<?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 Jonas Goddeeris on Medium]]></title>
        <description><![CDATA[Stories by Jonas Goddeeris on Medium]]></description>
        <link>https://medium.com/@jonas.goddeeris?source=rss-11ee79e3399f------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*iZyHVSmYs6WdRIw4.</url>
            <title>Stories by Jonas Goddeeris on Medium</title>
            <link>https://medium.com/@jonas.goddeeris?source=rss-11ee79e3399f------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 07:29:00 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@jonas.goddeeris/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 dashboard is a website, just like any other…]]></title>
            <link>https://medium.com/@jonas.goddeeris/a-dashboard-is-a-website-just-like-any-other-2cb08736cf99?source=rss-11ee79e3399f------2</link>
            <guid isPermaLink="false">https://medium.com/p/2cb08736cf99</guid>
            <category><![CDATA[analytics]]></category>
            <category><![CDATA[ux-design]]></category>
            <category><![CDATA[dashboard-design]]></category>
            <category><![CDATA[dashboard]]></category>
            <dc:creator><![CDATA[Jonas Goddeeris]]></dc:creator>
            <pubDate>Thu, 04 Nov 2021 09:02:42 GMT</pubDate>
            <atom:updated>2021-11-04T09:02:42.911Z</atom:updated>
            <content:encoded><![CDATA[<p>I’ve been making dashboards and optimizing websites for some time. <br>I came to a realization. Dashboards are websites.</p><h4>This means the same rules apply.</h4><h3>UX Design rules</h3><ul><li><strong>Information architecture / Navigation / Simplicity</strong></li></ul><p>Just like any website, the homepage should only contain the most important things (main company/project KPI’s) and it should be easy to navigate to where you wanna go. This <strong>gradual engagement</strong> gives you a clear view of the hierarchy of the information and reduces complexity. These different steps act like <strong>layers</strong>, <strong>each one getting into more detail</strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1CbA57R5n5GK0GyBiJuJyQ.png" /><figcaption>Homepage (dummy data)</figcaption></figure><p>Don’t make different versions of your “main” dashboards depending on the stakeholder group, make 1 version with several layers of detail.</p><p><strong>Organizational bonus:</strong> data transparency and consistency. <br>Everybody is looking at the same data, packaged the same way. <br>This also means different teams work towards the same main goals.</p><p><em>The operational level will rely on the supporting data but they can see what the management sees &amp; visa versa the c-level can have a look at the detailed leading metrics. (leading metrics = “Has an impact on a certain main KPI”)</em></p><p><strong>Maintenance bonus:</strong> no duplicate work maintaining several dashboards with similar data.</p><ul><li><strong>Lingo</strong></li></ul><p>Just as with any website, don’t use business lingo “The business lingo of analytics”. Use plain English. (e.g.: sessions = website visits)</p><ul><li><strong>Gestalt laws</strong></li></ul><p>User Experience is … using Gestalt theory to improve the experience.<br>These psychological principles explain high level how our brain sees patterns everywhere. <a href="https://uxdesign.cc/using-gestalt-theory-to-improve-the-user-experience-66dd71154790">Read more here</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Ey6rD7LEvVwGSgfD.png" /></figure><ul><li><strong>Speed</strong></li></ul><p>On several occasions, I’ve been told a dashboard doesn’t work while it actually does. Especially when you have a big dataset, dashboards can be slow. Cache or chunk it<strong>,</strong> don’t let users wait.</p><h3>Marketing rules</h3><blockquote>I made a new website, now I’ll just share the link and many relevant users will automatically come and visit it regularly.</blockquote><p>Everybody who’s ever made a website knows this doesn’t happen. <br>Driving traffic to a website is hard work. Quality content, Activation via campaigns, emails, reaching out, refining.</p><p><strong>A dashboard needs traffic generation</strong></p><ul><li>Weekly / monthly automated emails from the Dashboarding tool to all stakeholders.</li><li>An activation tour throughout the (relevant) teams &amp; stakeholders to promote the dashboards.</li><li>Feedback loops with the PO, team(s) working on improving the KPI’s.</li></ul><h3>Define success rule</h3><p>Above all Dashboards need a purpose.</p><p>My dashboard is a success if …?</p><ul><li>it provides actionable insights towards improving my main KPI’s</li><li>the people that influence KPI’s use the dashboards</li></ul><p>Add analytics to the dashboard: Introducing the <strong>META DASHBOARD</strong>.</p><h3>Measuring success</h3><p>How can an analytics/data team not use data for its own projects?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/860/1*wGbbfOoFoHOmbc9fcEm_Kw.png" /></figure><p><strong>Page counter:</strong> <br>The idea is simple.<br>Filter the specific dashboard page pageviews in a scorecard metric for the Dashboard analytics data source. Voila, you got a page counter on every dashboard page. Transparent and low effort to gain insights.</p><p><strong>Meta Dashboard</strong></p><p>Are people using this dashboard? <br>If not, talk with the teams, why aren’t they using it?<br>But also, invest your effort into Dashboards that are being used.<br>Build success stories.</p><blockquote><em>Did you notice the weekly peaks ;). Data Newsletter !</em></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DonjEigHS-E-OcgRCBaH4Q.png" /><figcaption>Analytics of the dashboard itself</figcaption></figure><p>Hope you liked it. Feel free to reach out on <a href="https://www.linkedin.com/in/jonasgoddeeris/">https://www.linkedin.com/in/jonasgoddeeris/</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2cb08736cf99" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Automating A/B test Dashboards in Google Data Studio using a CRO data model]]></title>
            <link>https://medium.com/@jonas.goddeeris/automating-a-b-test-dashboards-in-google-data-studio-using-a-cro-data-model-6bdceeee1c53?source=rss-11ee79e3399f------2</link>
            <guid isPermaLink="false">https://medium.com/p/6bdceeee1c53</guid>
            <category><![CDATA[dashboard]]></category>
            <category><![CDATA[data]]></category>
            <category><![CDATA[cro]]></category>
            <category><![CDATA[ab-testing]]></category>
            <category><![CDATA[google-analytics]]></category>
            <dc:creator><![CDATA[Jonas Goddeeris]]></dc:creator>
            <pubDate>Wed, 29 Sep 2021 07:58:07 GMT</pubDate>
            <atom:updated>2021-09-29T08:09:31.001Z</atom:updated>
            <content:encoded><![CDATA[<p>We’ll be looking at a <strong>CRO data model; </strong>using structured data to document Experiments and using this data to automate dashboards and perform a meta-analysis.</p><p>In this example, we’ll be using these tools to make the automated dashboard.</p><ul><li>Google Analytics (API)</li><li>Bayesian stats API</li><li>Google Photos (any photo hosting will do)</li><li>Google Optimize</li><li>Google Sheets</li><li>Google Data Studio.</li></ul><h3>Why should you automate A/B test Dashboards?</h3><p>The effectiveness of any Experimentation driven organization is highly dependent on the number of Experiments they can run.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AODRLG8iGS7aBXLHl_tvIw.png" /></figure><blockquote>This means a fluent &amp; automated process for A/B test (reporting) is necessary.</blockquote><h3>The ‘A/B test’ Dashboard</h3><p>I’m a fan of showing the hypothesis, goals, and variant screenshots in an A/B test dashboard. After all, a test might take some weeks and it’s always good the have some context instead of just showing RAW numbers.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*a8EXu9YzfRcevVITrXkR3w.png" /><figcaption>I do not work for IKEA, data is fictional!</figcaption></figure><p>There are 2 options to make a dashboard such as this in Google Data Studio (GDS).</p><p><strong>The manual way</strong></p><ul><li>Set up the A/B test dashboard template in GDS.</li><li>Use Google Analytics as the Data source.</li><li>Copy-paste the page for every new test</li><li>Change filters and segments for every metric.</li><li>Manually add the hypothesis, screenshots, and goals.</li><li>Make a blended calculated metric for the Uplift %</li><li>Calculate the Bayesian significance in a calculator and add it to the dashboard manually.</li></ul><p>I won’t elaborate on the manual way, play around with GDS, use the Experiment ID (with variant) dimension, and filters. You’ll get there quickly!</p><p><strong>The automated way</strong></p><ul><li>Set up an A/B test dashboard template in GDS.</li><li>Use a Google Sheet for a CRO DATA MODEL for all A/B tests and add all the A/B test info.</li><li>Use a separate Google Analytics API - A/B TEST SRM Google Sheets template. (auto import that data in CRO DATA MODEL sheet)</li><li>Use A/B test name as a filter for the individual A/B test result</li><li>Change the filter for all the (grouped) fields, and all the ‘data’ is ‘automatically’ uploaded, clickable screenshots, test data, hypothesis, uplift, bayesian significance,… everything.</li></ul><p>Paradoxically, the automated way seems more intensive. It’s true that it is a bit more to get your head around, but I assure you changing all metrics in a dashboard is highly time-consuming. ( 🐵 job )<br>While adding some text to a Google Sheet and leveraging APIs, formulas and calculated fields feel like magic 🎩 .</p><p>Like all automation, it requires more work in advance to gain marginal efficiency in the repeat process. So unless you’re planning on testing a lot, you probably shouldn’t bother. But you can test more with automated reporting ;)</p><h3>The CRO ‘DATA MODEL’</h3><p>Think about your CRO process and the data you would like to have for every test. Not only as individual tests but also as a whole. If you had 1000 tests 🤩 , what would you like to know?</p><p>How many tests won? What the success rate is? How many tests won in which part of the Customer Journey? Which psychological principles work well for our Customers, in which part of the funnel?</p><p>To be able to perform such an analysis, you need structured data. Documenting your test as plain <strong>text </strong>in Confluence or some closed tool is useless for this. You need a ‘database’.</p><p>Depending on the scale and maturity of your CRO program you could use a data warehouse like BigQuery, but a Google Sheet will work just fine to get started for most. ( With the experiment ID , you can make a relational database to GA in bigquery, ideal for e.g. getting order hit level revenue data and perform <strong>Mann–Whitney U test)</strong></p><p>Just like Google Analytics has Metrics and Dimensions with different aggregations, so will this.</p><p>Here is what I’ve got so far. (summer 2020)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/358/1*owTw_K8YCtFvCqJLrMi4JA.png" /><figcaption>datamodel blueprint</figcaption></figure><p><strong>DIMENSIONS</strong></p><ul><li>id</li><li>A/B test name</li><li>Hypothesis</li><li>Status (draft, live, ended, canceled)</li><li>Success (boolean)</li><li>Start-end Date</li><li>Hosted screenshot URL Variant A &amp; B (Google Photos with prof. google account)</li><li>Testlane ( Page template where you are testing)</li><li>Brand</li><li>Platform</li><li>Primary goal &amp; Secondary goal(s)</li><li>BMAT, LIFT MODEL, PERSUASION PRINCIPLES</li><li>Specific A/B test data studio URL</li><li>SRM google sheet URL</li><li>Experiment ID</li><li>Monthly missed revenue (if not implemented)</li><li>…</li></ul><p><strong>METRICS</strong></p><ul><li>Uplift %</li><li>Bayesian Significance %</li><li>Users A &amp; B (formula importing from SRM google sheet)</li><li>Converted Users A &amp; B</li><li>CR% A &amp; B</li></ul><h3>The CRO overview dashboard</h3><p>Because of the cleanliness of the data and the calculated fields.<br>Creating an overview of your A/B tests is effortless.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qsxUo6yR1tKi2ItZeOGYxg.png" /><figcaption>I do not work for IKEA , all fictional stuff!</figcaption></figure><p>The lists of AB tests will automatically expand as you add more rows to your CRO data model Google Sheet.</p><h3>Behind the Magic sneak peak: Google Data Studio calculated fields.</h3><p>You can make the filters based on the dimensions in your sheet.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*WOWFFWB-ezO9hQlInu0FOQ.png" /><figcaption>Add the screenshot URL as an image, but add a hyperlink to the same image URL again =&gt; click on thumbnail to open the screenshot full size in a new tab.</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*YTaEbI0PqZGuP4ck-1M2SA.png" /><figcaption>Add a link to the dashboard specific URL to the name so you can click on it for the detailed dashboard.</figcaption></figure><h3>That’s all for now, but i’ll update this post as I’m going.</h3><p>I started writing this post in 2020, but never published, as were getting close to 2022 I doubled down on perfecting the dashboard and removing manual steps. The data model has almost doubled in size.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6bdceeee1c53" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>