<?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 Aakash Mangla on Medium]]></title>
        <description><![CDATA[Stories by Aakash Mangla on Medium]]></description>
        <link>https://medium.com/@aakashmangla3?source=rss-7d07c2c1113d------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*4Qf5ILF4f63L_si0rXSqig.jpeg</url>
            <title>Stories by Aakash Mangla on Medium</title>
            <link>https://medium.com/@aakashmangla3?source=rss-7d07c2c1113d------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 00:54:19 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@aakashmangla3/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[Design thinking for people who sell]]></title>
            <link>https://medium.com/@aakashmangla3/design-thinking-for-people-who-sell-1f8eafa0502a?source=rss-7d07c2c1113d------2</link>
            <guid isPermaLink="false">https://medium.com/p/1f8eafa0502a</guid>
            <category><![CDATA[business]]></category>
            <category><![CDATA[sales]]></category>
            <category><![CDATA[design]]></category>
            <category><![CDATA[design-thinking]]></category>
            <dc:creator><![CDATA[Aakash Mangla]]></dc:creator>
            <pubDate>Thu, 08 Oct 2026 14:20:11 GMT</pubDate>
            <atom:updated>2026-10-08T14:20:11.894Z</atom:updated>
            <content:encoded><![CDATA[<p><em>This piece is the essence of a session I ran for the business team at </em><a href="https://vecmocon.com/"><em>Vecmocon</em></a><em> in September 2026. The slides were built for a room. This version is meant to be read, and come back to.</em></p><p>A customer emails on a Tuesday afternoon. Their team needs an export-to-Excel button on the dashboard, and they would like it before month end. The request is clear, the change looks small, and the reply almost writes itself: yes, here is a timeline, here is the cost.</p><p>That reply answers the question they asked and skips the one underneath it. Somebody on their side has already decided that a spreadsheet is the answer to something, and nobody has written down what the something is. Maybe the operations head wants a weekly number for her Monday review. Maybe finance needs to reconcile usage against invoices. Maybe someone is copying data into another tool because the two systems don’t talk to each other. Each of those is a different problem, and only one of them is solved well by a button.</p><p>Design thinking is a way of working that starts with the person who has the problem and works back towards what you might offer. Designers use it to decide what to build. The same habits hold up on a sales call, an RFP, or a deal that has gone quiet, because each of those is a moment where someone hands you their guess at a solution and waits to see whether you accept it.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*OWn1fyYkxRM-hBkLsxD9GA.png" /></figure><p>It is usually drawn as five phases: understand, define, ideate, prototype, test. In practice you loop back constantly. Something you learn while testing sends you back to understanding, and a sharper problem statement changes which ideas are worth having. Most of it comes down to slowing down on your first instinct, because the first instinct usually fits what you already sell.</p><h3>Understand: what they did last time</h3><p>On a first call you can collect two kinds of information: what people say they need, and what they actually did the last time the problem came up. The second is more useful. Stated needs are shaped by what the customer thinks you can offer and by how they want to sound, while behaviour is harder to dress up.</p><p>So ask about last time. “Walk me through the last time you needed this data” gets you a story with a date, a person, a deadline and a workaround. “Would an export button help?” gets a polite yes, because the person is now guessing about their own future, the same way you are.</p><p>Keep your product out of your questions. If your feature name or part number appears in the question, you are testing your pitch and learning nothing about them. Ask why the request matters, and when you get an answer, ask why that matters. Within a few rounds you usually move from a spec to a business reason: a review meeting, a cost someone is held to, a complaint that keeps coming back.</p><p>Then leave room. When someone finishes an answer, wait about three seconds before you speak, because the second thing people say is often the honest one. I picked up a version of this from research interviewing and use it on internal discovery calls as well. Instead of interrupting with questions, I write them in a notebook as they occur to me. On a recent call about a new module, nearly every question on that page got answered as my colleague kept talking.</p><p>Finish by saying the problem back in one sentence and letting them correct you. Corrections are free information. If you can’t say their problem back in a sentence, the call isn’t finished.</p><h3>Define: the request is already an answer</h3><p>A request for quote, or a request like the export button, is the customer’s best guess at a solution, written before they spoke to you. Take it at face value and you compete on the terms of that guess, against everyone else who received the same document.</p><p>The problem underneath is a pain. In the export example it might be that the operations head rebuilds the same table by hand every Monday morning before her review. That can be solved several ways, and some of them are better than a button.</p><p>One format helps when writing the problem down:</p><blockquote><em>[Customer] needs a way to [job] because [insight], but today [obstacle].</em></blockquote><p>The insight slot is where the work is. “The operations team needs a way to export data because they want Excel” fills the template and says nothing. “The operations head needs a way to see weekly utilisation before Monday’s review because her targets are set on it, but today she rebuilds it by hand from several screens” gives you something to design against. It is also something a competitor quoting on the button won’t have.</p><p>Sometimes defining the problem comes down to a single question. In a recent meeting with production, sourcing, process and ERP teams, we were stuck on whether consumable parts should be created in two of our systems or one. Creating them in both meant constant syncing and messy code to manage it. After setting context, I asked one yes-or-no question: will a consumable ever appear in a bill of materials? The answer was no, and that settled it. Consumables moved entirely into one system.</p><p>A problem statement only counts if the customer recognises it, and agreement in a meeting can be politeness. I have had decisions confirmed across two meetings and an email, then reopened weeks later by someone who had nodded through all of it. Looking back, those meetings were me explaining while others listened, and the email recorded my version of the problem rather than one we had built together. Saying it back works when the other person has room to correct you.</p><h3>Ideate: build the options with them</h3><p>Once the problem is clear, the instinct is to go away and come back with a quote. A single quote asks for a verdict, and yes or no is the only move it leaves the customer. Two or three options ask which, and the answer tells you what they value: speed, cost, control, or fit with something you didn’t know about.</p><p>Bring those options into the conversation and work on them with the customer. For the operations head, that might be a scheduled weekly report in her inbox, a saved dashboard view built for her review, or the export button with a template already shaped like her table. Put them in front of her on a call. She will pick one, combine two, or add one you hadn’t thought of.</p><p>That matters after the call ends. Someone who helped draw the plan will defend it inside their company when you are not in the room. The first idea anyone has is usually the obvious one, which is also the one your competitor brought, so range is where the difference between you shows up.</p><h3>Prototype and test: the cheapest way to be wrong</h3><p>Every deal runs on guesses. When a deal goes quiet, you form a belief about why: the budget froze, your champion moved teams, a competitor got in. The instinct is to build the next three weeks of work on that belief. You are already wrong about something, and the only thing that varies is how much you build on it before you find out.</p><p>So look for the cheapest way to be wrong, this week rather than this quarter. Write down the sentence “I am assuming that…” and finish it. If that assumption turns out false and your plan doesn’t survive, that is the one to test. Shrink the test until it fits in a week, because anything that needs engineering time isn’t a prototype. Before you send anything, decide what counts as a yes, a no and a maybe. Otherwise silence gets read as whatever you hoped for.</p><p>In sales, a prototype is rarely a product. It can be any of these:</p><ul><li>a one-paragraph email stating the problem as you understand it</li><li>a single slide with the reframed problem and three options</li><li>a calendar invite they are free to decline</li><li>a ten-unit pilot described in one line</li><li>a sketch of the deployment, drawn live on a call</li></ul><p>The useful ones invite correction. A test the customer can’t fail tells you nothing.</p><h3>Further reading</h3><p>The Mom Test by Rob Fitzpatrick covers how to ask questions that people can’t answer politely. To Sell Is Human by Daniel Pink makes the case that most work now involves moving people. Salesforce’s Trailhead also has a short <a href="https://trailhead.salesforce.com/content/learn/modules/design-thinking-for-sales/apply-design-thinking-tosales">Design Thinking for Sales module</a>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1f8eafa0502a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How I Rebuilt Our Design System’s Color Palette with OKLCH — and What I Learned the Hard Way]]></title>
            <link>https://medium.com/design-bootcamp/how-i-rebuilt-our-design-systems-color-palette-with-oklch-and-what-i-learned-the-hard-way-018ff626cb4c?source=rss-7d07c2c1113d------2</link>
            <guid isPermaLink="false">https://medium.com/p/018ff626cb4c</guid>
            <category><![CDATA[ux]]></category>
            <category><![CDATA[ui]]></category>
            <category><![CDATA[design-systems]]></category>
            <category><![CDATA[ux-design]]></category>
            <category><![CDATA[ui-design]]></category>
            <dc:creator><![CDATA[Aakash Mangla]]></dc:creator>
            <pubDate>Fri, 12 Jun 2026 11:11:16 GMT</pubDate>
            <atom:updated>2026-06-12T11:11:16.494Z</atom:updated>
            <content:encoded><![CDATA[<p>There’s a specific kind of frustration that comes from staring at your own design system and knowing something is <em>wrong</em> — but not being able to articulate exactly what.</p><p>That was me, working on a design system that had been around for two and a half years. The palette had been assembled over time by multiple hands, with good intentions but no underlying model. Some colors felt vivid and punchy. Others felt washed out next to them. Badge colors that looked fine in isolation became indistinguishable in practice. Data visualization hues that were supposed to be clearly distinct would merge at smaller sizes.</p><blockquote>The palette wasn’t ugly. It just wasn’t <em>coherent</em>. And the deeper I dug, the more I realized the problem wasn’t taste, it was the color model we’d been working in.</blockquote><p>This is the full story of how I fixed it: what OKLCH is, why it was the right model, the step-by-step process I followed, the tools I tried (including one that hit a wall), and — most importantly — the real problems I ran into that no tutorial warned me about.</p><h3>1. Why Legacy Palettes Break</h3><p>Most design tools — and most designers’ mental models — operate in HSL or HSB. These models are intuitive: Hue is the color, Saturation is how vivid it is, Lightness or Brightness is how light or dark it is.</p><p>The problem is that HSL and HSB are lies.</p><p>Not intentionally — they’re just mathematical transformations of RGB, and RGB was designed for screens, not for human perception. The result is that two colors with the same HSL “Lightness” value can look dramatically different in actual perceived brightness. A orange at HSL(30, 100%, 50%) and a blue at HSL(250, 100%, 50%) are technically identical in lightness. But to your eye, the yellow blazes and the blue recedes. They are not perceptually equal at all.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*PiqkWpJ6IuYs-rsxyBBBOQ.png" /><figcaption>In HSL model, colors with technically identical lightness are not perceptually equal (reference: oklch.fyi).</figcaption></figure><p>This is the root cause of most “something feels off” palette problems. When you build a color ramp in HSL, you’re distributing colors evenly through a mathematical space that has no relationship to how your eyes work. You end up with:</p><ul><li>Shade 500s across different color families that feel like completely different weight classes</li><li>Ramps where the jump from 400 to 500 feels massive for one hue and invisible for another</li><li>Dark shades that converge toward indistinguishable mud because the hue’s identity collapses at low lightness in RGB space</li><li>Light shades that feel “loud” for some hues and “washed out” for others at the same nominal saturation</li></ul><p>The fix is to use a color model that was designed around human perception — one where equal numbers actually mean equal perceived values.</p><p>That model is OKLCH.</p><h3>2. What is OKLCH — and Why It’s the Right Model</h3><p>OKLCH is a polar representation of the Oklab color space, which was specifically designed so that equal numerical distances correspond to equal perceptual differences. It describes every color with three values:</p><ul><li><strong>L — Lightness:</strong> 0 is pure black, 1 is pure white. A blue and a yellow at L 0.62 will genuinely <em>feel</em> like the same lightness to the human eye.</li><li><strong>C — Chroma:</strong> How colorful or vivid the color is. 0 is completely grey. Higher values are more saturated. Unlike HSL’s “Saturation”, chroma is perceptually linear.</li><li><strong>H — Hue:</strong> The angle on the color wheel, 0°–360°. Red is around 25–30°, yellow around 90°, green around 145°, blue around 230–265°.</li></ul><p>The key insight that makes OKLCH transformative for palette work: <strong>if you hold L constant across all your color families, every shade at that step will genuinely feel like the same visual weight.</strong> Your blue-500 and your red-500 and your green-500 will be perceptual siblings, not strangers who happen to share a name.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*vIAlWE1sC6nFKDGvfSx6Hw.png" /><figcaption>OKLCH offers perceptually equal colors for identical lightness (reference: oklch.fyi).</figcaption></figure><p>This is what HSL promises but cannot deliver.</p><p>OKLCH also has a direct line to the browser — it’s now a native CSS color function (oklch(0.62 0.18 240)), which means the palette you build maps directly to modern CSS without any conversion layer.</p><p>You can learn more about it at <a href="https://oklch.fyi/">oklch.fyi</a>.</p><h3>Why not Lab, LCH, or HSLuv?</h3><p>These are all better than HSL, and they’re all built on similar perceptual principles. OKLCH specifically is worth using because:</p><ol><li><strong>Oklab is more accurate than the older CIELAB</strong> for blue hues in particular — classic LCH has well-known hue shift issues around blues and cyans that Oklab corrects.</li><li><strong>Native CSS support</strong> as of 2023 across all modern browsers.</li><li><strong>Tooling has caught up</strong> — Figma, several plugins, and web tools now work natively in OKLCH.</li></ol><h3>3. The Guiding Principle: Evolutionary, Not Revolutionary</h3><p>Before I touch a single color value, I want to name the constraint that shaped every decision in this project.</p><p>We could not blow up our existing color identity.</p><p>The brand had been using the same primary blue for years. The green was established in user-facing success states. The red was in error messages across every product surface. Even if I could theoretically argue that a different hue would be “more correct,” changing it would have meant re-approving assets, updating marketing materials, and — most importantly — breaking the visual continuity that users had learned to associate with meaning.</p><p>The constraint was: <strong>the new palette must feel evolutionary, not revolutionary.</strong> Users and stakeholders should not notice the change. They should just notice, over time, that things feel more coherent and polished.</p><p>This is actually one of the best features of OKLCH. Because OKLCH separates hue from lightness and chroma, I could lock the hues close to their existing values and rebuild only the structural layer underneath. The identity stays. The inconsistency gets fixed.</p><p>Concretely: if your existing blue-500 sits at H 240°, your new blue ramp uses H 240° throughout. You’re not choosing a new blue. You’re choosing a new <em>structure</em> for the same blue.</p><h3>4. Step-by-Step: Building the Palette</h3><h3>Step 1: Audit your current palette in OKLCH</h3><p>Before building anything, convert every color in your existing palette to OKLCH. Tools like <a href="https://oklch.com/">oklch.com</a> let you paste a hex value and read back its L, C, and H. Do this for your entire palette and put the numbers in a spreadsheet.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*8jjT2P5G1Fcicv2LilAHUw.png" /></figure><p>What you’re looking for:</p><ul><li><strong>L values at the same shade step across families.</strong> Are all your 500s at a similar L? Or is red-500 at L 0.55 while yellow-500 is at L 0.72? That discrepancy is why they don’t feel like siblings.</li><li><strong>Chroma distribution within ramps.</strong> Does your blue ramp have a smooth arc of chroma values from light to dark, peaking somewhere in the middle? Or is it all over the place?</li><li><strong>Hue drift within a single ramp.</strong> Does your blue-100 and blue-900 share the same H value? In HSL-built ramps, the hue often drifts several degrees across the ramp, which makes the ramp feel muddy.</li></ul><p>This audit is the foundation. It tells you what you’re actually fixing, and it gives you anchors to work from rather than starting from scratch.</p><h3>Step 2: Lock your lightness scale</h3><p>This is the single most important decision in the entire process, and it should be made before you touch chroma or hue for any individual color.</p><p>Define one shared L value for each shade step (50 through 950). Every color family will use these exact values. Here’s a reasonable starting scale:</p><pre>50:  L 0.98<br>100: L 0.95<br>200: L 0.89<br>300: L 0.81<br>400: L 0.71<br>500: L 0.62<br>600: L 0.54<br>700: L 0.46<br>800: L 0.38<br>900: L 0.28<br>950: L 0.18</pre><p>Tune these against your contrast requirements — if you need shade 700 to pass a specific contrast ratio against white for body text, adjust L 0.46 accordingly. But once set, they are fixed across all families. This shared lightness scale is what makes blue-600 and red-600 feel like the same visual weight.</p><h3>Step 3: Anchor hues from your existing palette</h3><p>For each color family, take the “main” shade from your existing palette — typically your 500 or 600 — and read its H value from the OKLCH audit you did in Step 1. That becomes your anchor hue for the new ramp.</p><p>Round to a clean number if you want (e.g., 237° → 240°), but stay within ~10° of the original to preserve identity. Keep the hue constant across all steps in a ramp — no drift.</p><p>This is where the “evolutionary” constraint pays off mechanically. You’re not guessing new hues. You’re reading them from what already exists.</p><h3>Step 4: Design your chroma curves</h3><p>Chroma is where the real craft lives. For each family, you need to decide how chroma behaves across the 11 shade steps.</p><p>The general shape that works well is a <strong>low → high → low arc</strong> — low chroma at the light end (50, 100), rising to a peak somewhere around shade 400–600, then falling again at the dark end (800, 900, 950). This mirrors how colors naturally behave: very light tints and very dark shades both approach grey.</p><p>The peak chroma value is constrained by the sRGB gamut — not every OKLCH color is representable on screen. High chroma at certain hues will fall outside the triangle of colors your monitor can display. Your tooling needs to flag this (more on that in the Tools section).</p><p>The chroma values also don’t need to be identical across families. Yellow and green, for reasons covered in the “Problems We Hit” section, need lower peak chroma than red or blue to feel balanced. The goal is not mathematical equality — it’s perceptual harmony.</p><h3>Step 5: Validate and export</h3><p>Once your ramps are defined, validate two things before exporting:</p><ol><li><strong>sRGB gamut compliance.</strong> Every shade should be representable in sRGB, or you need to reduce its chroma.</li><li><strong>Contrast ratios.</strong> Run APCA checks on your planned text color / background color pairs (see Section 8 on APCA vs. WCAG).</li></ol><p>Then export as JSON or as hex values, ready to import into your design tool.</p><h3>5. Tools of the Trade</h3><p>I tried several tools during this process. Here’s an honest account of each.</p><h3>Harmonizer (Website)</h3><p>This was my first stop. <a href="https://harmonizer.evilmartians.com/">Harmonizer</a> is a web-based tool for building OKLCH-based palettes, which seemed ideal.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*GSx6qPmevRGZqyarHcvjiQ.png" /><figcaption>Harmonizer lets you control APCA contrast values and set chroma cap for each shade.</figcaption></figure><p>I hit a significant limitation: <strong>when you load multiple hue families simultaneously, Harmonizer clamps all families to the weakest hue’s chroma ceiling.</strong> If one of your hues has a lower maximum displayable chroma in sRGB (say, yellow), Harmonizer uses that ceiling for everything — including your red and blue, which could go higher. The result is a palette that’s more muted than it needs to be, because you’re penalizing every color family for your most constrained one.</p><p>The other limitation: Harmonizer only exposes the <em>peak</em> chroma value, not the per-shade chroma. You can’t precisely sculpt the curve shape for each ramp. And the contrast input only accepts APCA values rather than direct lightness values, which makes it harder to set up your lightness scale with precision.</p><p>Harmonizer isn’t bad — for a single-hue palette or a quick exploration, it works fine. But for a multi-family design system palette, it’s not the right tool.</p><h3>Huetone</h3><p><a href="https://huetone.ardov.me/">Huetone</a> is a free web tool, and it became the workhorse of this project.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_eaBE8rI_u1Po4VSmcae3g.png" /><figcaption>Huetone provides perfect amount of control and visualization</figcaption></figure><p>What makes it right for this job:</p><ul><li><strong>You edit L and C curves directly per family.</strong> You can see and manipulate the exact lightness and chroma value for every shade step.</li><li><strong>APCA contrast validation is built in.</strong> You can see contrast ratios update in real time as you adjust curves.</li><li><strong>Multiple hue families live side by side</strong> without the cross-family chroma clamping issue Harmonizer has.</li><li><strong>It exports to JSON</strong>, which becomes the input for the next step.</li></ul><p>The workflow in Huetone: create a hue family, set the hue angle, then manually tune the L curve to match your locked lightness scale, and tune the C curve to hit your target chroma arc. Repeat for each family. The side-by-side view lets you compare families and catch visual imbalances before they become token-level problems.</p><p>One thing to be aware of: Huetone doesn’t auto-clamp out-of-gamut colors. It will flag them (shades that fall outside sRGB turn red), after which you can make them displayable using an option.</p><h3>Token Studio (Figma Plugin)</h3><p>Once Huetone exports your palette as JSON, you need to get those values into Figma as variables. Token Studio is the bridge.</p><p>Token Studio is a Figma plugin that lets you import a JSON token file and generate Figma variables from it automatically. Instead of manually creating 11 × N color variables by hand (which would take hours and introduce transcription errors), Token Studio reads the JSON and creates the entire primitive variable set in one operation.</p><p>The workflow:</p><ol><li>Export palette from Huetone as JSON</li><li>Open Token Studio in Figma</li><li>Import the JSON as a token set</li><li>Sync to Figma variables</li></ol><p>The JSON that Huetone exports maps cleanly to what Token Studio expects for a flat color token structure. You may need to do a light clean-up pass on the JSON keys (Token Studio uses a / or . separator for naming hierarchy, e.g., blue/500), but it&#39;s minimal work.</p><p>This step took what could have been an afternoon of manual variable creation and reduced it to about fifteen minutes.</p><h3>6. Real Problems I Hit (and How I Solved Them)</h3><p>This is the section I wish had existed before I started. Every tutorial I read made the process sound clean and linear. The reality was messier — and more instructive.</p><h3>Problem 1: Brand Teal and Green were 10° apart and their dark shades merged</h3><p>In my initial audit, brand teal was sitting at H 165° and green at H 155°. That’s only a 10° separation on a 360° wheel. At light shades (50–300), the two families looked distinct because lightness and chroma differences carried the differentiation. But at dark shades (700–950), both colors converged.</p><p>The fix: I pushed green toward H 145°, which is closer to leafy green. At H 145°, red reads unmistakably as red at both light and dark shades, and the separation from brand became a comfortable 20°. The identity change was minimal — it still read as “our red” — but the dark-end collision disappeared.</p><p><strong>The lesson:</strong> Hue separation that looks adequate at medium shades can collapse completely at the dark end of your ramp. Always check your dark shades (700, 800, 900) side by side across all your feedback colors before finalising hue angles.</p><h3>Problem 2: Yellow looked like mustard at light shades</h3><p>Yellow at high chroma in light shades (50–300) doesn’t look like “light yellow.” It looks like mustard or old parchment. This is a quirk of how yellow sits in the color space — the hue around 90° has a relatively narrow band where it reads as “yellow” before tipping into green-yellow on one side or orange-yellow on the other, and at high chroma in light shades it tips into something muddy.</p><p>The fix: I applied a chroma reduction specifically at shades 50 through 300 for yellow. The light shades now have noticeably lower C values than the rest of the yellow ramp, which keeps them reading as pale yellow rather than mustard. The curve shape for yellow is asymmetric — it climbs more steeply from the low end than it descends at the dark end, because the problem lives specifically in the light shades.</p><p><strong>The lesson:</strong> Chroma curves are not required to be symmetric or consistent across families. Each hue has its own perceptual quirks, and your curves should accommodate them.</p><h3>Problem 3: Green felt visually heavier than its siblings</h3><p>At identical L and C values, greens feel more intense than reds, blues, or oranges. This is not a tooling artifact — it’s human biology. The human eye has more green-sensitive photoreceptors than red or blue, and we are more perceptually sensitive to variations in the green range of the spectrum. A green with C 0.15 and a red with C 0.15 are mathematically equal in chroma, but the green will feel louder.</p><p>The fix: I reduced peak chroma for green relative to the other feedback colors. Green peaked at C 0.13 where red and blue peaked at C 0.17–0.18. This brought the visual weight into balance without making green feel washed out.</p><p><strong>The lesson:</strong> Perceptual equality in OKLCH is much better than HSL, but it is not perfect. Green especially needs manual tuning downward. Trust your eyes at the end of the process, not just the numbers.</p><h3>Problem 4: Structural perfection produced a visually muted palette</h3><p>This was the most important lesson, and the hardest one to accept.</p><p>Halfway through the project, I had a palette that was structurally impeccable. Every family shared the same lightness scale. Chroma peaked at shade 500 across all families. Everything was symmetric and internally consistent. I was proud of it.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/942/1*YsESQX4Fji942EL-qR0mkA.png" /><figcaption>Technically the perfect palette but flat and muted</figcaption></figure><p>Then I put it in components and ran it past my team, and the feedback was clear: it looked flat. The feedback colors (red, orange, yellow, green, blue) — the ones that need to carry urgency and identity in badges, alerts, and status indicators — didn’t pop. At dark shades, everything looked like a variation on the same muted tone.</p><p>The structure was right. The result was wrong.</p><h3>7. The Two-Palette Compromise</h3><p>The resolution to the “structurally correct but visually muted” problem was an architectural decision that changed how I thought about the whole project.</p><p>I split the palette into two categories with different rules:</p><p><strong>Brand and neutral colors</strong> — these use the carefully tuned, structurally uniform palette. Consistent lightness scale, modest chroma, symmetric curves, chroma peaks disciplined to stay within a range that works across the whole group. These colors are the foundation of the UI: backgrounds, surfaces, borders, text. Consistency matters more than vividness here.</p><p><strong>Feedback and data-viz colors</strong> (red, orange, yellow, green, blue) — these use high chroma values, pushed as far toward the sRGB gamut ceiling as each hue allows. The chroma peak may drift from shade 500 depending on each hue’s individual gamut ceiling. The curves are tuned per-color based on visual output rather than mathematical consistency.</p><p>The key insight: <strong>feedback colors are not supposed to feel like comfortable siblings. They’re supposed to be instantly, unmistakably identifiable.</strong> Red needs to scream. Yellow needs to pop. Green needs to read as “good” at a glance. Perceptual uniformity works against this job.</p><p>This wasn’t a failure of the system. It was the system being used correctly — OKLCH gave me the building blocks and the vocabulary to make a deliberate architectural decision, rather than having the color model make that decision for me implicitly (and badly).</p><p>The practical result:</p><ul><li>Brand ramp: maximum chroma around C 0.12–0.14, tightly controlled</li><li>Neutral ramp: maximum chroma around C 0.02, barely tinted</li><li>Feedback ramps: maximum chroma between C 0.17–0.22 depending on hue, gamut-clamped, curves tuned individually</li></ul><h3>8. Contrast: APCA vs. WCAG</h3><p>While going through this process, I spent enough time in contrast validation tooling that it’s worth addressing directly: if you’re building a palette in 2025 and beyond, you should understand what APCA is and why it matters.</p><p><strong>WCAG (Web Content Accessibility Guidelines)</strong> defines contrast as a ratio based on relative luminance — the familiar “4.5:1 for normal text” rule. It’s been the standard for over a decade. The problem is that it was derived from research on CRT monitors in the 1990s and has known weaknesses: it treats dark-on-light and light-on-dark as equivalent (they’re not), it doesn’t account for font weight or size properly, and it can produce misleading results at the extremes (very dark colors passing when they shouldn’t, some mid-tone combinations failing when they’re actually readable).</p><p><strong>APCA (Advanced Perceptual Contrast Algorithm)</strong> is the successor, developed for WCAG 3.0. It models contrast differently: it outputs a signed value (positive for dark-on-light, negative for light-on-dark), accounts for the asymmetry between reading directions, and is more accurate for modern display technology.</p><p>Huetone’s built-in contrast display uses APCA, which is one reason it’s a good tool for this work.</p><p><strong>Practically:</strong> use APCA as your primary guide when designing the palette, especially for text color tokens. Keep WCAG compliance tracked as a secondary requirement since it’s still the legal and auditing standard in most jurisdictions. The two will usually agree; where they diverge, APCA is giving you the more accurate picture of real-world legibility.</p><h3>9. The Bigger Lesson</h3><p>When I started this project, I thought fixing the palette would be a technical problem with a technical solution: use the right color model, follow the right algorithm, get a correct answer.</p><p>The reality was more interesting.</p><p>OKLCH is genuinely transformative. The shift from HSL to OKLCH is not incremental — it changes what you can control and what “correct” even means. Having a lightness scale that’s perceptually honest is a different class of tool than one that’s mathematically convenient.</p><p>But the deeper lesson is that <strong>perceptual consistency is a tool, not a goal.</strong> A palette where every 500 shade is at L 0.62 is not inherently better than one where they’re not. It’s a better foundation for building semantic decisions on top of. The structural correctness at the primitive level is what enables you to make intentional, calibrated decisions at the semantic level — which shade of each color to use for which job — rather than making those decisions by feel and hoping they cohere.</p><p>The two-palette compromise — strict uniformity for brand and neutral, high-chroma freedom for feedback — wasn’t a concession. It was a design decision that the system made room for, precisely because I understood the structure well enough to break it intentionally.</p><p>If I were to reduce the whole project to one practical takeaway, it would be this: <strong>start with the lightness scale, and lock it down before you touch anything else.</strong> Everything else in OKLCH palette design is tunable, adjustable, and open to compromise. The lightness scale is the spine. Get that right and the rest of the work has a foundation to stand on.</p><p><em>If you’re working on a similar palette migration, the tools I used — </em><a href="https://oklch.com/"><em>oklch.com</em></a><em>, </em><a href="https://huetone.ardov.me/"><em>Huetone</em></a><em>, and Token Studio in Figma — are all free and available today. The hardest part isn’t the tooling. It’s being willing to sit with a palette that scores perfectly on paper and still say “this isn’t right yet.”</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=018ff626cb4c" width="1" height="1" alt=""><hr><p><a href="https://medium.com/design-bootcamp/how-i-rebuilt-our-design-systems-color-palette-with-oklch-and-what-i-learned-the-hard-way-018ff626cb4c">How I Rebuilt Our Design System’s Color Palette with OKLCH — and What I Learned the Hard Way</a> was originally published in <a href="https://medium.com/design-bootcamp">Bootcamp</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[5 Reasons I love Yulu as a designer]]></title>
            <link>https://medium.com/design-bootcamp/5-reasons-i-love-yulu-as-a-designer-and-want-more-such-services-around-us-65967b11a60b?source=rss-7d07c2c1113d------2</link>
            <guid isPermaLink="false">https://medium.com/p/65967b11a60b</guid>
            <category><![CDATA[design]]></category>
            <category><![CDATA[sustainability]]></category>
            <category><![CDATA[electric-vehicles]]></category>
            <category><![CDATA[urban-mobility]]></category>
            <category><![CDATA[user-experience]]></category>
            <dc:creator><![CDATA[Aakash Mangla]]></dc:creator>
            <pubDate>Thu, 28 Sep 2023 09:44:46 GMT</pubDate>
            <atom:updated>2023-09-29T22:06:35.520Z</atom:updated>
            <content:encoded><![CDATA[<h3>5 Reasons I love Yulu as a designer (and want more such services around us)</h3><p>I spent my last few months savoring the lovely weather in Namma Bengaluru, and it allowed me to experience something truly rare — <strong>traffic-free roads!</strong></p><figure><img alt="A meme showcasing the urge of readers to beat the writer" src="https://cdn-images-1.medium.com/max/498/0*1kB3CaPmvujWq83O.gif" /></figure><p>Wait, don’t roll your eyes just yet! 😂 Obviously, I’m joking.</p><p>The traffic situation in Bangalore is nothing short of legendary chaos. I couldn’t believe that a 5 km ride could take a whopping 30 minutes, yet it’s the norm here! Living a large part of my life in Delhi (and NCR), this was like stepping into a parallel universe for me (Oh! I so missed the Delhi metro while I was in Bangalore).</p><figure><img alt="A station guard pushing the crowd inside the metro train" src="https://cdn-images-1.medium.com/max/498/0*E4lzdvTltRss4GSg.gif" /></figure><p>And let’s not talk about ride-sharing apps like Uber, OLA, or Rapido. The apps that are literally made to help you travel without any hassle are as helpful as a chocolate teapot during peak hours. You practically need to schedule your trip with a 30-minute contingency because, well, you might be sipping coffee while waiting for a driver to accept your ride.</p><p>BUT, amidst all this traffic mayhem, there’s a hero in disguise that helped me reach my work daily without any hassle — the adorable pint-size hero called Yulu! 🛴</p><p>According to Wiki, <a href="https://en.wikipedia.org/wiki/Yulu_(transportation_company)"><strong>Yulu</strong></a> is an Indian technology-driven mobility platform enabling integrated urban mobility across public and private modes of transport. Essentially, it’s a shared electric bike service for short distances or last-mile connectivity.</p><figure><img alt="Yulu bikes standing in the foreground while a couple enjoying tea in the background" src="https://cdn-images-1.medium.com/max/1024/0*EliwGP4taBhLZTOE.jpg" /><figcaption><a href="https://www.yulu.bike/products/miracle/">Source</a></figcaption></figure><p>I first rode Yulu in Delhi around the parts of India Gate in 2020. But I could only realize its huge potential as a shared mobility when I shifted to Bangalore for a few months. And this post is about that.</p><p>Here are my top 5 reasons I love Yulu and want more such services around me in the future.</p><p><strong>Solving Real-World Problems:</strong> Design is about solving problems, and Yulu tackles some of the most pressing issues in urban life — pollution, last-mile connectivity, and traffic congestion. Its smaller footprint doesn’t take up much space on the road and helps you zip through the huge cars on the road. It’s a brilliant solution that simplifies the daily commute and makes our cities more livable.</p><figure><img alt="A dancing traffic police guiding a massive traffic" src="https://cdn-images-1.medium.com/max/498/0*TwH21HFFiuor6Y8A.gif" /></figure><p><strong>User-centered design:</strong> While it solves real-world problems, it doesn’t do so at the cost of the user’s experience. Yulu’s user-friendly app and intuitive experience put the user at the center of the service. It has managed to create a seamless and enjoyable experience during all the phases of a trip — planning, picking, riding, and dropping off.</p><p><strong>Democratizing EVs:</strong> Bangalore is known for its expensive cost of living, and transportation costs can add up quickly, especially when you don’t have a hybrid or WFH working mode. When you choose Yulu, you save almost 50% compared to an auto rickshaw and nearly 75% compared to cabs. So, it’s not wrong to say that Yulu is playing a big part in democratizing EVs in India.</p><p><strong>Building a community:</strong> Only a few products manage to build a community like Yulu has. You must have seen groups of young people riding Yulu around the city and having fun at night. Its unisex and approachable design invites everyone to ride it regardless of identity. The app shows how much carbon you have avoided releasing and fosters a sense of shared responsibility towards our cities. Yulu isn’t just a mode of transport; it’s a community builder. At its core, design is about creating connections, and Yulu does just that.</p><p><strong>Sustainable chic </strong>— Yulu’s design (the service and the product) aligns perfectly with the principles of circularity. It’s an eco-conscious choice that reduces our carbon footprint and showcases the power of green design thinking. It’s electric and tiny in size, much cleaner than any other motored traveling option. It works because it puts users and their needs over the product itself. Yulu is a perfect example of utilizing a circularity in a product.</p><p>In a world where design and innovation can shape the future of our cities, Yulu stands as a shining example of how thoughtful design can revolutionize urban mobility. I’m a fan of Yulu as a designer, and I’m excited to see more services like Yulu transforming the world.</p><p>Cheers!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=65967b11a60b" width="1" height="1" alt=""><hr><p><a href="https://medium.com/design-bootcamp/5-reasons-i-love-yulu-as-a-designer-and-want-more-such-services-around-us-65967b11a60b">5 Reasons I love Yulu as a designer</a> was originally published in <a href="https://medium.com/design-bootcamp">Bootcamp</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[5 reasons why a to-don’t list is as important as a to-do list]]></title>
            <link>https://medium.com/@aakashmangla3/5-reasons-why-a-to-dont-list-is-as-important-as-a-to-do-list-47a40497ccce?source=rss-7d07c2c1113d------2</link>
            <guid isPermaLink="false">https://medium.com/p/47a40497ccce</guid>
            <category><![CDATA[design]]></category>
            <category><![CDATA[notion]]></category>
            <category><![CDATA[time-management]]></category>
            <category><![CDATA[productivity]]></category>
            <dc:creator><![CDATA[Aakash Mangla]]></dc:creator>
            <pubDate>Wed, 01 Mar 2023 17:12:58 GMT</pubDate>
            <atom:updated>2023-03-01T17:12:58.524Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*OYTmt5ui7EYC9vdo" /><figcaption>Photo by <a href="https://unsplash.com/@teamnocoloco?utm_source=medium&amp;utm_medium=referral">Team Nocoloco</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>A few months back, when I embraced <a href="https://www.linkedin.com/company/notionhq/">Notion</a> as my daily productivity tool, it changed my life. It has a well-designed ‘Task List’ where I can track my tasks. It’s super easy, and I no longer have to remember everything.</p><p>However, as I was adding tasks to the list blindly, it grew enormously. I knew what I had to complete, but I didn’t have enough time. Seeing all the work under ToDo tab stressed me. Obviously, something was off my Task List, and I needed to fix it.</p><p>I realized that I was being over-ambitious, trying to complete every single thing without stopping to consider if it was vital or not. That’s when I added a ‘To-Don’t’ tab to my Task List — a tab to keep track of the things I wanted to accomplish but ultimately opted against doing because there were other important things to do.</p><p>Here’s why I think a ‘To-Don’t’ list is as crucial as having a To-Do list.</p><p>1. Re-aligning focus: By adding items to your ‘To-Don’t’ list, you will be forced to assess your current situation and consider your long-term objectives. This habit will help you identify superfluous tasks you dumped in your brain and realign your focus on what matters the most. This, in turn, will help you save unnecessary cognitive processes that are overloading your brain and keep you motivated.</p><p>2. Decision-making skills: Evaluating and crossing off unnecessary tasks will require you to make constant decisions, an important skill you’ll need in today’s time. Prioritizing things and eliminating tasks from your schedule will surely turn you into a good decision-maker over time.</p><p>3. Sense of control: How often do you feel inspired to pursue a task because others are doing it and you feel FOMO? Putting things on the to-don’t list is like saying no to these urges and regaining control of your thoughts. It gives you a sense of autonomy as you are performing tasks while evaluating their purpose.</p><p>4. Avoiding burnout: As mentioned previously, we often overburden ourselves with multiple tasks without thinking of their purpose. We might do this to challenge ourselves or get out of our comfort zone. While a certain stress level is important to perform efficiently, overloading ourselves can have long-term harmful effects. Making a to-don’t list is to avoid the same. It helps us keep track of our mental stress and avoid running into a state of burnout.</p><p>5. Long-term satisfaction: Looking back at the to-don’t list will provide you with a sense of satisfaction even if you skipped those tasks. It is because you have a reason for saying no to it. You know that you made a deliberate decision while skipping it. Compare this to the situation when you try hard to complete a task but cannot because you lack the required motivation or time.</p><p>What do you think about the ‘To-Don’t’ list now? Would you also incorporate one into your workflow?</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=47a40497ccce" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How structured-creativity will help you as a designer?]]></title>
            <link>https://medium.com/@aakashmangla3/how-structured-creativity-will-help-you-as-a-designer-dc6250f01cf1?source=rss-7d07c2c1113d------2</link>
            <guid isPermaLink="false">https://medium.com/p/dc6250f01cf1</guid>
            <category><![CDATA[design-thinking]]></category>
            <category><![CDATA[creativity]]></category>
            <category><![CDATA[innovation]]></category>
            <category><![CDATA[design-process]]></category>
            <category><![CDATA[productivity]]></category>
            <dc:creator><![CDATA[Aakash Mangla]]></dc:creator>
            <pubDate>Wed, 01 Mar 2023 17:06:47 GMT</pubDate>
            <atom:updated>2023-03-01T17:06:47.701Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*3f_Hrp7iIz_xlY2t" /><figcaption>Photo by <a href="https://unsplash.com/es/@amayli?utm_source=medium&amp;utm_medium=referral">Amélie Mourichon</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>As oxymoron-ic as it might sound, structured creativity is something that I believe every designer should learn early in their career. As a creative professional, I understand the importance of creativity in driving innovation and growth. However, unstructured creativity can be elusive and inefficient, especially when tackling complex design problems that require consideration from multiple, sometimes competing, perspectives. This is why I believe structured creativity is essential to unlocking our full creative potential. So, how can we adopt it in our design practice?</p><blockquote>Unstructured creativity can be elusive and inefficient</blockquote><ol><li>Understand the problem you’re trying to solve. Clearly state your goal and success criteria for the solution.</li><li>Break down the problem into smaller, more manageable parts. I call these sub-problems.</li><li>Ideate freely for each sub-problem while also taking care of other sub-problems. This is the step where you will move back and forth between different sub-problems while also realizing the relationship that exists between them.</li><li>Finally, integrate the solutions for each sub-problem such that it forms a holistic solution that passes the success criteria of the original design challenge.</li></ol><p>It’s somewhat like building a house while also shifting focus on individual rooms/spaces when required. It helps maintain focus and avoid unproductive brainstorming sessions, guiding our thinking toward both creative and practical ideas. Adopting a more structured approach to creativity leads to increased productivity and effectiveness.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=dc6250f01cf1" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>