<?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 Iman Khabazian on Medium]]></title>
        <description><![CDATA[Stories by Iman Khabazian on Medium]]></description>
        <link>https://medium.com/@imankhabazian?source=rss-84830807a9cd------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*qhke5yNJZo2P49Ia.jpg</url>
            <title>Stories by Iman Khabazian on Medium</title>
            <link>https://medium.com/@imankhabazian?source=rss-84830807a9cd------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 23:57:33 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@imankhabazian/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[A Quick Guide to the Emergence of LTV]]></title>
            <link>https://medium.com/@imankhabazian/a-quick-guide-to-the-emergence-of-ltv-d2e1cc64d7e4?source=rss-84830807a9cd------2</link>
            <guid isPermaLink="false">https://medium.com/p/d2e1cc64d7e4</guid>
            <category><![CDATA[apps]]></category>
            <category><![CDATA[ltv]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[mobile-game-development]]></category>
            <category><![CDATA[roi]]></category>
            <dc:creator><![CDATA[Iman Khabazian]]></dc:creator>
            <pubDate>Tue, 15 May 2018 21:17:31 GMT</pubDate>
            <atom:updated>2018-05-15T22:35:23.392Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*eWYm3Dm7s0t8lLWYRW4pow.jpeg" /></figure><p>Savvy business leaders are increasingly focusing on the lifetime value (LTV) of a customer to properly grow their business. Of course, LTV is the total amount of money a user will spend in a product over their lifetime and a product is considered profitable when a user’s LTV is greater than the cost it took to acquire that customer and optimal marketing campaigns are those that maximize the ratio between LTV and the cost to acquire a user. Product changes are only considered improvements if they increase LTV.</p><p>Yet as LTV can be sometimes hard to measure, the analytics industry has shifted its focus to reporting KPIs — e.g., the number of downloads, engagement time, and retention at various time intervals. However, these KPIs are often not well correlated to LTV and are never presented with the relevant statistical contexts (for example, statistical significance between 2 possible outcomes). A/B Testing solutions were built on top of this faulty framework. As a result, companies who embraced A/B testing but lacked in-house data science teams haven’t been able to realize the promised boost to profitability.</p><p>For these and other reasons, data-driven decision-making has seen mixed results. Two key facts should be kept in mind as it relates to data-driven decision-making:</p><ul><li>When done correctly, it serves as a major competitive advantage for any digital product. Not only do these products attain profitability at a much higher rate, the product teams take away invaluable learnings that feed into a virtuous improvement cycle.</li><li>When done incorrectly, it creates greater confidence towards inferior paths. To further confound things, product teams learn the incorrect lessons about how changes in their product and marketing efforts affect profitability.</li></ul><p>By borrowing the mathematical symbol &gt;&gt; used to denote “significantly greater than” we can sum up these facts as follows:</p><p>Teams that make data-driven decisions correctly &gt;&gt; Teams that rely on their gut &gt; Teams that make data-driven decisions incorrectly.</p><p>The obvious question becomes, how do you know if you’re making data-driven decisions correctly. Here is a list of criteria:</p><ol><li><strong>Predicted LTV serves as a key indicator by which decisions are made.</strong> On the marketing side, budgets are allocated based on ROI (with predicted LTV as the numerator). On the product side, predicted LTV is the only evaluation criteria for product performance and all other KPIs are simply used to understand drivers in LTV trends.</li><li><strong>When determining the winner between experiments, it’s not enough to understand the impact but also the confidence that one impact is greater than another.</strong></li><li><strong>LTV predictions accuracy can be reviewed. </strong>You should know how accurate LTV predictions have been historically and how much data it takes before they are reliably accurate.</li></ol><p>Once the above criteria are met, teams can expand on the benefits of data-driven decision-making by taking the following steps:</p><ol><li>Don’t just use LTV on marketing spend but also to determine product A/B test winners.</li><li>Utilize a method for running A/B tests that does not require an engineer so you can test at scale.</li><li>Automate LTV reporting on all factors that drive decision-making, like LTV for marketing spends, A/B test variants, and build versions.</li></ol><p>Meeting these criteria has historically been a costly undertaking. Top game publishers are the best example of companies that have made the investment and reaped the substantial rewards. Their data analytics capabilities, and in particular their ability to effectively put LTV correlated KPIs at the heart of all decision-making, effectively boxes out smaller competitors from the most lucrative revenue positions in the app market. In fact, companies like Kixeye now look to double down on data driven decision making by investing in machine learning powered LTV predictions.</p><p>Over the past 5 years, data analytics has played a key role in the operations of any digital product. But now, machine learning tech has enabled predictive LTV to take its rightful place as the heartbeat of decision-making, even for companies that can’t afford a team of full-time data scientists. While smaller companies find it difficult to take the leap because they’ve been burned by previous failed attempts to harness data-driven decision-making, those that find a way past this barrier will enjoy a significant competitive advantage. At Hitsfu, we specialize in filling this gap, helping companies that don’t have millions to invest in data-driven decision making utilize tools and processes to data drive their decisions the right way. If you’re serious about harnessing this competitive advantage for your app, mobile game, or ecommerce site, let’s work together to make it happen.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d2e1cc64d7e4" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to Architect Apps Like a Boss]]></title>
            <link>https://medium.com/@imankhabazian/how-to-architect-apps-like-a-boss-ed19576573e4?source=rss-84830807a9cd------2</link>
            <guid isPermaLink="false">https://medium.com/p/ed19576573e4</guid>
            <category><![CDATA[design]]></category>
            <category><![CDATA[ab-testing]]></category>
            <category><![CDATA[software-architecture]]></category>
            <category><![CDATA[game-development]]></category>
            <category><![CDATA[app-development]]></category>
            <dc:creator><![CDATA[Iman Khabazian]]></dc:creator>
            <pubDate>Mon, 19 Feb 2018 22:29:47 GMT</pubDate>
            <atom:updated>2018-03-09T05:21:25.426Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*nhaJ8RNscXMpj7rlFVRGPA.png" /></figure><p><em>As Mike Tyson says, ‘Everyone has a plan ’till they get punched in the mouth.’ For most app developers, that punch is delivered at launch. They quickly learn that developing an amazing user experience just isn’t enough to turn an app into a hit.</em></p><p><em>This article is the first in a series that aims to help app developers turn their apps into hits. We’ll talk about the secrets of harnessing a data-driven architecture, not just to run more A/B tests, but to significantly improve the overall development process.</em></p><p><em>If you’re ready to implement a data-driven architecture, skip ahead to the ‘</em><strong><em>How to Start Data-Driving Your App?</em></strong><em>’ section.</em></p><p><strong>The Business Case for A/B Testing</strong></p><p>Currently, ten app publishers dominate the $85B app market and rack in a mind-blowing 67% of all app revenues(1). Couple that with Gartner’s stat that 99.99% of apps fail(2) and small app publishers are left with a pretty discouraging outlook for the future of their apps.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/550/1*rbrbGZTGK1cWTTa7TwoHbQ.png" /></figure><p>But, if you explore why these publishers can achieve this level of success, you’ll quickly find that their use of the scientific approach is at the core of their success. Specifically, these publishers know that maximizing their app’s profitability requires that they run lots of A/B tests.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/508/1*VoU7HUdvAlBQ_s-0RWRVDw.png" /></figure><p>I witnessed Disney’s use of A/B testing first-hand while building games for them. We’d run between 20 and 30 concurrent A/B tests at any given time, each yielding tremendous improvements and learnings.</p><p>Netflix is another company whose success is closely tied to it’s devotion to A/B testing. Here’s a fun visual that provides a small peek into the A/B tests they run(3):</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/768/1*awOXpa2FR88jkUokNUwB7A.png" /></figure><p>And here’s a glimpse into Facebook’s view of A/B testing:</p><p><em>“At any given point in time, there isn’t just one version of Facebook running… there are probably 10,000.”</em></p><p><em>– Mark Zuckerburg</em></p><p>The more you look around at the top publishers and at companies like Netflix, Facebook, and Amazon, the clearer it is. Teams that dedicate themselves to running a large array of A/B tests to tune user experiences and thus, maximize profitability, will beat out those that don’t — PERIOD.</p><p><strong>A/B Testing Unchained</strong></p><p>While many development teams think A/B testing is an expensive proposition, top teams have learned how to A/B test while actually reducing the amount of engineering time it takes to push out features and builds. <strong>The key element to achieving this is the data-driven architecture.</strong></p><p>Data-driving your app is the process of moving definitions previously housed in an engineering-only space (e.g. code) into a space that’s accessible to the broader team.</p><p>In this article, we’ll cover the process of moving definitions out of code and into a spreadsheet. Making this change allows non-engineers to manipulate the build, thus allowing you to reap massive efficiencies (discussed further below).</p><p><strong>Why Put Data in Control?</strong></p><p>Apps built using a data-driven architecture have four significant advantages:</p><ol><li><strong>Faster development velocity</strong>: Developers will build systems, not features, which results in less code and fewer bugs. Though a larger upfront cost is required, subsequent development will be faster. This article details the code and tools that will mitigate that upfront cost.</li><li><strong>Faster product iteration</strong>: Product managers can make changes without needing to involve engineers. The benefits here cannot be stressed enough as product managers don’t often know what they want until they are able to iterate. When engineers are required to make those back and forth iterations, morale suffers.</li><li><strong>Feature decisions result in increased profitability: </strong>With the new-found ability to run A/B tests without engineering, teams can run numerous deep A/B tests. And by doing so, they’ll only deploy features that increase profitability.</li><li><strong>Content-rich apps</strong>: Impactful content can be created faster, which in turn allows teams to build a richer end user experience. This is especially important for games.</li></ol><p>Now, let’s take a look at some tangible examples of how a data-driven system would impact your product development processes.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/926/1*QfYhcrwtozn7D0HM_K3WCg.png" /></figure><p><strong>How to Start Data-Driving Your App</strong></p><p>The following steps will help you implement a data-driven architecture for your app. If you have any questions, ping me on <a href="http://m.me/hitsfu">FB Messenger</a>.</p><p><em>Step 1 — Create a new Google spreadsheet that will serve as your app’s content management system (CMS).</em></p><ol><li>Spreadsheets are a familiar interface for non-coders to create and edit complex data sets. Figure out the structure of the spreadsheet that will best empower non-engineers on the team to tune the user experience. Thoughtful layout of the spreadsheet creates an effective interface. Examples are provided below. Once the structure of the spreadsheet is defined, the code will then need to support that structure.</li><li>Think of each sheet of the CMS spreadsheet as a database table. Provided that you name your first column ‘ID’ and populate it with a unique ID, you can use the provided code to grab all of the fields in that row.</li><li>We will provide you with code that will bring this data into your client in a way that is parsed and easy to extract.</li></ol><p><em>Step 1 Example</em></p><p>To demonstrate what this CMS will look like, we’ve pulled together a <a href="https://docs.google.com/spreadsheets/d/1ecB8--bKjeY77wL9hB6MDifjpdiryTx371lBpxGkK7k/edit?usp=sharing">sample CMS based on a popular game, Candy Crush Saga</a>. Scanning through the file should give you a good sense of the structure you’re trying to build.</p><p>First, let’s demonstrate the structure for a tutorial flow for the game. We created a flows sheet where we defined a tutorial that has three steps:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/557/1*k7MDtp3TdZ-uBfnuMVfv5w.png" /></figure><p>We then defined a flow steps sheet where we break down the user gesture and app action associated with each step:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/471/1*CX7TIHrph1w2y-f75_pKlg.png" /></figure><p>Finally, we created a string sheet that defines the text associated with each app action where the app shows a pop-up with text to the user. The below sheet allows us to data-drive all of the text in the app. By doing this we can localize the app simply by switching out the values of this sheet with another, and referring to the IDs in our code:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/683/1*u_S7eN_EXjbU2GImxQHi0A.png" /></figure><p><em>Step 2 — Bring the data from the spreadsheet into your app.</em></p><ol><li>You’ll need to set up a data pipeline to easily publish data and bring it into your app. We’ll use the Clairvoyance service to do this. The CMS functionality is available with a 21-day free trial, after which it costs $8.00/month. To get started, go to<a href="https://clairvoyance.hitsfu.com/"> clairvoyance.hitsfu.com</a> and sign up for a free account.</li><li>Switch on the tutorial button in the upper left and follow the steps to create an app.</li><li>From there you will integrate with your CMS spreadsheet, and in doing so, bind your spreadsheet to the Clairvoyance service to give you a means to publish the spreadsheet data into a file that your app can access.</li><li>As you complete the CMS integration steps, you will be provided with Unity C# code and JavaScript code that allows you to access the spreadsheet data in code through the “static data object.” If you have any difficulties, ping me on <a href="http://m.me/hitsfu">FB Messenger</a>.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/768/1*n6m7luo6kOcLAh4NkihqNA.png" /></figure><p>Moving forward, whenever you need to build a new feature, figure out how you want to data-drive it before building. Check out the examples of data-driving functionality below.</p><p><strong>Data-Driven Examples</strong></p><p>The key architectural goal in a data-driven system is to design the tables and columns in the spreadsheet in a way that empowers non-engineers to drive business outcomes. Designing clear spreadsheets lets everyone, not just programmers, have direct input into the behavior of the product. Below we highlight some ways app developers have set up their spreadsheet to effectively accomplish this goal.</p><p><strong>Prices</strong></p><p>Any price (whether actual or virtual) should be data-driven. This will allow you to tune to the perfect price point.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/624/1*TJm79vop_fp_saDrGrUtYg.png" /></figure><p>Here, the actual prices are defined in the stores, but we can use the data sheets to configure which item from the store is shown to the user.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/618/1*CxS7IHPnYDgyiWP7eqVwQw.png" /></figure><p>More generally, altering any numerical values is a powerful lever. Everything from price to XP to action speeds can be controlled by a small set of numbers, and minor changes in these areas can have huge outcomes.</p><p><strong>Strings</strong></p><p>In order to test out different messages to the user, data-drive every string. Do this by creating a “Strings” sheet in the spreadsheet (tables and sheets are used synonymously here). The first column in all sheets should be “ID” (since values in the spreadsheet can be looked up by sheet name, id, and column name). The next column will be called “text.” Add all English text in the app. Beyond testing messages, we can also localize the app using the localization feature within the Clairvoyance tool.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/626/1*bJu0bpsoHMnngZ-4KmXl9Q.png" /></figure><p><strong>Features</strong></p><p>You may choose to simply create a feature sheet that lists out all features with a Boolean next to it like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/624/1*dNcnn2itK9QY1CS_gdtIQA.png" /></figure><p>Or…</p><p>Instead of just data-driving whether a feature is on or off, you can data-drive the mapping of UI elements to commands. This sheet determines which item should be included in the main menu, and what should happen when each item is clicked:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/557/1*CwQ2NkgP-yzfzVPsLBU-Ag.png" /></figure><p>By data-driving buttons and modals, you can speed up your UI creation process.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/624/1*Je0cnOb9XcuZjTRVZcukvg.png" /></figure><p><strong>Flows</strong></p><p>Data-driving UI sets the stage for data-driving flows.</p><p>In the below example, we have a 5-step account flow, but we can easily remove steps or make changes to the flow in an A/B test because of the daisy-chained structure. Finding the optimal new user flow is essential to having a successful app.</p><p><em>NOTE: Notice that the previous table utilized a comma list, whereas below we use a more complex structure. The more complex structure enables the daisy chaining of steps. In designing the spreadsheet layouts, you will often have to decide how much power is appropriate, and how much complexity you are willing to build into the interface to enable that power</em>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/624/1*bzy0Xs5katIrhmlOKaIjSQ.png" /></figure><p><strong>How to Setup an A/B Test</strong></p><p>One of the coolest things about a data-driven system is that non-engineers can setup A/B tests that override cells in the spreadsheet and deploy those tests without involving engineers. Since engineers are generally focused on fixing bugs and building features, this is realistically the only way to A/B test effectively.</p><p>Once you’ve integrated with the CMS in Clairvoyance, simply click on the ‘A/B Test’ menu item. This is where we’ll add a new A/B test or edit an existing one.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/205/1*iK_YXhFUZswrM4wE-vTuJA.png" /></figure><p>An A/B test is made up of two or more variations that we want to test. Clairvoyance’s CMS allows us to define a variation simply by overriding data from the base spreadsheet.</p><p>Let’s take a look at an example that will show you what screens you’ll be navigating. Below, you can see an existing A/B tests we’ve set up.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/679/1*zOFlz6kGD17h2Zl1csi8Ew.png" /></figure><p>As a quick background, we set up these A/B tests to test an avatar that shows up in our app. We wrote client code that uses the static data object to look up the avatar’s emotion and gender. The static data object is structured as a reflection of the CMS spreadsheet we integrated (refer to the ‘<strong>How to Start Data-Driving Your App’ </strong>section). Any values specified in the spreadsheet can be looked up through it. Also, the data object is defined such that it will also account for changes based on which variations a given user is in, and what cell overrides are specified for those variations.</p><p>Before we deploy our A/B tests (discussed further below), our main spreadsheet shows the emotion as happy and the gender as male. As a result, the app would show a smiling male avatar.</p><p>For our tests, we wanted to see if users spend more time on the app and spend more money in the app given two tests:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/925/1*4OwGCiGf_7KGtI9PHLf2JQ.png" /></figure><p>So for the ‘Emotion’ A/B test, we will set up a ‘sad’ variation. By default, A/B tests include a control variation (with no cell overrides). If you want to exclude the control variation, you can uncheck the ‘Control’ checkbox. Given “Control” is checked, that means the control of ‘happy’ will be included in the test.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/670/1*9aLr3ELaaYjaHy-_RUq4KQ.png" /></figure><p>The variation overrides are defined in the ‘Configure Variation’ screen below. This screen is preloaded with all of the values from the spreadsheet so you can easily specify cell overrides for that variant. Here we enter ‘Sad’ over what was previously a ‘Happy’ value.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/628/1*KeJWgUus4C9T-dsnVSOFVQ.png" /></figure><p>We then click ‘Publish’ once the variation is ready to be tested. From here, we’ll continue to set the next test.</p><p>For the ‘Gender’ A/B test, we set up a ‘female’ variation. Given control is checked, that means the control of ‘male’ will be included in the test.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/674/1*AegwrRcbZ2_l3QrI5ntImg.png" /></figure><p>For every test we set up, the user will be placed in one of the variations. Every user will only be exposed to one variation for a given test. So, in our example tests, users will fall into one of the following combinations of instances:</p><ul><li>Happy male</li><li>Sad male</li><li>Happy female</li><li>Sad female</li></ul><p>In this example, we deployed two tests with two variations in each test, and each variation was defined by overriding a single cell’s value.</p><p>Thus, if you simply set up four A/B tests and each test had three variations, you’d have 3⁴ = 81 instances of your app simply by defining as few as 2+2+2+2= 8 cell overrides.</p><p>A few notes you may find helpful:</p><ul><li>The Clairvoyance Unity and JavaScript sample apps also come with a tester UI that lets you flip between variations, and thus test the A/B tests.</li><li>By default, A/B tests include a control variation (with no cell overrides). If you want to exclude the control variation, uncheck the ‘Control’ checkbox.</li></ul><p><strong>Validation</strong></p><p>As you data-drive more and more, you may start to notice bugs arising from data being incorrectly input. This is a very solvable problem if attacked on two fronts:</p><ul><li><strong>Code Validation</strong>: First and foremost is code validation. Before a value from the spreadsheet is used by code, that code should verify that the value makes sense. If it doesn’t, it should log the value that doesn’t make sense, the reason that the value doesn’t make sense, and it’s full path (sheet name, column name, row ID).</li><li><strong>Validation Rules: </strong>We can also create validation rules from within the spreadsheet using Google script. You may choose to turn a cell red if it refers to a string ID that doesn’t exist. You may choose to have an images column which actually loads the images from the URL that is fed into it. That will load a big red X if that image doesn’t exist. You can even create a “validate” button that runs the entire data set over some set of rules and turns problematic cells red.</li></ul><p><strong>Tips and Tricks</strong></p><p>Often, code that implements A/B tests uses “if” statements to define behavior for each variant. In these cases, we should consider an alternative implementation that creates more powerful data interfaces. For example, the data file may specify our avatar is to have a “Sad” expression. Ideally, we would read the emotion from the static data helper and then simply load the art associated with the value that’s read. This would mean we could support additional emotions without adding code, all because we found a way to do it without the “if” statement.</p><p><strong>But Wait, There’s More</strong></p><p>I hope that this article helps you set the foundation for success for your app. Building a data-driven architecture and turbo boosting your A/B test capabilities is a must for teams who want a clear view of how to make their app more profitable. In our next article, we’ll discuss the next critical step to using the scientific approach to make your app profitability skyrocket…the holy grail: using user lifetime value prediction to pick A/B test winners.</p><p>In the meantime, I’d love to see your comments on what we’ve shared. I look forward to reading your reactions, concerns, and experiences with data-driving and A/B testing.</p><p>1. ‘Top App Publishers of Q1 2017: 10 Largest Accounted for 67% of All Revenue’, published on 4/20/17 (<a href="https://sensortower.com/blog/top-app-publishers-q1-2017">https://sensortower.com/blog/top-app-publishers-q1-2017</a>)</p><p>2. ‘Gartner Says Less Than 0.01 Percent of Consumer Mobile Apps Will Be Considered a Financial Success by Their Developers Through 2018,’<a href="http://www.gartner.com/newsroom/id/2648515"> http://www.gartner.com/newsroom/id/2648515</a></p><p>3. <a href="https://medium.com/netflix-techblog/its-all-a-bout-testing-the-netflix-experimentation-platform-4e1ca458c15">https://medium.com/netflix-techblog/its-all-a-bout-testing-the-netflix-experimentation-platform-4e1ca458c15</a></p><p><em>Originally published at </em><a href="https://www.hitsfu.com/architect-apps-like-boss/"><em>www.hitsfu.com</em></a><em> on February 19, 2018.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ed19576573e4" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>