<?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 Marieke Pots on Medium]]></title>
        <description><![CDATA[Stories by Marieke Pots on Medium]]></description>
        <link>https://medium.com/@marieke91?source=rss-88283bda5f98------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/2*jcEDHiDJO06Ohn0CmuNg3w.jpeg</url>
            <title>Stories by Marieke Pots on Medium</title>
            <link>https://medium.com/@marieke91?source=rss-88283bda5f98------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 06:28:59 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@marieke91/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[BigQuery + Cloud Functions: how to run your queries as soon as a new Google Analytics table is…]]></title>
            <link>https://medium.com/data-science/bigquery-cloud-functions-how-to-run-your-queries-as-soon-as-a-new-google-analytics-table-is-17fbb62f8aaa?source=rss-88283bda5f98------2</link>
            <guid isPermaLink="false">https://medium.com/p/17fbb62f8aaa</guid>
            <category><![CDATA[google-analytics]]></category>
            <category><![CDATA[bigquery]]></category>
            <category><![CDATA[cloud-functions]]></category>
            <category><![CDATA[data-analysis]]></category>
            <category><![CDATA[google-cloud-platform]]></category>
            <dc:creator><![CDATA[Marieke Pots]]></dc:creator>
            <pubDate>Fri, 27 Mar 2020 14:04:28 GMT</pubDate>
            <atom:updated>2020-12-17T09:44:43.056Z</atom:updated>
            <content:encoded><![CDATA[<h3>BigQuery + Cloud Functions: how to run your queries as soon as a new Google Analytics table is available</h3><h4>A step-by-step guide on how to make sure your tables, dashboards and data transfers are always up-to-date</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*VcWWrLv-TCm6_iSK9lkCQw.png" /></figure><p>If you have your Google Analytics view linked to BigQuery, you are probably enjoying the advantages it brings. There is no sampling in your data, you can bring in third-party data, you have more control over your data and much more.</p><p>But, there was one thing bothering me. And maybe it is bothering you too. The exact time of the daily export from Google Analytics to the BigQuery table(ga_sessions_YYYYMMDD) can be somewhat unpredictable. This makes it is impossible to run scheduled queries on this dataset without taking a small risk.</p><ul><li><em>You might set your scheduled queries too early. </em>If, for some reason, the Google Analytics export is delayed, you might miss a day of data. You will have to manually correct this every time this happens. Or you might consider using the intraday table. But this might not be as ideal either, since Google states: “Data for the current day is not final until the daily import is complete. You may notice differences between intraday and daily data” (see the documentation <a href="https://support.google.com/analytics/answer/3437719?hl=en&amp;ref_topic=3416089">here</a>).</li><li><em>You might set your scheduled queries too late.</em> You might be safe if the exports are delayed in this case. But, your underlying tables for your dashboards and/or possible transfers to other data systems may not be as up to date as they need to be for your decision makers.</li></ul><p>Of course, in these uncertain times of Corona, you might call this a first-world-problem. But because it is a problem that I can help fix, I will take you through a step by step guide on how to run your scheduled queries in the exact minute that a new daily export from Google Analytics is ready. The result is a much more fail-proof set-up in which your tables and dashboards are always as up-to-date as they can possibly be.</p><p>The steps that we will be taking are:</p><ol><li>Create a filter in Cloud Logging that isolates the daily log that confirms that a new Google Analytics table is ready.</li><li>Set up a Cloud Pub/Sub topic that collect these daily logs.</li><li>Deploy a Cloud Function that runs your scheduled query in BigQuery as soon as the Pub/Sub topic is being updated with a new log.</li></ol><h3>Step 1. Cloud Logging</h3><ul><li>OK. The first thing to do is to open <a href="https://console.cloud.google.com/logs">Cloud Logging</a>.</li><li>Choose ‘Convert to advanced filter’, by clicking on the little dropdown arrow on the right side of the search field. At least, that’s where it was at the time of this writing 😉.</li><li>You can copy/paste the code below to create a filter that will isolate the logs that confirm that a new Google Analytics table (ga_sessions_YYYYMMDD) has been created. Do not forget to fill in your own dataset ID and project ID in the first two lines.</li></ul><pre>protoPayload.serviceData.jobCompletedEvent.job.jobConfiguration.load.destinationTable.datasetId=&quot;[REPLACE_WITH_YOUR_DATASET_ID]&quot;</pre><pre>protoPayload.serviceData.jobCompletedEvent.job.jobConfiguration.load.destinationTable.projectId=&quot;REPLACE_WITH_YOUR_PROJECT_ID&quot;</pre><pre>protoPayload.authenticationInfo.principalEmail=&quot;<a href="/service/mailto:analytics-processing-dev@system.gserviceaccount.com">analytics-processing-dev@system.gserviceaccount.com</a>&quot;</pre><pre>protoPayload.methodName=&quot;jobservice.jobcompleted&quot;</pre><pre>protoPayload.serviceData.jobCompletedEvent.job.jobConfiguration.load.destinationTable.tableId:&quot;ga_sessions&quot;</pre><pre>NOT protoPayload.serviceData.jobCompletedEvent.job.jobConfiguration.load.destinationTable.tableId:&quot;ga_sessions_intraday&quot;</pre><ul><li>If you adjust the timeframe to the last 7 days, you should now see one message for each day. In the screenshot below, you see the unpredictability of the exact time of the export. In this example, sometimes the table was ready before 2pm and sometimes it’s after 3pm. If your screen looks like this, you are ready to go to the next step!</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1019/1*kP0c8th5aBFmEiESLBj8Tg.png" /><figcaption>Screenshot made in Google Logging</figcaption></figure><h3>Step 2. Cloud Pub/Sub</h3><p>Pub/Sub stands for Publish–Subscribe and can be considered as a messaging system between systems. In this case, we want to create a new Pub/Sub topic that collects the logs that we isolated in the previous step. This way, a message will be collected every time a new Google Analyicts table is available.</p><ul><li>Click on “Create Sink” in the screenshot above.</li><li>You can be as creative as you want to be when entering a sink name.</li><li>Choose “Pub/Sub” as Sink service.</li><li>Choose to create a new Cloud Pub/Sub Topic and give it a name that you can remember in the next step.</li><li>Click ‘Create Sink’ and you are ready for the next step.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1019/1*fQnpkCA6p2ZwEg89Bivezw.png" /><figcaption>Screenshot in Google Logging: creating a pub/sub sink and topic</figcaption></figure><h3>Step 3. Cloud Functions</h3><p>In this last step we will create a Cloud Function (written in Python) that runs every time the Pub/Sub topic is triggered. The function then sends a request to the <a href="https://cloud.google.com/bigquery-transfer/docs/reference/datatransfer/rest">BigQuery DataTransfer API</a> to start a manual transfer run on one of your scheduled (on demand) SQL queries.</p><ul><li>Open up <a href="https://console.cloud.google.com/functions">Cloud Functions</a></li><li>Choose to create a new function.</li><li>Choose “Cloud Pub/Sub” as trigger and select the Pub/Sub topic that you created in the previous step.</li><li>Choose “Python 3.7” (or perhaps higher if you read this in the future 🚀) as runtime.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/603/1*TdK4OjLL2SdHmvGAVTf6oA.png" /><figcaption>Screenshot of how to set up the Google Cloud Function</figcaption></figure><ul><li>Copy/paste the code below, and put this in the “Main.py” file.</li></ul><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/dc51377317d7e644d917029ab74d8a01/href">https://medium.com/media/dc51377317d7e644d917029ab74d8a01/href</a></iframe><ul><li>Replace the project ID and the transfer ID with the details of the scheduled query that you want to run.</li><li>You can find these IDs in the configuration tab of your <a href="https://console.cloud.google.com/bigquery/scheduled-queries">scheduled query</a> The resource name shows a URL in which you will find both IDs: “<em>projects/[THIS_IS_YOUR_PROJECT_ID)/locations/us/transferConfigs/[THIS IS YOUR TRANSFER_ID]</em>”.</li><li>In this case, I’m assuming that you have already created the scheduled query that you want to run. If you have not, this is the time to <a href="https://cloud.google.com/bigquery/docs/scheduling-queries">create your scheduled query</a> in BigQuery. It’s best to have the queries set to run ‘on demand’, otherwise it might run too often.</li><li>Add the little piece of code below in ‘Requirements.txt’, so that it has all the packages it needs to run the Python Code.</li></ul><pre>google-cloud-bigquery-datatransfer==1</pre><ul><li>Set “runQuery” as the function to execute.</li><li>Click on deploy!</li></ul><h3>And now: enjoy the automation 👍</h3><p>If you have correctly set up everything, your scheduled query will now start to run as soon as a new daily table has been exported to BigQuery! This means that you can sit back, relax and enjoy that your Google Analytics data is always as fresh as it can be.</p><p><em>If you have any tips, comments or questions, do not hesitate to ask me in the comments below.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=17fbb62f8aaa" width="1" height="1" alt=""><hr><p><a href="https://medium.com/data-science/bigquery-cloud-functions-how-to-run-your-queries-as-soon-as-a-new-google-analytics-table-is-17fbb62f8aaa">BigQuery + Cloud Functions: how to run your queries as soon as a new Google Analytics table is…</a> was originally published in <a href="https://medium.com/data-science">TDS Archive</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What does a long loading time mean for your online business results? (+free tool).]]></title>
            <link>https://medium.com/@marieke91/what-does-a-long-loading-time-mean-for-your-online-business-results-free-tool-d2d2c638fb8c?source=rss-88283bda5f98------2</link>
            <guid isPermaLink="false">https://medium.com/p/d2d2c638fb8c</guid>
            <category><![CDATA[performance-testing]]></category>
            <category><![CDATA[google-analytics]]></category>
            <category><![CDATA[load-time]]></category>
            <category><![CDATA[performance]]></category>
            <category><![CDATA[data-studio]]></category>
            <dc:creator><![CDATA[Marieke Pots]]></dc:creator>
            <pubDate>Mon, 20 Jan 2020 09:08:45 GMT</pubDate>
            <atom:updated>2020-01-20T09:20:54.586Z</atom:updated>
            <content:encoded><![CDATA[<h3>How do loading times impact <br>your online performance? (+ free tool)</h3><p><em>This article has been co-written with </em><a href="https://medium.com/u/fb1c8166822e"><em>Sander Heilbron</em></a></p><p><strong>It is no secret that loading times play a significant role in the experience of website visitors. Many (case) studies draw the same conclusion: the longer a visitor has to wait for a website to load, the greater the chance that he or she will leave quickly. But is this also the case for <em>your website</em>? And: how big is the impact of the loading time exactly? Discover it with the </strong><a href="https://datastudio.google.com/open/1RWJAbST0GE6Zr7O9gbWWsww8EZ1ugoCn"><strong>OrangeValley Loading Time Impact Calculator</strong></a><strong>.</strong></p><figure><img alt="The OrangeValley Loading Time Impact Calculator" src="https://cdn-images-1.medium.com/max/1024/1*hJ7LhjFdifv4pMfUcAxMdA.png" /><figcaption>The OrangeValley Loading Time Impact Calculator</figcaption></figure><h3>Loading time: an important part of the user experience</h3><p>As an online marketer, developer or designer we often pay a lot of attention to user-friendliness and the look-and-feel of a website. However, what that new functionality, that beautiful high-resolution photo or that new third-party tool does with the loading time, is a question that is (too) often neglected.</p><p>More and more studies show that website visitors are impatient. They expect a website to load quickly, that the requested information is quickly visible and that something can be clicked instantly. The longer this takes, the greater the chance that visitors will leave a website early. But, these studies are often very general or very specific. The conclusions can not be applied to all branches, target groups and websites. A legitimate question is therefore: what exactly is the impact of loading time on the behavior of<em> your website visitors</em>? The <a href="https://datastudio.google.com/open/1RWJAbST0GE6Zr7O9gbWWsww8EZ1ugoCn">OrangeValley Loading Time Impact Calculator</a> gives you this insight.</p><p>This <a href="https://datastudio.google.com/open/1RWJAbST0GE6Zr7O9gbWWsww8EZ1ugoCn">calculator</a> is a Google Data Studio dashboard where you can easily connect your own Google Analytics account. The dashboard gives you immediate insight into the impact of a possible slow loading time on your online results. In the following paragraphs we will discuss this dashboard step by step.</p><h3>Google Analytics data as a starting point</h3><p>There are many reasons (such as sampling, availability of mere technical performance metrics, and working with averages) why Google Analytics is not the most suitable tool for collecting data about loading times. Yet this popular tool provides an enlightening first insight into the impact of loading times on the behavior of your visitors. That is, if the following important condition is met: make sure you have collected enough data to draw valid conclusions.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bmO70cAU4lvV-6gAaZ5DBQ.jpeg" /><figcaption>The Loading Time Impact Calculator gives you an indication about the quality of your page speed data.</figcaption></figure><p>With a standard configuration, Google Analytics collects data about the loading time of only 1% of all page views. This is the so-called “sample rate”. This means that all information in Google Analytics about loading times is based on a sample. To have enough and therefore reliable data, there are two things you can do: (1) choose a date period that is long enough so that you have enough “page samples” at your disposal. Another option is to (2) increase the sample rate, which means that you collect more data in less time. In the OrangeValley Loading Time Impact Calculator we maintain a minimum of 10,000 samples as a benchmark, but the more the better. In the screenshot below you can see the effect of increasing the sample rate in mid-november. As a result, more samples per day were collected, the average sample rate for that period increased to 14,51% and we had more than 300,000 samples at hand.</p><h3>What does loading time mean for the engagement of your visitors?</h3><p>The loading time differs per visitor. One visitor has a fast laptop with a fast internet connection, while another visitor has poor reach with his or her smartphone at that time. It therefore makes little sense to see the average loading time in Google Analytics as a single and stand-alone measure: this number only means something when we start comparing the average loading time between groups of visitors (or between time periods).</p><p>In this context it is primarily interesting to create three segments that answer the following questions:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Mat4brkJprnK73vjRY8GNQ.jpeg" /><figcaption>The Loading Time Impact Calculator shows the difference between bouncers and non bouncers.</figcaption></figure><h4>1. What is the difference in loading time between visitors who leave your website immediately (bouncers) and visitors who stay?</h4><p>You often see that the average loading time is much higher among people who bounce, and this difference can sometimes be more than 2 seconds. It shows the importance of keeping the loading time of pages on which people land as low as possible.</p><h4>2. What is the difference in loading time between visitors that view multiple pages per session versus visitors that view only a few pages?</h4><p>Often a trend is visible here: the more pages your visitors view, the shorter their loading time. This trend can be clearly seen in the graph below. The challenge is to reduce loading time so that more visitors will view more pages.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4MQp4bi1WQ1m5PXn6NIMgA.jpeg" /><figcaption>The Loading Time Impact Calculator often shows that the loading time decreases among visitors who view more pages per session.</figcaption></figure><h4>3. What is the difference in loading time between visitors who make a purchase (or another conversion) versus visitors who don’t?</h4><p>Often you will see a similar trend as we saw with the bounce rate: the loading time is lower among visitors with a purchase.</p><p>It is obvious that several factors can be of influence to the figures above. But, it is often clear that improving the loading time can tempt a specific subset of your visitors not to leave the website directly, but to view more pages and even make a purchase.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zl0B8iiyE4HjVL-UpYvvBg.jpeg" /><figcaption>The Loading Time Impact Calculator shows the difference between converters and non-converters.</figcaption></figure><h3>Are there specific pages that deserve attention?</h3><p>We now know that the loading time differs for every visitor and that this has an effect on the experience of the user. The same applies to specific pages. The loading time on a website differs per page, which means that the user experience can also be different here. For example, that one page with that beautiful high-resolution photo can have a much longer loading time, with the result that more visitors drop out.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*xvAAbiQoyG8EHLzCtn7hqw.jpeg" /><figcaption>The Loading Time Impact Calculator creates a scatterplot that gives insight in the correlation between exit rate and average page load time.</figcaption></figure><p>It is therefore interesting to view the average loading time in relation to the exit percentage per page. In the <a href="https://datastudio.google.com/open/1RWJAbST0GE6Zr7O9gbWWsww8EZ1ugoCn">OrangeValley Loading Time Impact Calculator</a> this is shown with a scatter chart. Based on the size of the bulbs, an indication is also given of the most viewed pages. The trend line indicates whether there is a connection between a possible higher loading time and the exit percentage. By identifying outliers, you can immediately see on which pages the loading time is a potential bottleneck.</p><h3>Attention to mobile: also test on slow devices</h3><p>If we take a look at the results per device, it is often noticeable that mobile users behave differently than desktop and tablet users. It is also striking that the loading time is often higher on mobile. The fact that mobile users behave differently is of course not directly due to the loading time; nevertheless there are often great opportunities for optimization here.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*i8DezB88LV5K2b0jdbPnbQ.jpeg" /><figcaption>The Loading Time Impact Calculator compares the difference device categories.</figcaption></figure><p>If we zoom in on the mobile segment, it is notable that the loading time in this group shows large differences. Users with a high-end smartphone often have much lower loading times than smartphone users with a budget or mid-range smartphone. This is because websites are increasingly using JavaScript and the speed of processing JavaScript by processors of smartphones determines the loading time. Popular budget and mid-range smartphones have slower and outdated processors because they are cheaper. And as we know, the loading time has an effect on the involvement that these visitors will exhibit. It offers opportunities to test and develop on frequently used smartphones from your target group.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*nm0nDkmX-bTd32zKrzdu4A.jpeg" /><figcaption>The Loading Time Impact Calculator gives an indication in the type of phones among your visitors.</figcaption></figure><h3>Use RUM tooling for even more insight into loading times</h3><p>If you want to measure more and want to know more about loading times, then it makes sense to look further than data from Google Analytics. Various (free) tools are available. <a href="https://developers.google.com/speed/pagespeed/insights/">Google PageSpeed Insights</a>, <a href="https://developers.google.com/web/tools/lighthouse">Google Lighthouse</a> and <a href="https://www.webpagetest.org/">WebPagetest</a> are examples of a number of commonly used “synthetic” tools that you can use right away. Google’s relatively new Chrome User Experience Report (CruX) provides real user experience data from more domains and pages based on relevant performance metrics (and more and more metrics are becoming available). Based on this, you can also collect data from your competitors, for example, to compare your own website against them. We do this in our <a href="https://orangevalley.nl/dashboarding/site-performance-dashboard/">Site Performance Dashboard</a>, among other things.</p><p>For the most comprehensive insights, we rely on Real User Monitoring tools, also known as RUM. These specialist performance tools offer extensive insight into what users see and experience and therefore go far beyond the technical performance metrics that Google Analytics offers. RUM also has no restrictions in terms of sampling data. Moreover, many RUM tools use percentiles instead of a general and therefore often meaningless average loading time like Google Analytics does. The percentiles give us insight into the distribution (normal distribution) of the loading times, with which we can better analyze which users experience improvements after possible optimisations.</p><p>RUM tooling is also capable of correlating performance metrics with business metrics, but is often complex and relatively expensive. Whichever option you choose, one thing is certain: optimizing the loading time can provide a surprising boost in your results.</p><h3>Want to measure more and know more?</h3><p>The <a href="https://datastudio.google.com/open/1RWJAbST0GE6Zr7O9gbWWsww8EZ1ugoCn">OrangeValley Loading Time Impact Calculator</a> provides a first indication of the impact of the loading time on your online results. Ultimately, you want to offer every visitor a pleasant experience, because happy visitors tend to do more on your website. They are more active and view more pages per session. The chance that they leave the website directly or prematurely is smaller, and the chance that they convert is greater.</p><p>Whatever tool and approach you choose, one thing is for sure: optimizing the loading time provides a surprising boost in your results!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d2d2c638fb8c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What is the actual impact of ITP 2.1 and 2.2 on your Google Analytics data? (+ free tool)]]></title>
            <link>https://medium.com/data-science/what-is-the-actual-impact-of-itp-2-1-and-2-2-on-your-google-analytics-data-free-tool-99e42c5978a6?source=rss-88283bda5f98------2</link>
            <guid isPermaLink="false">https://medium.com/p/99e42c5978a6</guid>
            <category><![CDATA[google-analytics-tips]]></category>
            <category><![CDATA[google-analytics]]></category>
            <category><![CDATA[data-studio]]></category>
            <category><![CDATA[data-analytics]]></category>
            <category><![CDATA[analytics]]></category>
            <dc:creator><![CDATA[Marieke Pots]]></dc:creator>
            <pubDate>Thu, 19 Sep 2019 11:14:56 GMT</pubDate>
            <atom:updated>2019-10-29T12:52:41.960Z</atom:updated>
            <content:encoded><![CDATA[<p><strong>Quite a lot has been written about the impact of ITP 2.1 and 2.2 on Google Analytics data. These articles often remain on a theoretical level by explaining what <em>could be</em> the impact. Let’s make it practical today. What has really changed in <em>your</em> Google Analytics data with the release of ITP 2.1 and 2.2? Let’s find out.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rJUJF92PxN83MqKExAaDXA.png" /></figure><h4><strong>The theory</strong></h4><p>Google Analytics relies on a cookie that is set when you visit a website for the first time. If you would visit the same website again after a couple of days, the same cookie will make sure that you are being recognized as a returning user and your complete journey starts to map. At least as long as you do not delete your cookies, stay on the same device and visit the website at least once in <em>every two years</em>. This last condition changed dramatically with the release of ITP 2.1.</p><p>ITP 2.1 came into the picture when Apple released Safari 12.1 and iOS 12.2 around the end of March 2019. A few weeks later ITP 2.2 was released among iOS users. This impacts the Google Analytics cookie in two ways:</p><ul><li>The Google Analytics cookie will be deleted after 7 days of inactivity (ITP 2.1).</li><li>The Google Analytics cookie will be deleted after 1 day of inactivity if the visitor came from a cross-domain link (ITP 2.2) and only visits one page.</li></ul><p>So, instead of a cookie expiration time of two years, we are now facing a cookie expiration time of 1 to 7 days! This means that a returning visitor cannot be recognized as such if he or she has not visited your website in the last 1 or 7 day(s). The theoretical impact: more sessions by new users, shorter journeys per user and less channels to attribute your revenue to.</p><p>Let’s find out if the theory is right when it comes to <em>your</em> data.</p><h4><strong>Tool: ITP Impact Calculator</strong></h4><p>Do not hesitate to make use of the <a href="https://datastudio.google.com/open/1llqii9oyHXx2kLy6AOHI8AQnKIS1BrX6"><strong>ITP Impact Calculator</strong></a> to find out how ITP 2.1 and 2.2 may have impacted your Google Analytics data. Below you will find an explanation of the output of this tool. Of course, you can also continue reading without using the tool for yourself.</p><h4><strong>How many of your visitors actually use an ITP 2.1 and 2.2 browser?</strong></h4><p>First, it is important to have an idea about the number of visitors that actually use an ITP 2.1 or 2.2 browser. Only versions of Safari 12.1 or higher have ITP 2.1 or 2.2. By creating a segment of this browser and version, you will find out when your visitors have adopted this new browser and what its total share is among your visitors. The more people use such a browser, the more ITP 2.1 will make a mark on your data.</p><figure><img alt="Overview of the amount and total share of sessions from Safari 12.1 or higher" src="https://cdn-images-1.medium.com/max/1024/1*h876GeBuCR9SPhtc4ZBX9Q.png" /><figcaption>The graph on the left shows a timeline of when your visitors have adopted Safari 12.1 (or higher). The scorecards on the right provide insight in the total share of Safari 12.1 (or higher) sessions in the past 30 days.</figcaption></figure><h4><strong>How has ITP impacted the amount of returning users?</strong></h4><p>New or returning visitor? That is the question! With ITP 2.1 it is likely that more sessions will be classified as a session by a new user. That is because visitors who have not visited your website in the last 1 or 7 day(s), cannot be recognized anymore.</p><p>To find out if this is the case in your data, you can take a look at the dimension ‘user type’ and make two segments: one for users with Safari 12.1 or higher, and one for your visitors with any other browser. Often, you will see that the amount of new visitors is much higher among Safari 12.1 or higher users. The difference between the two provides insight into the way ITP 2.1 has impacted this data in your dataset.</p><figure><img alt="Overview of the difference in new and returning users between Safari 12.1 (and higher) users and all other browsers." src="https://cdn-images-1.medium.com/max/1024/1*tZNWxjRiX7BFpPyyQrAUaA.png" /><figcaption>The pie charts indicate how the amount of new and returning users may have changed. In this example: with ITP 2.1, only 30% can be recognized as a returning user. Among other users, 48% can be recognized as a returning user. This is a difference of no less than 18%</figcaption></figure><h4><strong>What changed in the length of your visitor’s journey?</strong></h4><p>Google Analytics keeps track of how many times somebody visits your website with the dimension ‘count of sessions’. With ITP 2.1 and 2.2, it is expected that returning users cannot be recognized. The consequence: their count of sessions will be reset. In other words: we would expect a higher percentage of your users with a session count of just one. By looking at this dimension in Google Analytics or the<a href="https://datastudio.google.com/open/1llqii9oyHXx2kLy6AOHI8AQnKIS1BrX6"> ITP Impact Calculator</a>, you will find out how big this difference is compared to other browsers.</p><p>If you are wondering why the provided percentages in the ITP Impact Calculator do not add up to 100% — I must compliment you on your alertness! This has two reasons. First, it is because the ‘count of sessions’ dimensions looks further in the past than the 30 days in the calculator. This means that somebody may have only had a second visit in this period. The second reason is that this dimension is incremental: somebody can have both a first, second and third visit in the last 30 days. The percentage thus shows which share of your users has had a first, second, third, fourth or fifth session in the last 30 days.</p><figure><img alt="The graph shows the percentage of users who with their 1st, 2nd, 3rd, 4th and/or 5th session count. With ITP 90% of the users" src="https://cdn-images-1.medium.com/max/1024/1*ysavJHn5lZW8ORs9ddtNAA.png" /><figcaption>The graph shows the percentage of users who with their 1st, 2nd, 3rd, 4th and/or 5th session count. With ITP 90% of the users had their first session in the last 30 days. This is was only 58% among other browsers.</figcaption></figure><h4><strong>What does this mean for your marketing channel attribution?</strong></h4><p>The consequences of ITP 2.1 and 2.2 on estimating the value of your different marketing channels is perhaps the biggest challenge. Since fewer users can be recognized, fewer touchpoints in their customer journey can be linked. Imagine this scenario: somebody with the latest version of Safari visits your website for the first time after clicking on a Facebook ad; and after 8 days, this same person searches for your brand via Google and places an order. In this case, this transaction will only be attributed to (organic or paid) search. The rather big role of the Facebook ad that triggered the first contact is neglected.</p><p>The ITP Impact Calculator gives insight into the percentage of revenue that may have been impacted by ITP 2.1 and 2.2 and may thus have been attributed to a wrong channel and/or incomplete customer journey. This estimation is based on (but slightly different to) an <a href="https://online-metrics.com/intelligent-tracking-prevention/">earlier article by Paul Koks</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ZWSbbsApOG_YhA7I6BDG3w.png" /></figure><p><strong>Help! ITP has impacted my data to a large extent, what now?</strong></p><p>So far, a few (technically challenging and time-consuming) workarounds have been proposed. But, as <a href="https://www.simoahava.com/analytics/itp-2-1-and-web-analytics/">Simo Ahava</a> correctly points out: “As workarounds for ITP are invented, new iterations of ITP will be introduced”. It has become a cat and mouse game. In this game, it is especially interesting how Google may or may not find workarounds for ITP in the future.</p><p>For now, I believe that two things are important when it comes to ITP:</p><ul><li>As an analyst, online marketer or marketing manager: be aware of the possible impact that ITP 2.1 and 2.2 may have had on your data and what conclusions you can draw from the data. The<a href="https://datastudio.google.com/open/1llqii9oyHXx2kLy6AOHI8AQnKIS1BrX6"> ITP Impact Calculator</a> gives you a first impression on how much your data has been impacted.</li><li>If possible, try to analyse on session scope rather than user scope. Analyzing sessions is still perfectly reliable in Google Analytics. If you would still like to analyse on user-level, consider leaving ITP 2.1 and 2.2 browsers out of your dataset.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=99e42c5978a6" width="1" height="1" alt=""><hr><p><a href="https://medium.com/data-science/what-is-the-actual-impact-of-itp-2-1-and-2-2-on-your-google-analytics-data-free-tool-99e42c5978a6">What is the actual impact of ITP 2.1 and 2.2 on your Google Analytics data? (+ free tool)</a> was originally published in <a href="https://medium.com/data-science">TDS Archive</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>