</a> can schedule a new theme or checkout configuration for a set time, swap one in temporarily with an automatic revert, release it gradually to a percentage of visitors, and A/B test two entirely different versions. Shopify creates a copy of your published theme automatically, so the live version can keep being edited while an experiment runs. You will find it under Markets &gt; Rollouts in the admin.</p><p>A September 22 <a href="https://changelog.shopify.com/posts/plan-launches-events-and-tests-with-clearer-and-more-granular-controls-in-rollouts">update</a> redesigned the setup flow: you choose timing and traffic percentage, see how traffic is split between versions, and are warned about conflicting settings. One rollout can now bundle theme, checkout, and customer-account changes.</p><p>The practical consequence is cultural as much as technical. A redesign used to be a single high-stakes launch. It can now be a controlled release with a way back. On X, CRO specialists such as <a href="https://x.com/oliverkenyon/status/2115487717588914533">Oliver Kenyon</a> were already reading Shopify’s Fall ’26 announcements through an experimentation lens within hours of release. That is practitioner commentary, not evidence, but it shows where the conversation has moved.</p><h3>What Shopify Rollouts can and cannot prove</h3><p>Documented capability and statistical reliability are different things. Shopify’s changelog describes what Rollouts <em>does</em>. It does not promise that every store will get a trustworthy answer from it. The following points are my inference, not Shopify’s claims.</p><ul><li><strong>Traffic volume decides whether a test can conclude.</strong> A store with a few hundred orders a month may wait a long time for a difference to separate from noise.</li><li><strong>Two whole themes differ in many ways at once.</strong> If a new theme wins, you will not necessarily know whether the product page, the navigation, the speed, or the imagery did the work.</li><li><strong>Timing matters.</strong> Running a test across a sale, a paid-media burst, or a stock-out can distort a result no matter how cleanly traffic was split.</li><li><strong>The changelog does not describe reporting depth.</strong> Before committing a redesign to a test, check what the Rollouts results screen actually reports for your plan, and whether it covers the metric you care about, such as revenue per visitor or margin.</li></ul><p>Used well, Rollouts is a safer launch mechanism and a decent way to compare two directions. It is not a substitute for a hypothesis. If you cannot say what you expect to change and why, a test will only produce a number you cannot interpret. Our approach to <a href="https://prateeksha.com/conversion-rate-optimization">evidence-led conversion optimisation</a> starts from the same premise: verify the measurement first, and run controlled tests only where traffic can support them.</p><h3>Reset your conversion baseline after Shopify’s analytics change</h3><p>Before you test anything, check your yardstick. Shopify updated how it measures sessions, with the change rolling out between September 21 and 23, 2026. According to Shopify’s <a href="https://help.shopify.com/en/manual/reports-and-analytics/discrepancies/session-measurement-update">help documentation</a>, sessions now continue through customer activity instead of ending at midnight UTC, end after 30 minutes of inactivity, include sessions without a pageview (such as a direct checkout from a cart link), and exclude identified bot sessions from session-related reports by default.</p><p>Because conversion rate is calculated from sessions, it can move even when orders and sales do not. More sessions can lower the rate; fewer, after bot filtering, can raise it. Orders, sales, and customer counts are not affected.</p><p>The implication is simple and easy to forget:</p><ul><li>Treat post-update data as a new baseline for session-based metrics.</li><li>Do not compare a post-update conversion rate with last quarter’s figure and declare a win or a loss.</li><li>Read sessions alongside orders and revenue, which did not change definition.</li><li>Check custom reports and any apps that rely on Shopify session data.</li></ul><p>If your redesign case was built on “our conversion rate fell,” re-test that claim against the new baseline first. Sometimes the problem is the measurement, not the store.</p><h3>Product-page UX: resolve objections before adding visual novelty</h3><p>Most product pages lose buyers to unanswered questions, not to unfashionable typography. A shopper deciding on a purchase is quietly checking a short list:</p><ul><li><strong>Fit and suitability:</strong> Is this right for me? Sizing, compatibility, materials, and use cases should be answerable without leaving the page.</li><li><strong>Variants:</strong> Are options clear, and does the image change when they select one?</li><li><strong>Delivery:</strong> When will it arrive, and what will it cost?</li><li><strong>Returns:</strong> What happens if it is wrong?</li><li><strong>Trust:</strong> Are there reviews, clear contact details, and a business that looks real?</li><li><strong>Payment:</strong> Can I pay the way I prefer?</li><li><strong>Availability:</strong> Is it in stock, and will it be?</li></ul><p>Shopify’s <a href="https://www.shopify.com/news/fall-26-edition-merchant">Fall ’26 edition</a>, announced on October 28, reinforces why the last two items deserve attention. Shopify added seven local payment methods to Shopify Payments (Alipay, WeChat Pay, MB WAY, OXXO, Satispay, Scalapay, and GrabPay), most of which also work with Shop Pay. It also expanded Shop Promise, a free program for eligible US stores that shows trusted delivery and returns badges on product pages, at checkout, and in the Shop app. Both are about reducing doubt at the moment of decision. Whether either is relevant depends on your markets and eligibility, so check before designing around them.</p><p>The order of work matters. Fix the answers first, then the aesthetics. A <a href="https://prateeksha.com/blog/ecommerce-web-design-priorities">product-page UX checklist</a> built around the objections above will usually deliver more than a new typeface, and it gives any later A/B test a sharper hypothesis.</p><h3>Standard, modified, or custom Shopify theme?</h3><p>Custom is not automatically better. The right choice depends on what your store needs that a theme cannot already do, and on who will maintain it.</p><h4>Standard theme</h4><p>Best when your catalogue is straightforward and the theme’s existing sections cover your product story. It is the fastest to launch, benefits from the theme developer’s updates, and carries the least maintenance risk. The trade-off is sameness and limits: if your product needs a configurator, comparison layout, or unusual merchandising, you will be fighting the theme.</p><h4>Modified theme</h4><p>Best when a good theme gets you most of the way and you need targeted changes: a custom product-page section, a revised collection filter, a tuned cart. It keeps the benefits of a maintained foundation while fixing the specific places where your store differs. The risk is drift, where many small edits make future theme updates harder.</p><h4>Fully custom theme</h4><p>Best when your catalogue structure, content model, or brand experience genuinely cannot be served by modifying an existing theme. It offers the most control and the most responsibility: you own the performance, accessibility, testing, and upgrade path. If you are weighing <a href="https://prateeksha.com/shopify">Shopify development and theme customisation</a> options, ask for a clear reason why custom is needed. “It will look unique” is not enough on its own.</p><p>A useful rule: start with the least custom option that can solve your actual conversion problem, then prove it before spending more.</p><h3>What Shopify development actually costs</h3><p>No honest article can give a universal price for a Shopify build, because the work varies far more than the platform does. What you can do is understand the cost drivers, so quotes become comparable:</p><ul><li><strong>Catalogue size and structure:</strong> a 30-product store and an 800-SKU store with complex variants are different projects.</li><li><strong>Custom sections and templates:</strong> each bespoke component needs design, build, and testing.</li><li><strong>Data readiness:</strong> clean product data, imagery, and copy save time; messy data consumes it.</li><li><strong>Integrations:</strong> ERP, inventory, shipping, subscriptions, reviews, and marketing tools each add scope.</li><li><strong>Migration:</strong> moving products, customers, orders, and redirects from another platform is its own workstream.</li><li><strong>Checkout constraints:</strong> how much of checkout you can customise depends on your Shopify plan.</li><li><strong>Analytics setup:</strong> events, consent handling, and reporting that you can trust.</li><li><strong>QA and accessibility:</strong> device testing and accessible markup take time that is easy to under-budget.</li><li><strong>Performance work:</strong> auditing and removing what slows pages down.</li><li><strong>Ongoing maintenance:</strong> updates, monitoring, and iteration after launch.</li></ul><p>When you compare proposals, check that each one prices the same items. A cheap quote that omits migration, analytics, and QA is not cheaper; it has simply moved the cost later. Prateeksha’s own <a href="https://prateeksha.com/pricing">scope and development costs</a> page works the same way, with pricing depending on factors such as product count, integrations, and content readiness, confirmed in writing after a short discovery conversation.</p><h3>The hidden conversion cost of scripts, apps, and page builders</h3><p>Every script that loads on a storefront competes for the browser’s attention. Apps, tracking tags, chat widgets, and page builders can all add JavaScript that affects Largest Contentful Paint (how quickly the main content appears) and Interaction to Next Paint (how responsive the page feels when someone taps). Google’s <a href="https://web.dev/articles/vitals">Core Web Vitals guidance</a> uses 2.5 seconds for LCP and 200 milliseconds for INP as “good” thresholds, measured on real users at the 75th percentile.</p><p>It is tempting to read that as “apps are bad” or “page builders are slow.” Neither is a safe general claim. Impact varies by app, by how it loads, by the theme, and by the device. Any speed comparison you read, including community tests of page builders, reflects one store, one setup, and one set of conditions; treat it as a prompt to test your own store, not a benchmark.</p><p>A sound approach:</p><ol><li>Measure real-user data first, not just a single lab score.</li><li>List every app and script, and what revenue or insight each one earns.</li><li>Duplicate the theme, remove or defer suspects one at a time, and compare.</li><li>Test representative journeys on mobile: home, collection, product, cart, and checkout.</li></ol><p>If you would rather have this done as a defined piece of work, our <a href="https://prateeksha.com/speed-optimization">Shopify speed improvements</a> service covers audits, Core Web Vitals fixes, and monitoring. Either way, remember that speed work and conversion work are connected: a faster page that still hides its delivery and returns information will not convert much better.</p><h3>A 90-day Shopify optimization roadmap</h3><h4>Days 1–15: Measurement audit</h4><p>Confirm which conversion events fire and where. Document your new post-update baseline for sessions, conversion rate, add-to-cart rate, and revenue per session. Note seasonality and campaigns that could contaminate comparisons.</p><h4>Days 10–30: Product-page diagnosis</h4><p>Review top product pages against the objection list. Use session recordings, customer-service questions, and reviews to find what shoppers cannot answer on the page. Write specific hypotheses, such as “shoppers cannot find delivery cost before checkout.”</p><h4>Days 20–40: Performance audit</h4><p>Run real-user and lab tests on mobile across the five key journeys. Inventory apps and scripts. Rank issues by likely effect on LCP and INP.</p><h4>Days 35–60: Prioritised implementation</h4><p>Fix the highest-value, lowest-risk items first. Decide whether a standard, modified, or custom approach fits what you found. Resist bundling unrelated changes, because it makes results harder to read.</p><h4>Days 55–80: Controlled rollout</h4><p>Use Rollouts to release to a portion of traffic, or to A/B test where volume allows. Define in advance what you will measure, how long you will run, and what result would change your decision.</p><h4>Days 75–90: Post-launch monitoring</h4><p>Watch performance, errors, and conversion after full release. Record what worked, what was inconclusive, and what to try next. Inconclusive is a valid result; it is better than a false win.</p><p>If you want an outside team involved at any stage, we are open about the limits: no theme, tool, or agency can guarantee a conversion lift. Results depend on demand, price, traffic quality, and many factors outside the storefront.</p><h3>Build a measurable storefront system, not a one-time redesign</h3><p>Shopify’s 2026 tooling makes it easier to ship changes safely and to compare options. It does not make weak product pages persuasive, it does not make heavy scripts light, and it does not repair a baseline that quietly changed in September.</p><p>The stores that benefit most will treat the theme as a living system: a trustworthy baseline, a clear list of shopper objections, a lean script budget, and a habit of releasing changes in controlled steps. The redesign becomes less an event and more an ongoing practice, one that you can defend with evidence rather than taste.</p><p>If you remember one thing, make it this: before you commission a new theme, decide how you will know it worked.</p><p><em>Prateeksha Web Design is a Mumbai-based web design and development agency. This article reflects publicly available information as of October 29, 2026 and does not promise any specific conversion or revenue outcome.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cfabb480332b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Dark Mode That Sells: A Practical Guide for Small Business Websites]]></title>
            <link>https://medium.com/@prateekshawebdesign/dark-mode-that-sells-a-practical-guide-for-small-business-websites-1750c796cad4?source=rss-c4c4cda9f964------2</link>
            <guid isPermaLink="false">https://medium.com/p/1750c796cad4</guid>
            <dc:creator><![CDATA[Prateeksha Web Design]]></dc:creator>
            <pubDate>Thu, 29 Oct 2026 09:47:37 GMT</pubDate>
            <atom:updated>2026-10-29T09:47:37.080Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6DjFbFSZTikJa-l_4XqMFg.jpeg" /></figure><p>won’t drown you in technical details. I’ll walk you through what dark mode actually means, where people typically mess it up, and a step-by-step framework you can follow to add a polished dark theme to your website. You’ll get practical tips for colors, accessibility, rollout, and simple ways to test results — plus a short example of a small shop owner who made the switch and saw measurable improvements.</p><p>By the end, you’ll know whether dark mode fits your brand, what to do first, and the exact next steps to move forward this week.</p><h3>Where most people go wrong</h3><p>Before you start, it helps to know the common mistakes so you don’t repeat them.</p><p><strong>- Thinking dark mode is just “invert colors.” </strong><br> Many sites simply flip white to black and black to white. That creates harsh contrast, washed-out images, and unreadable text. Dark mode needs its own color choices, spacing, and touch-ups.</p><p><strong>- Designing for aesthetics only, not readability. </strong><br> A sleek design is great, but if your headings, buttons, or links vanish into the background at night you’ve lost the point. Accessibility and legibility must come first.</p><p><strong>- Rolling it out everywhere at once without testing. </strong><br> Designers sometimes launch a global dark mode and then get feedback about broken icons, logos, or poor contrast. Gradual rollout with user testing prevents that.</p><h3>Main framework — 5 simple steps to a dark-mode-ready website</h3><p>Follow these five pillars. Think of them as a checklist you can actually use.</p><h4>1. Decide if dark mode suits your brand and users</h4><p>Short explanation: Dark mode works great for visual, premium, and tech-forward brands. It’s less suitable for long-form reading sites where high brightness helps readability.</p><p><strong>Practical tips:<br></strong>- Look at your analytics: Do many users visit from mobile at night? If yes, dark mode can help.<br>- Consider imagery: If your product photos are dark or moody (jewelry, fashion, art), dark mode enhances them. If your photos are light and airy (bakeries, daytime services), test before committing.<br>- Ask customers: A simple poll in your newsletter or a website survey can reveal whether they want a dark option.</p><h4>2. Build a thoughtful color palette (don’t just use black and white)</h4><p>Short explanation: Dark themes need softer blacks, off-whites, and carefully tuned accents so elements remain distinguishable and comfortable to view.</p><p><strong>Practical tips:<br></strong>- Use dark gray backgrounds (#111–#222 range) instead of pure black. Pure black increases eye strain and kills depth.<br>- Choose off-white text colors (#E6E6E6 or #F5F5F5) rather than bright white.<br>- Tone down saturated brand colors for accents — bright neon looks harsh on dark backdrops.<br>- Create a simple color map: background, surface (cards), primary text, secondary text, accent. That will keep things consistent.</p><h4>3. Prioritize contrast, typography and spacing</h4><p>Short explanation: Readability is a non-negotiable. Good spacing and font choices often matter more than flashy visuals.</p><p><strong>Practical tips:<br></strong>- Aim for a contrast ratio of at least 4.5:1 for normal text (WCAG recommendation). Use an online contrast checker.<br>- Increase line-height slightly for paragraphs on dark backgrounds to aid reading.<br>- Make buttons and links obvious with subtle shadows, borders, or brighter accents.<br>- Test faulty states like disabled buttons, error messages, and hover interactions — they must remain visible.</p><h4>4. Implement progressively, with tech-light options</h4><p>Short explanation: You don’t need a full engineering sprint to add dark mode. Use a combination of system preference auto-detection and a simple “toggle” so users can choose.</p><p><strong>Practical tips:<br></strong>- Respect system settings: modern browsers offer a “prefers color scheme” setting. If a user prefers dark, default to it.<br>- Provide a visible toggle (e.g., sun/moon icon) and remember the choice for returning visitors.<br>- If you don’t manage your site code, many website builders and themes (Shopify, Squarespace, Wix, WordPress themes) now offer built-in dark variants or plugins.<br>- Keep a single source of truth for colors (CSS variables or theme settings) so you can adjust easily later.</p><h4>5. Test with real people and devices, then iterate</h4><p>Short explanation: Small issues will crop up — icons that disappear, logos that clash — and real users will spot them before you do.</p><p><strong>Practical tips:<br></strong>- Test on multiple devices (iOS, Android, desktop) and in different lights (daylight, dim room).<br>- Run a short user test or get feedback from 5–10 customers. Ask them to complete a task (find product, add to cart) and note friction.<br>- Use grayscale testing to spot low contrast: switch your design to gray and see if elements still read clearly.</p><h3>Short case study — Before → After: Luna Candle Co.</h3><p>Luna Candle Co. is a one-woman online candle shop. Her website was bright white with pastel images. Customers loved the candles, but evening shoppers (a big audience for candles) bounced quickly at night.</p><p><strong>Before:<br></strong>- Bounce rate for evening visitors: 68%<br>- Mobile add-to-cart rate: 1.8%<br>- Brand feedback: “Your site feels too bright at night.”</p><p><strong>What she did:<br></strong>- Followed the five-step framework. Decided dark mode matched her product (cozy candles).<br>- Switched to a dark-gray background, off-white text, and muted amber accents that echoed candlelight.<br>- Added a toggle linked to cookie-based preferences and respected system dark-mode settings.<br>- Tested with eight regular customers and adjusted logo variants to sit well against dark background.</p><p><strong>After (6 weeks later):<br></strong>- Evening bounce rate: 42% (down 26 points)<br>- Mobile add-to-cart rate: 3.4% (up 89%)<br>- Customer feedback: “The site feels cozier and more like real candlelight.”</p><p>Not every result will be this dramatic, but the logic is repeatable: align design with product and user context, then test and refine.</p><h3>Common questions / objections</h3><p><strong>Q: Will dark mode hurt my SEO or page speed?<br></strong>A: No. Color choices don’t directly affect SEO. What matters is user engagement — if dark mode improves time on site or conversion, it can indirectly help. Regarding speed, a properly implemented theme swap has negligible impact. Avoid huge image re-sizes or extra scripts if speed is a concern.</p><p><strong>Q: Won’t implementing two themes double my workload?<br></strong>A: Initially, yes, there’s extra work for design and QA. You can minimize overhead by using design tokens (single color variables) and starting with key pages (home, product, checkout). Many website builders now have built-in theme switches that simplify the process.</p><p><strong>Q: Is dark mode accessible for visually impaired users?<br></strong>A: It can be — if you follow contrast guidelines and test. Dark mode is actually preferred by some users with light sensitivity. The key is ensuring contrast ratios and not relying on color alone to convey information.</p><p><strong>Q: What about product images and logos?<br></strong>A: Prepare alternate logo versions (light and dark) or use a subtle outline/shadow to separate the logo from the background. For images, add gentle card backgrounds or borders so images with transparent areas don’t disappear.</p><p><strong>Q: Should every page offer dark mode?<br></strong>A: Start with critical pages (home, product, checkout, blog). If tests are positive, extend site-wide. For content-heavy pages, test readability carefully — some long-form readers still prefer light mode.</p><h3>Conclusion</h3><p>Dark mode isn’t a gimmick — done right, it improves comfort, highlights imagery, and can lift evening engagement. To get it right:</p><p>- Decide if it fits your brand and customer habits.<br>- Build a considered palette (dark grays, off-whites, muted accents).<br>- Prioritize contrast, spacing, and clear UI elements.<br>- Implement progressively: auto-detect preferences + user toggle.<br>- Test with real customers and iterate.</p><p>This week’s action steps:<br>- Check your analytics for night-time mobile traffic.<br>- Pick one page (home or top product) and create a mockup in dark gray.<br>- Run a quick contrast check and share the mockup with 5 customers for feedback.</p><p>Try one small test this week. You don’t need a full redesign to see if dark mode resonates — and the results could make evening visitors stick around a lot longer.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1750c796cad4" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Next.js in 2026: Faster Apps, Bigger Security Responsibilities]]></title>
            <link>https://medium.com/@prateekshawebdesign/next-js-in-2026-faster-apps-bigger-security-responsibilities-c02c1f33672c?source=rss-c4c4cda9f964------2</link>
            <guid isPermaLink="false">https://medium.com/p/c02c1f33672c</guid>
            <dc:creator><![CDATA[Prateeksha Web Design]]></dc:creator>
            <pubDate>Thu, 29 Oct 2026 09:23:00 GMT</pubDate>
            <atom:updated>2026-10-29T09:30:48.495Z</atom:updated>
            <content:encoded><![CDATA[<h4>What Next.js 16.4, React 19.3, critical October patches, and real-world hosting debates mean for teams choosing a modern React stack.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1eqN_KunE8ZIA0LiZRRfIQ.png" /></figure><p>October 2026 compressed the full cost of modern frontend development into three weeks. <a href="https://nextjs.org/blog/next-16-4">Next.js 16.4</a> shipped on October 6. Version 16.4.1 followed on October 14 with fixes for two critical remote code execution (RCE) vulnerabilities and a high-severity denial of service (DoS) issue. On October 27, the maintainers announced another scheduled security release for November 3. Next.js is getting faster, but the responsibility of operating it is accelerating too.</p><p>That isn’t a reason to avoid it, and frequent advisories don’t make a framework unsafe. They make patch ownership, monitoring and maintenance essential. <strong>Next.js and React development</strong> is not a one-time build; it carries a standing duty to patch, cache correctly, host sensibly and keep JavaScript lean. Our <a href="https://prateeksha.com/blog/website-security-maintenance-checklist-mumbai-2026">website security and maintenance checklist</a> covers the general version of that duty. This guide is the Next.js-specific one, for owners, product leads and technical decision-makers.</p><h3>What Next.js 16.4 and React 19.3 actually changed</h3><p>Four terms get thrown around without explanation, so here they are in plain language.</p><ul><li><strong>Server Components</strong> render on the server and send finished output to the browser. They ship no JavaScript of their own, which is why they’re the App Router default.</li><li><strong>Client Components</strong> run in the browser: menus, live-validated forms, anything with state or event handlers. Each one adds JavaScript the visitor must download and run.</li><li><strong>Cache Components</strong> are Next.js’s component-level caching model. You mark a function or component with &#39;use cache&#39;, which works like a component-level Cache-Control header and replaces the implicit caching of earlier App Router versions.</li><li><strong>Server Actions</strong> are server-side functions your UI can call directly, usually for forms and mutations. They’re convenient, but each one is an endpoint that accepts requests from the internet.</li></ul><p>Next.js 16.4 recommends Cache Components for every app and enables them by default for new create-next-app projects. They become the default in Next.js 17, so existing apps have a migration to plan rather than a surprise to absorb. The release also adds ensureStatic, which fails the build if dynamic content sneaks into a route meant to stay static. There are also smaller production bundles, an upgraded Turbopack Bundle Analyzer and next upgrade --agent. Its companion experimental.agentUpgrade option defaults to a &quot;security&quot; policy that reminds you about upgrades fixing known vulnerabilities in your installed version.</p><p>It ships with <a href="https://react.dev/blog/2026/09/09/react-19-3">React 19.3</a>, released September 9, 2026. View Transitions and Fragment Refs are stable, it adds Trusted Types support in React DOM, and Transitions now render independently so one slow update is less likely to hold up another.</p><h3>The security lesson: patch cadence is part of the architecture</h3><p>Next.js 16.4.1 (Active LTS) and 15.5.28 (Maintenance LTS), released <a href="https://nextjs.org/blog/nextjs-security-update-october-14-2026">October 14</a>, fixed three issues:</p><ul><li>A <strong>critical, unauthenticated RCE</strong> in App Router apps using Cache Components, triggered when a use cache function returns multiple typed arrays. There&#39;s no known workaround; the advice is to upgrade immediately.</li><li>A <strong>critical, unauthenticated RCE in the Image Optimization API</strong> when using SVG. An upstream library caused it, and it affects Next.js on glibc-based Linux. The workaround is sharp 0.35.5 or later, and upgrading Next.js pulls in a fixed version.</li><li>A <strong>high-severity DoS</strong> in App Router apps with at least one Server Action, where crafted requests cause excessive CPU use. Pages Router apps and apps without Server Actions aren’t vulnerable.</li></ul><p>The two critical fixes had been postponed from the <a href="https://nextjs.org/blog/september-2026-security-release">September 30 release</a> because of upstream dependency delays. That detail matters: your framework’s security timeline depends on libraries you never chose directly. September’s release had covered an image-optimization SSRF and cache-poisoning issues in self-hosted setups, among others.</p><p>On October 27 the team <a href="https://nextjs.org/blog/upcoming-nextjs-security-release-november-2026">announced</a> a scheduled release for Tuesday, November 3, 2026, covering seven medium-severity vulnerabilities. Details aren’t published yet, so the only useful step is to put the upgrade window in the calendar now.</p><p>Don’t read that list as proof of a liability. A project that publishes detailed advisories, patches multiple supported lines and gives advance notice is behaving like a mature one. The exposure was specific: one feature used one way, apps with Server Actions, one Linux libc. Teams that can answer “are we exposed?” in ten minutes know their features, hosting environment and image-processing setup.</p><p>So treat patch cadence as architecture. Decide who owns upgrades, which release line you track, how fast each severity ships, and how you roll back.</p><h3>Is Next.js still a sensible choice?</h3><p>Yes, for the right projects. Server rendering by default, fine control over what reaches the browser and a large React hiring pool are real strengths. It suits product-led sites, authenticated dashboards, ecommerce storefronts and anything where behaviour matters as much as content.</p><p>What’s changed isn’t the case for Next.js. It’s the cost of pretending it’s maintenance-free. Pick a supported LTS line, track advisories and budget engineering time every month, not just at launch. It’s a weaker choice when no one will own the stack, or when the site is mostly pages edited by non-technical staff.</p><h3>Next.js vs WordPress: choose based on the operating model</h3><p>“Which is better?” is the wrong question. We cover the detailed trade-offs in our comparison of <a href="https://prateeksha.com/blog/nextjs-vs-wordpress-which-is-better-for-business-websites-in-2026">Next.js and WordPress for business websites</a>. Each gives you a different operating model, and you should pick the one your organisation can actually run.</p><ul><li><strong>Editorial workflow.</strong> Next.js: Depends on the CMS you connect. WordPress: Mature and familiar out of the box.</li><li><strong>Custom application behaviour.</strong> Next.js: Strong: authenticated flows, dashboards, complex interactions. WordPress: Possible, but awkward as logic grows.</li><li><strong>Maintenance skills.</strong> Next.js: React, Node, build tooling, caching. WordPress: PHP, themes, plugin management.</li><li><strong>Security ownership.</strong> Next.js: Framework and npm patching, plus your hosting and image stack. WordPress: Core, theme and plugin updates. Plugin quality varies.</li><li><strong>Hosting complexity.</strong> Next.js: Managed platform to multi-instance Node with shared-cache concerns. WordPress: Typically simpler; caching still needs attention.</li><li><strong>Total cost.</strong> Next.js: Higher build effort and steady engineering time. WordPress: Often lower for content-led sites; costs shift to plugins and upkeep.</li></ul><p>Neither column wins universally. A team publishing weekly with no custom logic will probably be happier on WordPress. A product with accounts and integrations will probably outgrow it.</p><h3>The Next.js SEO checklist that matters</h3><p>Next.js doesn’t automatically guarantee SEO. It gives you good tools: server rendering, a metadata API and route-level control. You still have to use them correctly. This is the checklist we work through:</p><ol><li><strong>Rendering.</strong> Confirm every indexable page delivers its main content in the initial HTML. Use ensureStatic for routes that should stay static.</li><li><strong>Metadata.</strong> Give every page a unique title and description built from real page data. Our guide to <a href="https://prateeksha.com/blog/seo-nextjs-metadata-og-canonical">Next.js metadata, Open Graph and canonical tags</a> shows the patterns.</li><li><strong>Canonicals.</strong> Set a self-referencing canonical on each indexable page, and point parameterised URLs at the right version.</li><li><strong>Robots controls.</strong> Staging and preview deployments shouldn’t be indexable, and production pages shouldn’t inherit a stray noindex.</li><li><strong>Structured data.</strong> Add JSON-LD that matches visible content, and validate it.</li><li><strong>Sitemaps.</strong> Generate them from real routes or your CMS and list only canonical indexable URLs.</li><li><strong>Internal links.</strong> Use real anchors with real href values. Orphaned pages are common in script-heavy builds.</li><li><strong>JavaScript visibility.</strong> Test what a crawler sees with and without scripts.</li><li><strong>Crawlable error states.</strong> Return genuine 404 and 5xx status codes. A “not found” page that returns 200 is a soft 404.</li></ol><h3>Core Web Vitals and JavaScript budgets</h3><p>Per <a href="https://web.dev/articles/vitals">web.dev</a>, “good” means a Largest Contentful Paint (LCP) of 2.5 seconds or less, an Interaction to Next Paint (INP) of 200 milliseconds or less and a Cumulative Layout Shift (CLS) of 0.1 or less. They’re judged at the 75th percentile of page loads, split by mobile and desktop. A fast result on your office laptop says little about a phone on a congested connection.</p><p>JavaScript is the lever you control most directly. More script means more to download, parse and execute, which hurts INP in particular. A <strong>JavaScript budget</strong> is an agreed ceiling for what a route may ship, checked in your build or review process so growth gets noticed before users feel it. The right number depends on the page and audience, but you need one.</p><p>The 16.4 Bundle Analyzer helps, with route summaries and snapshot diffs between builds. On October 28, 2026, Vercel <a href="https://nextjs.org/blog/ship-less-javascript-with-ai">published its own case study</a>: its agents cut nearly two megabytes of compressed JavaScript across v0 and vercel.com using next-bundle-optimizer, including 150 KB on most v0 pages from one eagerly loaded code editor. That&#39;s Vercel&#39;s result on Vercel&#39;s codebases, not a promise for yours. The useful lesson is that a lazy import still loads upfront if any other module statically imports the same component. If your scores are already slipping, our <a href="https://prateeksha.com/blog/core-web-vitals-checklist-2026">Core Web Vitals checklist</a> covers the fixes in order.</p><h3>Key facts</h3><ul><li><strong>2.5 s / 200 ms / 0.1:</strong> the “good” thresholds for LCP, INP and CLS at the 75th percentile of page loads, per <a href="https://web.dev/articles/vitals">web.dev’s Core eb Vitals guide</a>.</li><li><strong>Nearly 2 MB of compressed JavaScript removed</strong> across v0 and vercel.com, according to Vercel’s own report on <a href="https://nextjs.org/blog/ship-less-javascript-with-ai">shipping less JavaScript with an AI agent</a> (October 28, 2026). It is one vendor’s case study, not a benchmark.</li><li><strong>7 medium-severity vulnerabilities</strong> are scheduled for the November 3, 2026 release, per the Next.js team’s <a href="https://nextjs.org/blog/upcoming-nextjs-security-release-november-2026">advance notice</a> of October 27, 2026.</li></ul><h3>Where React performance problems usually begin</h3><p>Slow Next.js apps tend to share a short list of causes, and it’s rarely the framework itself.</p><ul><li><strong>Excessive Client Components.</strong> A &#39;use client&#39; directive high in the tree pulls everything beneath it into the browser bundle. Push it down to the smallest interactive piece.</li><li><strong>Large shared dependency graphs.</strong> One heavy library imported from a shared file lands on every route. Date libraries, charting packages and rich-text editors are frequent offenders.</li><li><strong>Eager imports.</strong> Modals, editors and maps used only after a click shouldn’t load with the page. Check the lazy import is actually lazy.</li><li><strong>Unnecessary state and effects.</strong> Derivable state, effects that fetch what the server could have rendered, and re-renders from broad context changes add up on slower devices.</li><li><strong>Hydration mismatches.</strong> When server and client output differ, users can see flicker or broken interactivity. Our <a href="https://prateeksha.com/blog/hydration-error-nextjs-fix">hydration error guide</a> covers the common causes.</li><li><strong>Unoptimised images and third-party scripts.</strong> Oversized hero images and stacked tag, chat and analytics scripts can outweigh your application code. See our guide to <a href="https://prateeksha.com/blog/nextjs-image-optimization-responsive-ecommerce-portfolio">Next.js image optimisation for ecommerce and portfolios</a>.</li><li><strong>Incorrect caching or revalidation.</strong> Pages that should be cached rendering on every request, or stale content that never updates, trace back to caching decisions. Our explainer on <a href="https://prateeksha.com/blog/nextjs-caching-explained-route-cache-data-cache-revalidation-tags">Next.js caching, revalidation and tags</a> helps.</li></ul><h3>The real cost of Next.js hosting</h3><p>The monthly invoice is only part of hosting cost. Think in total cost of ownership across the options:</p><ul><li><strong>Managed platform.</strong> Less operational work, usage-based billing. As an illustration only, Vercel’s pricing page (checked October 29, 2026) lists Pro at $20 per month, with usage-based charges beyond included amounts for data transfer, function compute and image transformations. The point isn’t the plan; cost tracks traffic and compute, so it needs watching.</li><li><strong>Container or Node server.</strong> Many apps simply run next start on a Node.js server. You gain control, and scaling, restarts, deploys and TLS become your job.</li><li><strong>VPS self-hosting.</strong> Often the cheapest line item and the most expensive in attention, since OS, runtime and app patching all fall to you. Our <a href="https://prateeksha.com/blog/deploy-nextjs-on-aws-guide">guide to deploying Next.js on AWS</a> covers cost, scaling and observability for the cloud route.</li><li><strong>Static export, where appropriate.</strong> A genuinely static site loses the server runtime and a class of risk, but not features that need server behaviour.</li><li><strong>Everything around the server.</strong> CDN and cache costs, image transformation, logs, monitoring, backups and the engineering hours spent patching rarely appear in the first estimate and always appear in the second year.</li></ul><p>Two items deserve explicit planning. First, <strong>multi-instance cache consistency</strong>. Next.js 16.2 shipped a stable Adapter API, built with OpenNext, Netlify, Cloudflare, AWS Amplify and Google Cloud, plus a public test suite. As of that <a href="https://nextjs.org/blog/nextjs-across-platforms">March 2026 announcement</a>, Vercel and Bun adapters were shipping while Netlify, Cloudflare and AWS adapters were still in development. The practical point: with several instances, cached content, on-demand revalidation and streaming must stay correct across all of them. Test that before launch, not after a stale page reaches customers.</p><p>Second, <strong>image-optimisation memory</strong>. Image processing can be memory-hungry and sat behind two of the issues above, so size instances accordingly.</p><p>Not sure what hosting and upkeep a Next.js project would really cost you? Talk to our team about a setup you can maintain.</p><p><a href="https://prateeksha.com/nextjs">Plan your Next.js setup</a></p><h3>A practical operating checklist for production teams</h3><p>Whether you run it in-house or with an agency, this is the baseline we’d expect from any production Next.js app:</p><ol><li><strong>Name a dependency owner.</strong> One person, with a backup, accountable for framework and package updates. “The team” isn’t an owner.</li><li><strong>Subscribe to advisories.</strong> Follow the Next.js blog and GitHub security advisories via RSS or notifications, routed to a channel someone reads. Leave the agentUpgrade security reminders on.</li><li><strong>Stay on a supported LTS release.</strong> The October fixes landed on both 16.4.1 (Active LTS) and 15.5.28 (Maintenance LTS). Know which line you’re on and when it stops receiving fixes.</li><li><strong>Keep staging and a rehearsed rollback.</strong> Every upgrade passes through staging, and you’ve practised reverting.</li><li><strong>Write down a secret rotation procedure.</strong> After a suspected exposure you should know where each key lives and how to rotate it.</li><li><strong>Set a patch SLA by severity.</strong> Define in advance how fast critical, high and medium fixes must ship. The numbers are your call; the discipline is what counts.</li><li><strong>Keep a dependency and image-processing inventory.</strong> Which features you use (App Router, Cache Components, Server Actions, SVG handling), your Linux base image and your sharp version. It turns an advisory into a ten-minute answer.</li><li><strong>Log and rehearse incident response.</strong> Keep logs long enough to investigate, alert on CPU spikes and error rates, and agree who does what when something looks wrong.</li></ol><h3>When Prateeksha recommends Next.js, and when we don’t</h3><p>We build with Next.js, so we’ll be direct about where we’d use it and where we wouldn’t.</p><p><strong>We recommend it</strong> when the product has real application behaviour: logged-in experiences, dashboards, complex search, integrations or ecommerce flows with custom logic. We also recommend it when front-end performance matters to the business and someone, us or you, is committed to maintaining it properly.</p><p><strong>We’d steer you elsewhere</strong> when the site is mostly content with simple needs and the editors already live in a WordPress workflow. There, WordPress is often the better business decision: cheaper to build, familiar to the team and easier to hand over. We’d also hesitate if nobody will own updates after launch. An unmaintained Next.js app is a bigger risk than an unmaintained brochure site on a managed host.</p><p>Already on Next.js and unsure about its health? Start with your version, enabled features, bundle sizes and Core Web Vitals field data.</p><p>Already on Next.js and worried about slow pages or heavy JavaScript? We can review your bundles and Core Web Vitals and show you where the weight is.</p><p><a href="https://prateeksha.com/speed-optimization">Get a speed review</a></p><h3>Reference reading</h3><ul><li><a href="https://nextjs.org/blog/next-16-4">Next.js 16.4 release notes: Cache Components, smaller bundles and upgrade tooling</a></li><li><a href="https://nextjs.org/blog/nextjs-security-update-october-14-2026">Next.js security update, October 14, 2026: affected configurations and fixed versions</a></li><li><a href="https://nextjs.org/blog/upcoming-nextjs-security-release-november-2026">Advance notice of the November 2026 Next.js security release</a></li><li><a href="https://web.dev/articles/vitals">Core Web Vitals explained: metrics and thresholds on web.dev</a></li></ul><h3>Frequently Asked Questions</h3><h3>Is Next.js secure enough for a business website?</h3><p>Yes, when it’s maintained. Every widely used framework publishes vulnerabilities, and Next.js publishes detailed advisories with fixes across supported release lines. The risk comes from running an outdated version, so someone has to be responsible for applying patches promptly.</p><h3>Did the October 2026 vulnerabilities affect every Next.js app?</h3><p>No. The Cache Components RCE applied only to App Router apps whose use cache function returns multiple typed arrays. The DoS required at least one Server Action, and the SVG image-optimisation RCE affected Next.js on glibc-based Linux. Checking your features and environment against the advisory tells you whether you were exposed.</p><h3>Which Next.js version should I be running right now?</h3><p>Run a supported LTS release that includes the October 14 fixes: 16.4.1 on the Active LTS line or 15.5.28 on the Maintenance LTS line. A scheduled release on November 3, 2026 will address seven medium-severity issues. Its details weren’t published when this article was written, so check the official announcement before you upgrade.</p><h3>Does Next.js guarantee better SEO than WordPress?</h3><p>No. Next.js gives you strong tools for rendering, metadata and performance, but nothing is automatic. Canonicals, robots controls, structured data, sitemaps and crawlable error states still have to be implemented correctly, and a well-configured WordPress site can rank just as well.</p><h3>What are the thresholds for good Core Web Vitals?</h3><p>According to web.dev, LCP is good at 2.5 seconds or less, INP at 200 milliseconds or less and CLS at 0.1 or less. These are judged at the 75th percentile of page loads and reported separately for mobile and desktop. Field data from real visitors is what counts, not a single test on your own device.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c02c1f33672c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Web Design and SEO Services in Thane: What Local Businesses Need to Know]]></title>
            <link>https://medium.com/@prateekshawebdesign/web-design-and-seo-services-in-thane-what-local-businesses-need-to-know-37e0559fa450?source=rss-c4c4cda9f964------2</link>
            <guid isPermaLink="false">https://medium.com/p/37e0559fa450</guid>
            <category><![CDATA[web-design]]></category>
            <dc:creator><![CDATA[Prateeksha Web Design]]></dc:creator>
            <pubDate>Mon, 12 Oct 2026 21:43:57 GMT</pubDate>
            <atom:updated>2026-10-12T21:43:57.486Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="Thane street with a glowing map pin beside a local storefront and apartment buildings" src="https://cdn-images-1.medium.com/max/1024/1*tkreMw5AB-6v92A7Y2taDg.jpeg" /></figure><p>Thane gets treated as a Mumbai suburb in most marketing strategies, but it has its own commercial identity entirely — Ghodbunder Road, Naupada, Vartak Nagar, Manpada — each with its own traffic patterns that shape how people actually search for local services.</p><p>Thane’s buyer base has grown fast enough to support serious competition, while staying local enough that word of mouth and “near me” searches still decide who gets the enquiry over a competitor with a bigger ad budget.</p><p>We wrote a guide on what web design and local SEO actually need to look like for businesses competing in Thane specifically — not a copy-pasted Mumbai strategy applied to a different pin code.</p><p>Read the full guide on Prateeksha Web Design: <a href="https://prateeksha.com/blog/web-design-seo-thane">https://prateeksha.com/blog/web-design-seo-thane</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=37e0559fa450" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Website Maintenance Plans in India 2026: What Should You Expect and What Does It Cost?]]></title>
            <link>https://medium.com/@prateekshawebdesign/website-maintenance-plans-in-india-2026-what-should-you-expect-and-what-does-it-cost-d471a678245a?source=rss-c4c4cda9f964------2</link>
            <guid isPermaLink="false">https://medium.com/p/d471a678245a</guid>
            <dc:creator><![CDATA[Prateeksha Web Design]]></dc:creator>
            <pubDate>Thu, 24 Sep 2026 13:16:16 GMT</pubDate>
            <atom:updated>2026-09-24T13:16:16.719Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2pMfzcbFdeyvzVO-3wm1bw.jpeg" /></figure><p>Website maintenance plan cost in India typically runs ₹3,000 to ₹20,000 a month. That range alone tells you almost nothing, because the number on the invoice doesn’t reveal what’s actually included.</p><p>Two agencies can quote the exact same ₹5,000 a month for completely different scopes of work. One plan covers backups and nothing else. The other covers backups, security patching, uptime monitoring, and a monthly performance report.</p><p>We wrote a practical breakdown of what a real website maintenance plan should include, realistic pricing for different levels of coverage, and how to evaluate a maintenance quote before signing — so the invoice actually matches the work being done.</p><p>Read the full guide<br><a href="https://prateeksha.com/blog/website-maintenance-plans-india-2026">https://prateeksha.com/blog/website-maintenance-plans-india-2026</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d471a678245a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What Is AEO (Answer Engine Optimisation)? Why Mumbai Businesses Need to Care Abo]]></title>
            <link>https://medium.com/@prateekshawebdesign/what-is-aeo-answer-engine-optimisation-why-mumbai-businesses-need-to-care-abo-0f936ff79869?source=rss-c4c4cda9f964------2</link>
            <guid isPermaLink="false">https://medium.com/p/0f936ff79869</guid>
            <dc:creator><![CDATA[Prateeksha Web Design]]></dc:creator>
            <pubDate>Mon, 21 Sep 2026 08:41:25 GMT</pubDate>
            <atom:updated>2026-09-21T08:41:25.978Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*E_bMpUeFx_t8JJ_W_Igf4Q.jpeg" /></figure><p>Ask ChatGPT a question today and you’ll often get a direct answer with a specific business name attached — no blue links, no scrolling, no ten tabs open comparing options side by side.</p><p>That’s Answer Engine Optimisation (AEO): structuring your business information so AI systems like ChatGPT, Google AI Overviews, Gemini, and Perplexity can find it, trust it, and recommend it by name, rather than defaulting to a generic or better-known competitor.</p><p>For Mumbai businesses, this is becoming as consequential as traditional search rankings — but almost nobody has structured their information for it yet. We wrote a guide explaining what AEO actually is, why some businesses get cited by AI systems and others get ignored, and the concrete steps to fix that.</p><p>Read the full guide<br><a href="https://prateeksha.com/blog/what-is-aeo-answer-engine-optimisation">https://prateeksha.com/blog/what-is-aeo-answer-engine-optimisation</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0f936ff79869" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>