<?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 Jack Underwood on Medium]]></title>
        <description><![CDATA[Stories by Jack Underwood on Medium]]></description>
        <link>https://medium.com/@JUnderwood?source=rss-e205131f1c3e------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*Yj5f6twIn5BNV5djl0wi4w.png</url>
            <title>Stories by Jack Underwood on Medium</title>
            <link>https://medium.com/@JUnderwood?source=rss-e205131f1c3e------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 20:22:45 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@JUnderwood/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[SaaS: How increasing the price of our cheaper plan improved sales of our more expensive, target…]]></title>
            <link>https://medium.com/@JUnderwood/saas-how-increasing-the-price-of-our-cheaper-plan-improved-sales-of-our-more-expensive-target-b051a8d1e534?source=rss-e205131f1c3e------2</link>
            <guid isPermaLink="false">https://medium.com/p/b051a8d1e534</guid>
            <category><![CDATA[vc]]></category>
            <category><![CDATA[saas]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[software]]></category>
            <dc:creator><![CDATA[Jack Underwood]]></dc:creator>
            <pubDate>Tue, 06 Aug 2019 12:42:37 GMT</pubDate>
            <atom:updated>2019-08-06T17:50:45.829Z</atom:updated>
            <content:encoded><![CDATA[<h3>SaaS: How increasing the price of our cheaper plan improved sales of our more expensive, target plan by ~100%</h3><p><a href="http://getcircuit.com/"><em>Circuit</em></a><em> is a Techstars ’19 SaaS company building route optimization software that improves the efficiency of last-mile delivery.</em></p><h4>Pricing Plans</h4><p>Circuit’s B2C product has 2 pricing options:</p><ul><li>A $20 plan, allowing drivers to optimize an unlimited amount of stops.</li><li>A $10 plan, limiting drivers to 200 stops per month.</li></ul><p>The idea behind the $10 plan is that it provides value for money for those who have a need for Circuit, but don’t use it all day every day (think Florists, bakeries, etc), without cannibalizing sales to our target customers, who require many more than 200 stops per month.</p><p>The mix for these plans has been relatively steady, with the cheaper $10 plan making up roughly 3/5 of sales, some of whom would eventually upgrade to the $20 plan.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jRjA6hvBNWpUz0t1e5pAmA.png" /><figcaption>Sales per day — Typical product mix</figcaption></figure><h4>1st experiment (Decoy)</h4><p><a href="https://medium.com/u/533a9f1f84f9">Pol Kuijken</a> and I were considering the role of psychology when making a buying decision. Pol suggested that by adding a third “unlimited” plan for $30, and reducing the current $20 plan to ~5000 stops, we’d make our target plan look better value for money, resulting in an improved product mix.</p><p>(Providing 5000 stops is more than enough for our average user)</p><p>I argued that by adding this third plan, we’d actually see a decrease in the overall conversion rate, as the decision on which plan to buy would suddenly become significantly more complicated.</p><p>Upon confirming that 5000 stops/mo was indeed enough for 95% of our customers, we applied the change to see what would actually happen.</p><p>30 days later, it was clear there was:</p><ul><li>A minor improvement in our product mix.</li><li>The overall conversion rate had remained steady.</li></ul><p>Overall, applying that change would have been a small win!</p><h4>2nd experiment</h4><p>Despite the minor win, we remained confused by how the conversion rate could possibly remain steady, despite the significantly harder decision for users. We dove into the data to figure out what we were missing.</p><p>In doing so, we suddenly had a realization. What if the friction that comes with having to decide which plan you needed, was effectively just canceled out by making the $20 look better value for money? If that was the case, could we achieve the same improvement in perceived value for money, but without the friction?</p><p>During our time at Techstars, <a href="https://twitter.com/HugoLisbon">Hugo Augusto</a> (Our MD) had previously spoken to us about focusing more on our target customers (those on the $20 plan), and worrying less about those on the cheaper, $10 plan.</p><p>That conversion sparked an idea. What if the $10 plan was making the $20 look bad value for money? And if that was the case, how could we fix it?</p><p><em>We decided to increase the price of the $10 plan to $15.</em></p><p>The expectation being that whilst fewer people would buy the cheap plan, more people would now buy the $20 plan, as it now looked much better value for money.</p><h4>Results</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*91fVY3EChLVEcDb0ceJXkw.png" /><figcaption>Sales per day — Significant improvement in product mix</figcaption></figure><p>It took less than 2 hours to see the effect of this change - There was an immediate improvement in the product mix.</p><p>Once the dust had settled, the final results suggested a ~100% increase in the sales of our $20 plan, and a ~50% reduction in sales of our cheaper plan.</p><p>Accounting for all changes (including changes to retention), <strong>this change increased our average revenue per install by ~45%.</strong></p><h4>Conclusion</h4><p>We thought our cheaper plan was just capturing an additional market segment for free, but it turns out we were cannibalizing the sales of the segment we actually cared about in the process.</p><p>Interested in more learning like this?<br><a href="https://twitter.com/JUnderwood_1">Follow me on Twitter</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b051a8d1e534" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What we learnt bootstrapping a consumer SaaS startup to $250k in ARR]]></title>
            <link>https://medium.com/@JUnderwood/what-we-learnt-bootstrapping-a-consumer-saas-company-to-250k-in-arr-c7c6978eefbc?source=rss-e205131f1c3e------2</link>
            <guid isPermaLink="false">https://medium.com/p/c7c6978eefbc</guid>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[saas]]></category>
            <category><![CDATA[mobile]]></category>
            <category><![CDATA[route-optimization]]></category>
            <category><![CDATA[delivery]]></category>
            <dc:creator><![CDATA[Jack Underwood]]></dc:creator>
            <pubDate>Mon, 05 Mar 2018 17:57:40 GMT</pubDate>
            <atom:updated>2018-03-07T15:53:05.890Z</atom:updated>
            <content:encoded><![CDATA[<p>Today, <a href="https://getcircuit.com/">Circuit</a> finally crossed the huge milestone that is $250k in ARR!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DK5FHb-hJvRD4UdwnRt32g.png" /></figure><h4>A quick overview</h4><p>Circuit’s a consumer app developed to make delivery drivers’ jobs easier.</p><p>Drivers enter their assigned route for the day, along with any extra constraints such as specific time-windows and priorities, and Circuit uses the data to reorder their route in a more efficient way.</p><p>The result of this reordering results in a ~10–30% decrease in driving time and distance (depending on the quality of the original route), saving drivers money on fuel (if they pay it), and getting them home significantly earlier.</p><p>We let drivers create routes up to 8 stops for free, and charge a monthly subscription to allow them to create and optimize routes up to 200 stops long.</p><h4>#1: Data’s great, but it can be dangerous at low volume</h4><p>I’m in a firm believer in using data to make decisions, and whilst at Circuit we’re using data to make more of our product decisions every week, that hasn’t always been the case.</p><p>In the early months, your usage volume is likely to be be so low that’d be impossible to get signifcant, actionable results from your data. At 50 installs per day, it’ll take you months to figure whether two different layouts of the same feature perform significantly different from each other, so it’s just not worth the additional development time that comes with building and supporting AB testing.</p><p>Not only that, because the early data you receive can be off by factor 2–3 or more, it can send down the completely wrong path, and you won’t even be aware it’s happening.</p><h4>#2: You definitely don’t need to ‘build something you’d use’</h4><p>Trying to build and sell something you don’t use, in a market you know nothing about is hard, no doubt. I can personally count the number of times I’ve used Circuit in the real world on one hand, and even then it was the basic usage that I could have just used Google Maps for. Not only that, I’ve still yet to speak to a user of Cirucit face-to-face, the oppertunity has just never arisen.</p><p>We’ve made a lot of wrong product decisions building Circuit so far, almost all of which can be attributed to a lack of understanding around our customers and market. Many were simple mistakes that could have been prevented if either of us had spent even just a week doing Amazon Flex delivery.</p><p>But ultimately making these mistakes doesn’t matter if you’re able to both realise the issue, and learn from the process.</p><p>At the first signs of potential viability of Circuit we switched focus from improving the product for our users to increasing the amount of usable feedback we were able to collect from them:</p><ul><li>We forced users to sign in with an email address, so we could email when they signed up, subcribed and cancelled.</li><li>We implemented Intercom’s in-app messaging directly into the app, and made it available in multiple places to ensure users could find it when they were having an issue.</li><li>We purposely didn’t expose a way to cancel the subscription in-app, and whilst the majority of users canceled through the Play Store, a minority would ask us to cancel their subscription through Intercom, giving us a direct line of real-time communication with unhappy customers.</li></ul><p>~5000 conversions later, we have a great understanding of our market, despite never using the product ourselves and we make significantly fewer mistakes, although we still make a lot!</p><p>Although it’s true having zero knowledge of the market has hindered our development speed, it was by no means a reason not to attempt to build Circuit.</p><h4>#3: The competition doesn’t always having their pricing right</h4><p>When we first started, we figured the entrenched competition had a better understanding of what users were willing to pay then we did, so we just copied it. That led us to begin with a simple $5/mo subscription.</p><p>Fast forward 15 months, we’ve now been through two separate price hikes, and charge our users $20/mo for unlimited access to Circuit, whilst our main competition still charges ~$5. Both increases led to an immediate and significant increase in average revenue per install.</p><p>We believe pricing for a product should be a function of the value it creates, which in our case is easy to measure. Here’s some (super-conservative) numbers:</p><p><em>A driver spends 40 hours a week delivering.</em></p><p><em>Circuit allows drivers to complete their allocated deliveries 10% faster, saving them 4 hours per week, or ~16 hours per month.</em></p><p><em>At $10 an hour, thats $160 a month of ‘value’ created.</em></p><p>Charging 20% of this figure leads to a price of $32/mo, still signficantly more than we’re charging now, and 6x higher than the competition.</p><p>We were aware of this value we were able to create early on, but our competitor’s pricing made us doubt pricing increases for new customers would be a profitable change.</p><p>Keeping the price at $5 so long was one of our largest mistakes we’ve made so far.</p><h4>#4: PPC/CPA advertising is perfect for bootstrapping new SaaS products</h4><blockquote>If you can make PPC viable, even at super-low volume you know you’re on to something.</blockquote><p>PPC at low volume is super cheap, you can aquire a fair amount of app installs for as low as ~$0.30 each. As soon as you’re able to get average revenue per install (ARPI) higher than your cost per aqusition (CPA), you can kickstart a cycle that for us led to incredible product growth and speed of learning.</p><p>As soon as your ARPI is higher or equal to your CPA, it means you can spend everything you can on advertising at that price, whether that’s limited by the volume available a that price or your budget doesn’t matter.</p><p><em>Even if it’s only $100 a month, this is likely to be significantly more traffic than you’re getting organically right now.</em></p><p>Spending on adverising leads to more installs, more installs leads to more learning and to product improvements, product improvements lead to higher ARPI, and a higher ARPI lets you pay a higher CPA.</p><p>This cycle repeats itself indefinitely, with each cycle fueling further product improvements and revenue.</p><p>Before this inflection point, you might struggle to get enough meaningful feedback to know what to build next, but as soon as the PPC cycle becomes viable, you’re set.</p><h4>More Learnings</h4><p>I hope you found this subset of learnings we’ve shared useful! I’ll try to write more about each (and all the learnings we didn’t cover here) in substantially more detail over the coming months.</p><p>Follow me on <a href="https://twitter.com/JUnderwood_1">Twitter</a> and here on Medium to make sure you don’t miss them!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c7c6978eefbc" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Platform Update: 16–05–2017]]></title>
            <link>https://medium.com/appcam/platform-update-16-05-2017-edcd9767cb98?source=rss-e205131f1c3e------2</link>
            <guid isPermaLink="false">https://medium.com/p/edcd9767cb98</guid>
            <category><![CDATA[android]]></category>
            <dc:creator><![CDATA[Jack Underwood]]></dc:creator>
            <pubDate>Tue, 16 May 2017 18:25:33 GMT</pubDate>
            <atom:updated>2017-05-16T19:53:16.565Z</atom:updated>
            <content:encoded><![CDATA[<h4>An update to the AppCam platform that makes AppCam that much better for UX Designers.</h4><p>A bunch of new features in this update, some of which I really wanted to get into the initial lunch, but ran out of time.</p><h4>Microphone Recording</h4><p>For devices running API 23+ (Marshmallow), the option to record both the screen and microphone is presented. This requires the RECORD_AUDIO permission, however because we’ve limited the feature to API 23+, AppCam can used the &lt;uses-permission-23&gt; tag, and therefore won’t show up as requiring the permission in the Play Store.</p><p>On 23+ devices, AppCam will handle the request permission for you. User’s can choose to opt out of microphone recording.</p><h4>Annotation Mode</h4><p>During a recording, the user can shake their device once to enable ‘annotation mode’. Annotation mode lets the user write on the screen with their finger or stylus.</p><blockquote>Combined with the microphone recording feature, this makes development between remote teams extremely easy.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/360/1*F0HXcNeac9kvWzXByIZLPQ.gif" /><figcaption>Example annotation (without sound)</figcaption></figure><h4>Instructions</h4><p>Those using link-based recording can now provide instructions for their users to carry out during a recording. This is perfect for testing out the UX of tiny parts of your app with simple instructions like “Try adding a papaya to your shopping cart.”</p><h4><strong>Link-generator</strong></h4><p>With all these extra features, we wanted to offer the option to pick and choose between which are enabled. We also wanted a way to make sure the link was correct every-time, and there was no mistypes, so we built a generator.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*cPHVhSXuXATwwm5JdCu2ZQ.png" /><figcaption>AppCam’s Link-Generator in action.</figcaption></figure><h3>That’s it!</h3><p>That’s a wrap for this Platform Update. If you like the look of this feature, you should definitely <a href="http://appcam.io/">try out AppCam in your own app</a>, it’s free.</p><p>To update to this version of AppCam, just include the following:</p><pre>compile &#39;com.appcam.android-sdk:core:0.4.0.1&#39;</pre><p>If you’ve got feedback or feature suggestions that you need to begin using AppCam in your apps, please let us know!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=edcd9767cb98" width="1" height="1" alt=""><hr><p><a href="https://medium.com/appcam/platform-update-16-05-2017-edcd9767cb98">Platform Update: 16–05–2017</a> was originally published in <a href="https://medium.com/appcam">AppCam</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AppCam Platform Update: 10th May 2017]]></title>
            <link>https://medium.com/appcam/appcam-platform-update-10th-may-2017-58f69b485dfb?source=rss-e205131f1c3e------2</link>
            <guid isPermaLink="false">https://medium.com/p/58f69b485dfb</guid>
            <category><![CDATA[appcam]]></category>
            <category><![CDATA[android-app-development]]></category>
            <category><![CDATA[androiddev]]></category>
            <category><![CDATA[android]]></category>
            <dc:creator><![CDATA[Jack Underwood]]></dc:creator>
            <pubDate>Wed, 10 May 2017 10:29:24 GMT</pubDate>
            <atom:updated>2017-05-10T17:11:00.466Z</atom:updated>
            <content:encoded><![CDATA[<h4>Summary of AppCam’s first platform update.</h4><p><em>For anyone not yet familiar with AppCam, </em><a href="https://blog.appcam.io/introducing-appcam-io-929ca1375c8b"><em>here’s an intro.</em></a></p><h3>Crash Uploading</h3><p>AppCam now automatically uploads crash reports for any uncaught exceptions that occur during a recording. Exceptions are linked to the corresponding recording and can be viewed in the dashboard.</p><p>Uploads following crashes will automatically begin uploading 10–15 seconds after the crash, providing the device is connected to WiFi.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3YrJbdLAdlp0h69oaS6UBA.png" /><figcaption>Option to download uncaught exception if one was found.</figcaption></figure><h3>Automatic touch tracking</h3><p>The Android-SDK no longer requires developers to Override onDispatchTouchEvent() in each of their Activities in order to record touch events.</p><p>Touch tracking is now handled silently by the SDK itself, requiring zero implementation from the developer.</p><h3>Implementation</h3><pre>dependancies {<br>   compile &#39;com.appcam.android-sdk:core:0.3.+&#39; // can use .0 instead<br>}</pre><h4>Notes</h4><ul><li>App’s shipping with 0.2.0 will continue to work as usual.</li><li>Method count increased from 343 — &gt;413.</li></ul><p><a href="http://twitter.com/appcam_io">Follow AppCam on Twitter for updates.</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=58f69b485dfb" width="1" height="1" alt=""><hr><p><a href="https://medium.com/appcam/appcam-platform-update-10th-may-2017-58f69b485dfb">AppCam Platform Update: 10th May 2017</a> was originally published in <a href="https://medium.com/appcam">AppCam</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Preliminary Performance of Apple’s Search Ads]]></title>
            <link>https://medium.com/@JUnderwood/preliminary-performance-of-apples-search-ads-90dd553b06c3?source=rss-e205131f1c3e------2</link>
            <guid isPermaLink="false">https://medium.com/p/90dd553b06c3</guid>
            <category><![CDATA[apple]]></category>
            <category><![CDATA[search-ads]]></category>
            <category><![CDATA[marketing]]></category>
            <category><![CDATA[adwords]]></category>
            <category><![CDATA[advertising]]></category>
            <dc:creator><![CDATA[Jack Underwood]]></dc:creator>
            <pubDate>Sat, 06 May 2017 19:09:10 GMT</pubDate>
            <atom:updated>2017-05-06T19:32:01.222Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*WHoy_Lq9bnF5ZLoG2DhP5w.jpeg" /></figure><p>For just over a week I’ve been experimenting with advertising <a href="https://itunes.apple.com/gb/app/circuit-route-planner/id1198232244?mt=8">Circuit for iOS</a> on Apple’s new ‘Search Ads’ platform.</p><p>I now have enough data to confidently say the <strong>performance/dollar it provides is nothing short of incredible.</strong></p><p>Here’s a breakdown of what makes it so good.</p><h3>Costs-Per-Click Performance</h3><p>Based on data from just the UK and US App Stores, Search Ads managed to provide ads at an extremely low cost per click, specifically:</p><ul><li><strong>£0.16/click</strong></li></ul><p>Although these figures are for low volume, and the price would almost certainly increase with volume, these results are incredible when you compare them to the competition:</p><ul><li>Facebook couldn’t provide a single click for less than £0.75.</li><li>Adwords struggled to provide clicks for £0.50.</li></ul><p>Diving a little deeper, Search Ads’ average TTR (Tap through Rate), was 10.5% over the week period, whereas Adwords equivalent (CTR) was 1.83%.</p><p>This actually makes the CPM of search ads ~50% higher than Adwords at £16.81 vs £9.15. I guess that explains the low cost per click?</p><blockquote>It’d be interesting to see if Search Ads continues to provide such a incredibly high CTR after the actual Ads are toned down from their current, overwhelmingly intrusive state.</blockquote><h3>Tap -&gt; Install Performance</h3><p>Again based on data from just the UK and US, Search Ads’ continued to impress.</p><ul><li><strong>43% Conversion Rate</strong></li><li><strong>£0.37/install</strong></li></ul><p>Conversion rate will obviously vary between apps, and unfortunately I don’t have any data from Facebook or Adwords to compare here, but these are again truly impressive figures.</p><p>Firstly, a 43% conversion rate from paid clicks is really impressive when compared to the ~20% long term trend I was seeing from all other sources.</p><p>Perhaps even more incredible is that according to research done by AdEspresso the average cost per app install from Facebook Ads in the US is $2.74. That’s 5.7x higher after currency conversion, and would explain why I was struggling to get a single click for £0.75.</p><h3>Install -&gt; Subscription Performance</h3><p>This data contains slight estimates as Search Ads doesn’t currently provide any information about whether or not installs result in in-app purchases, however for the purposes of this review we can consider it accurate enough.</p><ul><li><strong>5.5% Conversion Rate</strong></li></ul><p>This will again vary from app-app, but this figure is pretty much in-line with the Circuit’s long run average of 5.9% from all other sources.</p><h3><strong>Conclusion</strong></h3><p>Although this data is still firmly from the “low volume” segment, Ad performance far exceeded any other low-volume medium I’d previously tried.</p><p><strong>Search Ads produces a the type of ROI I figured was impossible.</strong></p><p>As a result, I’ll be ramping up the budget up aggressively over the next few weeks to see how much volume can be extracted as reasonable cost.</p><p>I’ll be writing about the results of this ramp up, so follow me here and on <a href="http://twitter.com/junderwood_1">Twitter</a> to a make sure you don’t miss it.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=90dd553b06c3" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Android bug-squashing: AppCam vs Bugsee?]]></title>
            <link>https://medium.com/appcam/android-bug-squashing-appcam-vs-bugsee-6b363338af90?source=rss-e205131f1c3e------2</link>
            <guid isPermaLink="false">https://medium.com/p/6b363338af90</guid>
            <category><![CDATA[debugging]]></category>
            <category><![CDATA[android]]></category>
            <category><![CDATA[app-development]]></category>
            <category><![CDATA[android-app-development]]></category>
            <dc:creator><![CDATA[Jack Underwood]]></dc:creator>
            <pubDate>Fri, 05 May 2017 19:25:45 GMT</pubDate>
            <atom:updated>2017-05-10T10:54:02.155Z</atom:updated>
            <content:encoded><![CDATA[<p>How does the newcomer <a href="http://appcam.io/">AppCam</a> stack up against <a href="http://bugsee.com/">Bugsee</a> when it comes to squashing bugs in Android apps?</p><p>Both offer the ability to automatically record and upload user recordings to an online dashboard, but that’s pretty much where the similarities end.</p><p><a href="https://blog.appcam.io/introducing-appcam-io-929ca1375c8b">Here’s an intro to AppCam, incase you’re lost.</a></p><h3>Basic Recording</h3><p>Both AppCam and Bugsee provide great tools for bug fixing during internal testing. Both pair recordings with static device info like Device model, Android version and App version, however Bugsee goes one step further and also provides the dynamic CPU + RAM usage alongside the recording.</p><p>Both SDK’s can show touch-points on the recordings automatically.</p><p>AppCam’s recordings are significantly higher frame-rate + quality, but that comes at the expensive of recording size and therefore data usage. AppCam solves this problem by scheduling uploads for the next-time the device connects to wifi, and deleting recordings from the device once the upload completes.</p><p>Bugsee has additional features that AppCam does not yet posses, such as the ability to block our certain views (like password fields) for privacy reasons, and ability to also upload console logs and system traces.</p><h3>Crash Recording</h3><p>Both AppCam and Bugsee can record the moments leading up-to a crash, and both SDK’s also automatically upload the stack-trace to their corresponding online dashboards for viewing.</p><p>AppCam will only capture these traces when the recording is running, whereas Bugsee has the option to automatically capture all crashes, or only those that occur during video recording.</p><h3>Recording Initiation</h3><p>Bugsee requires recording initiation to happen in the Applications’s onCreate() function, which means recording has to begin when the app is started. This means each time the app is launched, the user will be greeted by a “Sample App wants to record your screen” dialog, and a persistent notification that Bugsee is running.</p><p>Unfortunately, this pretty much makes shipping Bugsee in a production app a no-go.</p><p>AppCam doesn’t have such a limitation. It can be initiated at any time, from anywhere in the app. Users will still see the dialog (both app’s use the same underlying API that forces this), and there’s no persistent notification. Not requiring to be launched in the Applications onCreate() means that AppCam can be shipped in production, as for 99% of users it’ll sit idle, using zero resources.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*94Im_9W-O1-nNvYia5HEkg.png" /></figure><h4>AppCam can also be launched via a link.</h4><p>AppCam’s got it’s own unique way of starting a recording. Simply send a special link containing your API key to a user, and when clicked AppCam will intercept the link, inform the user of what’s about the happen and start recording.</p><p>This means you can send support emails like this:</p><p>“Hi User,</p><p>Thanks for the bug report!</p><p>Would you mind sending us a recording of this bug in action?</p><p>You can click _this link_ to automatically send us a recording.”</p><p>Instead of exposing A “Start Recording” button inside the UI.</p><h3>Implementation Difficulty</h3><p>As mentioned before, Bugsee requires developers to initate Bugsee in the app’s Application class, so you’ll need to create one(if they haven’t done so already) to initiate Bugsee in.</p><p>If you’re using link-based initiation, AppCam requires not a single line of code. You just need to add the Gradle dependency and AppCam does some fancy internal stuff to initiate itself when needed. If you’re using manual initiation, AppCam requires one line of code to start the recording.</p><h3>SDK Size</h3><ul><li>Bugsee’s SDK contains 11,700 methods, and adds 1.7mb to your APK.</li><li>AppCam’s SDK contains 407 methods and adds 50kb to your APK.</li></ul><p>Bugsee’s SDK is pretty huge. It’s not so much of an issue if you’re only deploying it locally, but it’ll still hurt build-times if it forces you into Multi-Dex territory.</p><h3>Pricing</h3><p>Bugsee allows unlimited recording from the first 5 devices it’s used on, after that the you’ll need a subscription, with the cheapest package starting at $99/mo, for up to 10,000 devices.</p><p>AppCam allows 30 minutes of recordings for free each month, after that you’ll need a subscription, with the cheapest package start at $19/mo for 2 hours of monthly recording uploads.</p><h3>Conclusion</h3><ul><li>If you need super detailed information about crashes, I.E not only a recording, basic device information and a stack trace, but also information about recent network requests, CPU and RAM usage, you should probably go with Bugsee.</li><li>If you want to record bugs in a production environment, but want 99% of users to have an unimpeded experience, you should consider AppCam.</li><li>If want higher quality recordings, don’t want to hit the DEX limit, want minimize the size of your APK, or want to initiate recordings with a link, then <a href="http://appcam.io/">AppCam</a> is also probably better suited.</li></ul><p>Regardless of your option, if you found this useful please leave a *HEART*!</p><p>Also be sure to <a href="http://twitter.com/appcam_io">follow AppCam on Twitter</a> as we’ve got a ton of powerful new features in the pipeline that’ll you’ll definitely want to check out regardless of whether you chose AppCam or Bugsee.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6b363338af90" width="1" height="1" alt=""><hr><p><a href="https://medium.com/appcam/android-bug-squashing-appcam-vs-bugsee-6b363338af90">Android bug-squashing: AppCam vs Bugsee?</a> was originally published in <a href="https://medium.com/appcam">AppCam</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Introducing AppCam.io!]]></title>
            <link>https://medium.com/appcam/introducing-appcam-io-929ca1375c8b?source=rss-e205131f1c3e------2</link>
            <guid isPermaLink="false">https://medium.com/p/929ca1375c8b</guid>
            <category><![CDATA[android-app-development]]></category>
            <category><![CDATA[android]]></category>
            <category><![CDATA[android-development]]></category>
            <category><![CDATA[ux]]></category>
            <dc:creator><![CDATA[Jack Underwood]]></dc:creator>
            <pubDate>Thu, 04 May 2017 11:50:06 GMT</pubDate>
            <atom:updated>2017-05-05T11:53:04.723Z</atom:updated>
            <content:encoded><![CDATA[<p>I’m so excited to finally reveal what I’ve been working on for the last month.</p><p>It started as a side-project for <a href="https://medium.com/u/59b6046f39f2">Circuit</a>, but things escalated quickly as early feedback suggested AppCam had the potential to be useful in all aspects of app development, from bug fixes and to UX reviews to client feedback.</p><h3>What is AppCam?</h3><p>AppCam is a platform built for Android development teams that makes it incredibly easy to record users interacting with your apps — You just send your users a link to click!.</p><p>AppCam provides a tiny SDK that handles the requesting permission, actual recording, processing and uploading of the screen recordings to an online dashboard for review.</p><p>The SDK will intercept the link and start the recording process. It’ll then automatically upload the video to your dashboard the next time the user connects to wifi. Simple right?</p><h3>Get Started with AppCam</h3><h4>Implementing the Android SDK</h4><p>Implementing AppCam’s Android SDK couldn’t be simpler. It’s literally just one line of code.</p><p>Put this line in the build.gradle file:</p><pre>dependencies {<br>     compile &#39;com.appcam.android-sdk:core:0.2.0&#39; // &lt;-- This line<br>}</pre><p>That’s it — you’re fully implemented.</p><h4>Building your recording link</h4><p>Now you’ve implemented the SDK, the next step to build the link you’ll use to request recordings from users.</p><p>Login to <a href="https://appcam.io/">AppCam.io</a> and go to the instructions page. At the bottom you’ll find an API key, take a note of this.</p><p>You’ll also need to your apps package name, it’ll be something like ‘com.company.app’. If you don’t know it, ask your developer.</p><p>Once you’ve got both pieces of information, swap out the [ ]s in the following link for the respective information. Be sure to remove the [ ]s.</p><pre><a href="https://appcam.io/record.html?package=[YOUR">https://appcam.io/record.html?package=[YOUR</a> APPS PACKAGE]&amp;key=[YOUR API KEY]</pre><h4>Recording a video</h4><p>Email the link we just built to the device that has the updated app installed. When you click the link, the AppCam SDK that’s now included in your app will intercept the link, and show you this activity:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*94Im_9W-O1-nNvYia5HEkg.png" /></figure><p>Follow the instructions and start recording. AppCam will record until the app is sent to the background via the home/back button, or the screen is turned off.</p><p>The recording will be scheduled to upload the next time the device connects to wifi.</p><p>You’ll receive a on-device notification once the upload has completed.</p><h4>Viewing the recording</h4><p>Once the recording has been uploaded, it’ll be available is your <a href="https://appcam.io/">online dashboard</a> 10–15 seconds later.</p><p>The free version of AppCam allows you 30 minutes of recording uploads for free each month.</p><p>This’ll be more than enough for vast majority of indie developers that will occasionally use AppCam to solve a difficult bug.</p><p>If you’re a small team consistently using AppCam for bug fixes, or if you’re running UX tests, you might need to <a href="https://appcam.io/usage">upgrade to a paid plan</a> — the $19/mo plan is enough to get started.</p><blockquote>That’s it. You’ve successfully made use of AppCam!</blockquote><p>AppCam’s got some other cool features like showing touch points, and adding additional meta-data to links, you can view detailed instructions on how to implement this feature on the <a href="https://appcam.io/instructions">instructions page</a>.</p><p>That’s a quick summary of AppCam! I’d love to hear your feedback on the platform and how easy/difficult you found the implementation instructions.</p><p>You can follow <a href="http://twitter.com/appcam_io">AppCam on Twitter</a>, we’ve got some seriously cool stuff planned for the near future. Oh, and be sure to share AppCam if you think others would fine it useful!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=929ca1375c8b" width="1" height="1" alt=""><hr><p><a href="https://medium.com/appcam/introducing-appcam-io-929ca1375c8b">Introducing AppCam.io!</a> was originally published in <a href="https://medium.com/appcam">AppCam</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The quality of your app’s organic traffic will fall as it grows.]]></title>
            <link>https://medium.com/@JUnderwood/the-quality-of-your-apps-organic-traffic-will-fall-as-it-grows-6d1ea5e542b2?source=rss-e205131f1c3e------2</link>
            <guid isPermaLink="false">https://medium.com/p/6d1ea5e542b2</guid>
            <category><![CDATA[apps]]></category>
            <category><![CDATA[mobile]]></category>
            <category><![CDATA[ios]]></category>
            <category><![CDATA[android]]></category>
            <category><![CDATA[saas]]></category>
            <dc:creator><![CDATA[Jack Underwood]]></dc:creator>
            <pubDate>Sat, 08 Apr 2017 17:11:20 GMT</pubDate>
            <atom:updated>2017-04-08T18:50:11.622Z</atom:updated>
            <content:encoded><![CDATA[<p>I’m writing about my learnings founding and growing <a href="https://getcircuit.com/">Circuit</a>, my ‘Route Optimization as a Service’ startup. Follow me on <a href="https://twitter.com/JUnderwood_1">Twitter</a> for future learnings.</p><p><em>This applies to apps that solve problems, such as those in the tools, productivity, business categories. It probably applies to other apps too, but it doesn’t apply to games.</em></p><h3>The early days are the best days.</h3><p>In the early days, as you grow from 1 — &gt; 10 organic app installs per day, things are going to be great! The quality of your app installs will be really high, and the few users you do manage to acquire will probably convert pretty well.</p><p>As you grow from 10 — &gt; 100 organic installs per day though, the quality of the users installing your app is probably going to go down, and there’s nothing you can do about it, it’s just something you’ve got to accept.</p><p>When I talk about quality, I’m referring to things like:</p><ul><li>How likely that user is to make it through your onboarding without giving up.</li><li>How likely they are to try and figure something out if they’re stuck.</li><li>How likely they are to convert to a paid customer.</li></ul><h3>Why’s this going to happen?</h3><p>Which of the following users is more likely to put in the time to figure out exactly how to use your app?</p><ul><li>A user who came across your app in a “related apps” of the store.</li><li>A user who found your app at #14 in the search listings.</li></ul><blockquote><strong>It’s the search listings user, because they put in far more effort to find the app to begin with.</strong></blockquote><p>For a user to be installing an app at #14 in the search listings, there’s a good chance they’ve installed at least one other app that might have solved their problem before yours.</p><p>If they’ve already tried another app, decided it didn’t work for them, and then moved onto another, there’s a good chance they put in some effort to figure out if that first app could actually solve their problem before uninstalling it. If they did that for the first app, they’ll do that for yours too.</p><p>Even if they haven’t installed another app, they’ve made it to #14 in the search listings, that’s further than most. They’re obviously invested in solving the problem they’re having, and willing to give potential apps the time necessary to provide them with the value they’re looking for.</p><h3>Your app gets easier to find over time.</h3><p>To increase your organic installs, your app must be easier become easier to find in the store. Once is becomes more visible however, users won’t have to put as much effort in to actually find it anymore.</p><p>This means users who are not likely to put any real effort into understanding your app and the value it provides, wouldn’t have been weeded out yet.</p><p>This doesn’t mean you you’ll miss out on the high-quality users you were getting when you were ranking lower, you’ll keep those and then some, but you’ll also gain new, lower quality users, and the average quality of each install will fall.</p><h3>It’s not something you’ve done.</h3><p>If you’re not ready for this to happen, like I wasn’t, you’ll probably see you metrics suffer and think it’ss your fault.</p><p>You’ll probably think it was one of the recent updates you made. You’ll spend hours looking through all the metric data you’ve got, trying to figure out what’s happening. You’ll roll back to when your app was actually worse than it is now, just to see if it fixes things.</p><p>Don’t. Nothing is actually wrong.</p><p>This is just what happens when apps grow from nothing to something.</p><h3>Circuit’s Experience:</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pVzDyDsdKpejhSq9Cr7KNQ.png" /><figcaption>Percentage of users making it to each stage of the app.</figcaption></figure><p>The graph show’s a comparison of two cohorts of users. One that includes every new user in the past 90 days, and one that only includes the past 30.</p><p>As you can see,<strong> users that installed in the past 30 days actually converted lower at every stage of the funnel.</strong></p><p>The difference amounted to over 25% by the time users reach the end.</p><h3>Takeaway:</h3><ul><li>Quality of organic app installs will drop in the transition from the super early days to the early days. It’s not your fault and there’s nothing you can do about it.</li></ul><blockquote>If you found this useful, leave a like and follow me on <a href="https://twitter.com/JUnderwood_1">Twitter</a> and here on Medium for future learnings.</blockquote><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6d1ea5e542b2" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why ‘Churn’ needs rethinking.]]></title>
            <link>https://medium.com/@JUnderwood/why-churn-needs-rethinking-fcdf97bc1568?source=rss-e205131f1c3e------2</link>
            <guid isPermaLink="false">https://medium.com/p/fcdf97bc1568</guid>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[metrics]]></category>
            <category><![CDATA[saas]]></category>
            <category><![CDATA[churn]]></category>
            <dc:creator><![CDATA[Jack Underwood]]></dc:creator>
            <pubDate>Wed, 08 Mar 2017 08:43:10 GMT</pubDate>
            <atom:updated>2021-05-12T13:23:34.762Z</atom:updated>
            <content:encoded><![CDATA[<p><em>I’m blogging about my learnings founding and growing </em><a href="https://www.getcircuit.com/"><em>Circuit, my Route Optimizing SaaS startup.</em></a><em> I’m doing this partly so I can refer to them later, partly to promote Circuit, and partly so others can learn from my mistakes.</em></p><p>The most basic definition of the churn rate is the number of customers that were lost over a period of time, divided by the total number of customers at the beginning of that period.</p><p>The problem with this simplified version of churn (which for the purposes of this article I’m renaming “Average Churn”) is that the results it provides are extremely misleading.</p><h3>New Users vs Existing Users</h3><p>Let’s take two random customers from an imaginary SaaS company. This company charges it’s users monthly, and has no free trial period.</p><ul><li>Customer 1 has been paying for the service for 8 months.</li><li>Customer 2 is only 15 days in, and as a result has only paid for 1 month.</li></ul><p>Which customer is more likely to renew next month?</p><p>Customer 1, by far.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bbwcDH4pyAoOnoiydOLn0A.png" /><figcaption>Example “Churn Chance vs Customer Age” Graph</figcaption></figure><p>Customers are far more likely to churn when they first start using the product, for a number of reasons. Maybe your onboarding sucks, maybe marketing’s messaging doesn’t align with the product, or maybe your customers just can’t derive enough value from your product to justify the continued use.</p><p>Whatever the reason, an individual customer’s chance of churning is by far the highest in the initial “New Customer” period.</p><p>A customer that has already made it to month 8 has already figured out how to use your product, found workarounds for it’s bugs and figured out how to derive value from it. As an “Existing Customer”, there’s now only a handful of reasons left as to why this customer may churn, versus the hundreds as a “New Customer”.</p><blockquote>Average Churn fails to differentiate between these two segments of customers.</blockquote><p>Because of this failure to differentiate, Average Churn has a number of glaring problems.</p><h3>1.) Average Churn can hide significant issues.</h3><p>Tracking only your Average Churn can will leave you open to missing otherwise painfully obvious issues with certain areas of your company.</p><p>Let’s look at this extremely simplified example of 2 companies, growing at 100 customers per month, for 12 months:</p><h4>Company A</h4><ul><li>20% churn first month (new customers)</li><li>6% churn thereafter (existing customers)</li></ul><h4>Company B</h4><ul><li>35% churn first month (new customers)</li><li>3% churn thereafter (existing customers)</li></ul><p>These are clearly two very different companies. Company B has an obvious problem with initial churn, likely to do with issues with on-boarding, or the message being put across by the companies marketing department. Company A is far more balanced.</p><p>Yet, despite the obvious problems with Company B, <strong>both companies have an Average Churn rate of 8%.</strong> (<a href="https://medium.com/@JUnderwood/why-churn-needs-rethinking-fcdf97bc1568#.bftcslqtq">Here’s the math</a>)</p><p>If you’d only seen the Average Churn figure for both companies, you wouldn’t have been able to identify the obvious issue Company B has with it’s first month.</p><blockquote>Segmenting between new and existing customers makes it far easier your problem areas.</blockquote><h3>2.) Average Churn naturally goes down over time.</h3><p>The grouping of new and existing users together when calculating churn will also lead you to be believe (at least in the early days) that you’re making progress, when really you aren&#39;t.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rO3PxVkeBN2loZyYl7EgOw.png" /><figcaption>Company A, Churn Over Time</figcaption></figure><p>Using the same data for Company A (growing at 100 customers/month, new/existing churn always steady) you can clearly see that between month 2 and month 11, despite zero improvement to product, marketing or customer success, the ‘Average Churn’ fell from 14% to 8%.</p><blockquote>This is a trap.</blockquote><p>What you’re actually seeing is a greater and greater percentage of your customer base move in the “Existing Customer” rather than “New Customer” segment, where the chance of them churning is far lower.</p><p>Your product has not got any better and your customer success team are not doing a better job of keeping customers happy, it’s just a by-product of using Average Churn.</p><p><em>Note: In a startup that’s growing customers month-month, the effects will be less pronounced, as each month there will be more and more customers added to the ‘New Customers’ segment than lost.</em></p><p>3.) It messes up the LTV formula.</p><p>This is the big one particularly for new SaaS companies. You’re still trying to figure out how much each customer is worth to you, but you only have a few months data.</p><p><em>If you use the standard LTV formula of [Revenue Month / Average Churn], you’re going to get a completely wrong figure.</em></p><p>Let’s again use Company A, and we’ll assume they’re selling at $49/mo. Here’s their metrics:</p><h4>Month 3</h4><ul><li>Average Churn: 13.78%</li><li><strong>LTV: $356</strong></li></ul><h4>Month 11</h4><ul><li>Average Churn: 7.96%</li><li><strong>LTV: $615</strong></li></ul><p>Despite zero changes to the segmented churn rates (either New / Existing), the LTV estimate using the Average Churn measured during month 3 was 43% lower than when measured during month 11.</p><blockquote>You can’t make good business decisions with this kind of data.</blockquote><blockquote>If you’re growing revenue each month, by definition you’re always going to have a larger proportion of users in “New User” stage, so your LTV, as derived from Average Churn, will always be a low estimate.</blockquote><h3>Takeaways</h3><ul><li>Work out your ‘Segmented Churn’, rather than your ‘Average Churn’ if you want to use the data.</li><li>LTV derived from average churn will always be a low estimate if you’re growing.</li><li>Average churn naturally goes down over time, it doesn’t mean your making progress.</li></ul><p>Do you agree? Let me know!</p><p>Follow me on <a href="https://twitter.com/JUnderwood_1">Twitter.</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fcdf97bc1568" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[You might not have the credibility to pull off annual pricing.]]></title>
            <link>https://medium.com/@JUnderwood/you-might-not-have-the-credibility-to-pull-off-annual-pricing-42d4a44deacd?source=rss-e205131f1c3e------2</link>
            <guid isPermaLink="false">https://medium.com/p/42d4a44deacd</guid>
            <category><![CDATA[saas]]></category>
            <category><![CDATA[bootstrap]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[pricing]]></category>
            <dc:creator><![CDATA[Jack Underwood]]></dc:creator>
            <pubDate>Tue, 28 Feb 2017 11:44:33 GMT</pubDate>
            <atom:updated>2017-02-28T13:34:22.038Z</atom:updated>
            <content:encoded><![CDATA[<p><em>I’m blogging about my learnings founding and growing </em><a href="https://www.getcircuit.com/"><em>Circuit, my Route Optimizing SaaS startup.</em></a><em> I’m doing this partly so I can refer to them later, partly to promote Circuit, and partly so others can learn from my mistakes.</em></p><h3>Doing the Research</h3><p>Being new to SaaS, when it came to making a decision on pricing for Circuit, I looked elsewhere at what all the other successful SaaS companies were doing.</p><p><em>Edit: A</em>dmittedly<em>, I should have looked specifically at B2C SaaS companies, rather than SaaS in general, but I would have got similar results.</em></p><p>For the most part, most startups we looked at:</p><ul><li>Had multiple plans.</li><li>Had an obvious “preferred plan” that they wanted to direct users to.</li><li>Pushed users towards annual subscriptions.</li></ul><p>Eventually we settled on (a variation of) this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/360/1*tcKphYXvLz8zy-cn-9M1kw.png" /></figure><p>Two upgrade plans, with the current free plan at the top so we could centre the plan we thought would be most popular. Two payment options for each plan, with the annual option prioritised.</p><h3>Unexpected results</h3><p>Over the next two months, we sold ~90 subscriptions to Circuit, 80% of which were for the Plus plan (which was to be expected).</p><p><strong>Out of ~90 subscriptions we sold, zero were annual subscriptions. Zero.</strong></p><p>Surely this had to be a mistake. There was obviously some bug preventing users from buying the annual subscription, right?</p><p>Nope. Users were just not purchasing the annual subscription.</p><h3>Talking to customers</h3><p>After managing to talk to a few customers who did purchase the monthly subscription, it couldn’t have been more obvious the reason why were struggling to sell the annual (which we could definitely use, being bootstrapped).</p><blockquote>“What if you’re not still around in a year?”</blockquote><blockquote>“I don’t mean to be rude, but I don’t trust you with a year upfront.”</blockquote><p>Thinking about it, I’m not sure how I’d feel about trusting a brand new company that had only been around for &lt;2 months with a years upfront payment either.</p><h3>Fixing the problem</h3><p>We removed the annual payment option once we figured out users weren’t ever going to use it.</p><p>Once we’d waited long enough for the statistics to be meaningful, we could see that dropping the annual payment options had increased Circuit’s conversion rate by ~20%.</p><p>Whether that was because of the now simpler pricing page, or because we were no longer forcing the users to question “Can I even trust this company with my money?” is debatable, but it was definitely the correct decision.</p><h3>Takeaway</h3><p>When you’re a brand new company, your app has less than 1000 downloads, your website sucks and your customers have never heard of you before, trying to get them to pay a year upfront probably just isn’t going to happen.</p><p>This lesson probably applies to a bunch of other things too. Things that work for large startups with millions of dollars in revenue and venture capital funding aren’t necessarily going to work for your brand new startup.</p><p><a href="https://play.google.com/store/apps/details?id=com.underwood.route_optimiser&amp;hl=en_GB">Check out Circuit on the Play Store</a><br><a href="https://itunes.apple.com/gb/app/circuit-route-planner/id1198232244?mt=8">Check out Circuit on the App Store</a><br>Follow me on Medium and <a href="https://twitter.com/JUnderwood_1">Twitter</a> for more.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=42d4a44deacd" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>