<?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 Rubén Jiménez Baz on Medium]]></title>
        <description><![CDATA[Stories by Rubén Jiménez Baz on Medium]]></description>
        <link>https://medium.com/@ruben-jimenez-baz?source=rss-193d6aad3437------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*7ra4ltLBF1aw1KSz8AdYZg.jpeg</url>
            <title>Stories by Rubén Jiménez Baz on Medium</title>
            <link>https://medium.com/@ruben-jimenez-baz?source=rss-193d6aad3437------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Mon, 16 Nov 2026 23:51:50 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@ruben-jimenez-baz/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[CXL Institute — Growth Marketing Minidegree Review | Week 4]]></title>
            <link>https://ruben-jimenez-baz.medium.com/cxl-institute-growth-marketing-minidegree-review-week-4-8d99995417d?source=rss-193d6aad3437------2</link>
            <guid isPermaLink="false">https://medium.com/p/8d99995417d</guid>
            <dc:creator><![CDATA[Rubén Jiménez Baz]]></dc:creator>
            <pubDate>Sun, 14 Mar 2021 21:29:52 GMT</pubDate>
            <atom:updated>2021-03-14T21:29:52.611Z</atom:updated>
            <content:encoded><![CDATA[<h3>CXL Institute — Growth Marketing Minidegree Review | Week 4</h3><p><strong>Module 3</strong> covers <strong>Data and Analytics</strong>, being Google Analytics (GA) its main element of discussion for this part of the <a href="https://cxl.com/institute/programs/growth-marketing-training/">Growth Marketing Minidegree</a>, due to its unquestionable relevance for businesses when uncovering insights and making decisions.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*wI9ObdA8n8KP1fKk" /><figcaption>Photo by <a href="https://unsplash.com/@hostreviews?utm_source=medium&amp;utm_medium=referral">Stephen Phillips - Hostreviews.co.uk</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p><strong>What is covered on Module 3?</strong></p><ul><li><a href="https://cxl.com/institute/online-course/google-analytics-beginners/"><strong>Google Analytics for beginners</strong></a><strong> | </strong>Introduction on how to start using GA, setting up the platform, and start extracting data out of it.</li><li><a href="https://cxl.com/institute/online-course/intermediate-google-analytics/"><strong>Intermediate Google Analytics</strong></a><strong> | </strong>Advance level on attribution, custom reports, metrics, and insights from GA.</li><li><a href="https://cxl.com/institute/online-course/google-tag-manager/"><strong>Google Tag Manager for beginners</strong></a><strong> | </strong>Review<strong> </strong>of Tag Manager fundamentals to ensure organized tracking to extract all its power.</li><li><a href="https://cxl.com/institute/online-course/attribution/"><strong>Attribution</strong></a><strong> | </strong>Consumer behaviour across different marketing channels to ensure proper attribution.</li></ul><p>Additionally this module includes an <a href="https://cxl.com/institute/online-course/excel-sheets-for-marketers/">Excel and Google Sheets</a> course to ensure proper management of data and formulas, to extract additional insights from GA information.</p><p>I might slow my learning speed on this part for a few weeks, as I’m a total newbie to GA on the in-hands experience part of it. I’ve used GA in the past with view permissions, learn data out of it, and know what it is capable of doing. However, I have never deep dive into the details of setting things up, or the logic behind how the data is structured.</p><p>While my experience for this week is reduced to the first course of the module (<a href="https://cxl.com/institute/online-course/google-analytics-beginners/"><strong>Google Analytics for beginners</strong></a><strong>), </strong>that’s enough to say that Chris Mercer is a fantastic teacher: really easy to follow, tons of trips and tricks in all his lessons, all of them ready to be applied in your projects.</p><h3>Google Analytics: lessons and tips for newbies</h3><p>To better understand what do I mean by the learning experience I’ve had on this course, on this week post I’ll be sharing some of the notes I’ve taken from Chris’ lessons, and how immediate they can be if you are starting from 0 on GA.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*mmUwm1NyHyZnx-WZXr0yww.png" /><figcaption>Screenshot of <a href="https://support.google.com/analytics/answer/6367342?hl=en">Google Analytics demo account</a> | Google Merchandise Store</figcaption></figure><h4>On reporting</h4><ul><li>Real time reports answers the “what” of testing questions (whether something is well configured or not)</li><li>Audience reports answers the “who” of your site (the type of user that visits and purchases on your site)</li><li>Acquisition reports answers the “where” (the channel more successful attracting traffic into your site)</li></ul><h4>On setting up</h4><ul><li>There are 3 levels of configuration (from top to bottom): (1) Account level, (2) Property level, and (3) View level.</li><li>When defining the kind of information to be included or not at a Property level, be aware that Property 1 from Property 2 will be separate silos (their information won’t be mixed on Analytics). Consider that when separating or unifying different domains for later View.</li><li>Property names follow alphabetical order (create your own system to label them)</li><li>Enable users button: this will categorize your information based on users rather than sessions. As you know, sessions is any visit to the website (whether from the same person multiple times or from multiple persons just one time) whereas user’s is the tracking of the same person across different moments in time and devices.</li></ul><h4>On session settings</h4><ul><li><strong>Session timeout |</strong>When a person arrives to your site, it will be counted as a page view hit. If no other page view hit happens during that session, after the timeout the session will be closed. This is something important to considered based on your type of website and content within it, because if there’s a video or documentation that takes time to review, you may encounter multiple sessions within the same visit (as the person goes from one content to the other expiring your timeout). Of course this can also be solved with video tracking events (like click play, video milestones, etc.), but keep this session timeout concept in mind.</li><li><strong>Campaign timeout |</strong>This is how long are we going to grant with attribution any traffic source. As an example, when someone comes from Facebook because of a post we posted, should a visit 3 months later from the same user be attributed to that post or considered organic? That’s what is being defined here. As a rule of thumb, a couple of days is generally within a reasonable timeframe for hits of this kind.</li><li><strong>Referral exclusion</strong> | By this way you can avoid attribution from websites that are not really adding traffic from your site (i.e. traffic coming from PayPal, where in reality people leave your site to PayPal for payment and then comes back). To do this, it is only necessary to link the main URL form the referral we want to avoid.</li><li><strong>Search Term Exclusion</strong> |It would work similar to the referral exclusion, but applied to search engines and keywords. For those cases where people is clearly looking four our website but end up typing it on the search engine, we may want to not include this search as Google organic but rather direct traffic (as that was the intention).</li></ul><h4>On view settings</h4><ul><li>There are three minimum levels of View any Property should have: (1) Production view, (2) Testing view, and (3) Backup view.</li><li><strong>Production View </strong>| (or Master view) — the one that will answer most of your marketing questions and will use on a daily basis (tip: don’t forget to check the ‘Bot filtering’ box, and if you have site search don’t forget to activate its tracking with GA).</li><li><strong>Testing View </strong>| the one to be use when testing things, not for your answer your daily questions.</li><li><strong>Backup View </strong>| An emergency view just in case something goes off the rails so bad that you cannot have access from other views data. A configuration like this is also key to avoid situations where you may have filtered data in a way that you might be loosing strategic information. Without any backup view, that information will be lost forever and you won’t be able to look backwards and restore that info because it never existed.</li><li>First thing to do when creating view is to leave the automatic view created by GA on any property as it is, renaming it as BACKUP VIEW.</li><li>Once done that, it is time to start configuring the real production view with its goals, filters, etc. Notice that any view we create can be duplicated and all preferences from the copied view would be transferred into the copycat (goals, filters, etc.). However, data on any view will start to flow in once the view has been created. It won’t display previous information to the date the view was created. That’s why is so important to create views early on.</li><li>Before moving any view into the trash, change its name with a label that signals everyone within the team/company that it is there on purpose (you may avoid one or two heart attacks by doing it).</li></ul><h4>On filtering setting</h4><ul><li>Filters is the most powerful, yet the most dangerous tool within analytics. This is so because it permanently alters the data we are going to visualize (and that’s why is so important to have a backup view).</li><li>The reason why filters are created (although we might be missing some data on the view) is to make that specific data being visualized easier to read, cleaner to present for that specific goal, and also helping to fix any fractured reporting.</li><li>Filters can be created at View level of Account level. We should consider two things about creating filters at View level: (1) they are only available for that specific view, and (2) we might fall into duplication when the same filter is created in two or more views. In the case of the generic filters, which would be subject to apply to any view we create afterwards, it would be far more interesting to create them at Account level. By doing so we would avoid filter management chaos, while keeping the power to decide to which view the filter may apply by including or removing it from any view available.</li></ul><h3>Some generic filters worth creating early on</h3><p>Chris Mercer shares with us three generic filters worth to be considered to include on any GA set up to make our lives easier:</p><h4>To make data easier to read</h4><p>One proposal we get here is to create a view that allows us to show the domain of the different pages we may have. To do so we would have to configure it under the view we want to filter and then:</p><blockquote>Filter / Add Filter -&gt; Name filter (e.g. “Show domain”) / Filter type -&gt; Custom/Advance:</blockquote><blockquote>* Field A to “Hostename” = (.*)</blockquote><blockquote>* Field B to “Request URL” = (.*)</blockquote><blockquote>* Output to “Request URL” = $A1$B1</blockquote><p>As long as the “Override output field” box is checked, the result will get is the complete URL of the website, and not just what comes after the “.com/“ of any URL. Something very useful if you have created a Property that connects different hostnames (e.g. <a href="https://www.yourwebname.com/">https://www.yourwebname.com/</a>).</p><h4>To clean data</h4><p>In order to confirm that the data flowing is useful and not spam traffic, another filter we might want to create is a hostname filter. What it will do is to just include the specific domains we defined, avoiding copycats site’s information. To do so we would have to go to:</p><blockquote>Filter / Add Filter -&gt; Name filter (e.g. Include NameWeb Hostname) / Filter type -&gt; Custom -&gt; Include -&gt; Hostname / Filter pattern -&gt; <a href="http://yourwebname.com/">yourwebname.com</a></blockquote><p>We can define or not what comes after the name (.com / .net /.app) depending on the filtering level we want to apply. If we have more than one version, by not specifying it, GA will cover all of them by just including the name (i.e. “yourwebname” would cover the .com, the .net, and another we own in which we have included our tracking code).</p><p>In order to add more hostnames to this filter (as may happen with checkout services, etc.) we would have to include on the “Filter pattern” the symbol “ | “ which means “OR”. This way GA will include any data of the hostnames defined there (excluding any third party copycat).</p><h4>To fix fractured reporting (1)</h4><p>We may encounter on our GA that the same page is reported duplicated in two different pages whose only difference is whether a slash ( / ) at the end of the URL was included or not. To avoid this situation, we would have to create a filter that merges both into just one outcome. To do so we would have to go to:</p><blockquote>Filter / Add Filter -&gt; Name filter (e.g. Add Slash) / Filter type -&gt; Custom/Advance:</blockquote><blockquote>*Field A to “Hostename” = ^(/[a-zA-Z0–9/_\-]*[^/])$ (it basically finds any URL that doesn’t end with a slash /)</blockquote><blockquote>*Field B to “Request URL” = (this stays blank)</blockquote><blockquote>*Output to “Request URL” = $A1/ (this way will include a slash to any URL without it)</blockquote><h4>To fix fractured reporting (2)</h4><p>Another filter we may initially include is “Lowercase source”. Since GA is case sensitive, applying a filter of this kind will avoid that traffic form the same channel gets split into multiple different sources (i.e. Facebook, and facebook would appear as different sources without this filter). To do we have to simply:</p><blockquote>Filter / Add Filter -&gt; Name filter (i.e. Lowercase — Source) / Filter type -&gt; Custom/Lowercase -&gt; Filter field: Campaign Source</blockquote><p>Similar case would apply for search terms, as people may use or not an uppercase when writing whatever term they might be looking for. To clean such data and avoid duplications for the same term, we would apply a similar filter as before. In this case:</p><blockquote>Filter / Add Filter -&gt; Name filter (i.e. Lowercase — Search) / Filter type -&gt; Custom/Lowercase -&gt; Filter field: Search Term</blockquote><p>This use of lowercase filter might be also applied to other areas within GA such as Medium, Content, Campaign, or Site search.</p><h3>In a nutshell</h3><p>Chris Mercers makes a great job introducing any newbie to Google Analytics (GA) with a highly tactical approach to real world situations and practical tips to be implemented right away on any project. Very instructive and useful lessons learnt this week.</p><p><em>Stay tuned for week 5 post.<br>// If you want to review the previous week post, </em><a href="https://ruben-jimenez-baz.medium.com/cxl-institute-growth-marketing-minidegree-review-week-3-cdb01dc39236"><em>click here</em></a><em> //</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8d99995417d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[CXL Institute — Growth Marketing Minidegree Review | Week 3]]></title>
            <link>https://ruben-jimenez-baz.medium.com/cxl-institute-growth-marketing-minidegree-review-week-3-cdb01dc39236?source=rss-193d6aad3437------2</link>
            <guid isPermaLink="false">https://medium.com/p/cdb01dc39236</guid>
            <dc:creator><![CDATA[Rubén Jiménez Baz]]></dc:creator>
            <pubDate>Sun, 07 Mar 2021 20:44:50 GMT</pubDate>
            <atom:updated>2021-03-07T20:44:50.972Z</atom:updated>
            <content:encoded><![CDATA[<h3>CXL Institute — Growth Marketing Minidegree Review | Week 3</h3><p>Second part of Module 2 — Running Growth Experiments, it’s covered by Ton Wesseling on its <a href="https://cxl.com/institute/online-course/ab-testing-mastery/">A/B Testing Mastery Course</a>. On it he amplifies the concepts covered on the first part of this second module of the <a href="https://cxl.com/institute/programs/growth-marketing-training/">Growth Marketing Minidegree</a>, deep diving on the different steps of the A/B testing phases, from planning to execution and results.</p><p>For this reason, I would focus this time on two elements that complements my <a href="https://ruben-jimenez-baz.medium.com/cxl-institute-growth-marketing-minidegree-review-week-2-57d55296b2a4">previous post</a> of this module: KPI definition, and 6V Conversion Canvas.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*m3blCkOCSG5cCs3Q" /><figcaption>Photo by <a href="https://unsplash.com/@srd844?utm_source=medium&amp;utm_medium=referral">Stephen Dawson</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><h3><strong>Which KPI to pick?</strong></h3><p>KPI stands for Key Performance Indicator. It is a metric that defines the goal to achieve by a business or team. Although there exist many different types of KPIs, attending to its nature on optimization these could be grouped as follow, from the least impact/relevancy to the most relevant/desired outcomes.</p><p><strong>Clicks</strong> | Increasing its number within a website/page. A part from being very easy to manipulate (to meet the bar), its outcome won’t go further than that and into more business impactful goals.</p><p><strong>Behavior</strong> | Achieve a change in the way customers relates to your business site. This is the kind of metrics that starts to look for a longer-term outcome.</p><p><strong>Transaction</strong> | Impact on the revenue of the business. This is the kind of KPI that any business should aim to defined and start caring, as it is the one that drives any business success.</p><p><strong>Revenue Per User</strong> | increasing the average purchase order per user. As level of maturity of the business on KPIs increases, their focus should shift from the previous to this one to increase the impact of every conversion on the site (if the acquisition cost stays the same, this could drive the business income to the moon).</p><p><strong>Potential Lifetime Value</strong> | creating longer relationship with customers. While the previous two relates to more short-term outcomes, at this level of maturity of the company the goal growths from today’s into tomorrow and years from now. Even though the actions driving these metrics may impact negatively on short-term KPIs (i.e. transactions) if repetition increases, in the long-run the result might be beneficial (think of Amazon and its customer experience as driver of growth at the expense of returning costs).</p><p>KPIs should be defined in a binary way (0/1), meaning the goal is achieved or not. When the outcome to measured is not binary (i.e. average purchase order, # pages viewed per session, etc.), these have to be converted to meet the criteria. For example, if the # pages viewed per session is 3, either the experiment increases this result or it doesn’t.</p><p>However, a success result for a KPI within an optimization experiment may impact negatively on a different KPI. This type of conflict is very common within businesses whose teams have different goal metrics for their teams. That’s why North Star Metrics are key for any business. A shared-clear common goal for all teams across the organization helps to later defined at team level the short-term indicators towards that long-term value success indicator.</p><p>When the business success cannot be defined in one metric, an alternative can be based on a weight combination of a group of them. Defined that way, any optimization experiment can be binary measured as successful or not taking into consideration all its different impact outcomes (positive, neutral, and negative) across the business. This way negative impact on low weight KPIs may not invalidate experiments whose outcome is higher on more relevant KPIs for the business.</p><h3><strong>The 6V Conversion Canvas</strong></h3><p>An alternative to the previous ResearchXL presented by Peep, Ton walks us through his 6V Conversion Canvas for desk research phase, as first step prior to any experimentation. It consists on a six part canvas:</p><ul><li><strong>Value |</strong> What company values are important and relevant? What focus delivers most business impact? What are the short and long term goals?</li><li><strong>Versus |</strong> Who are the competitors? What competitor analysis and market best practices can be found? What changes have they made on their website? (we can use services like <a href="https://visualping.io/">Visualping</a> to tack our competitors changes)</li><li><strong>View</strong> | What insights can be found from web analytics and web behavior data? Where people come from? What are their flows? How it differs from mobile to desktop? Based on users flow behavior within our website, we can segment them into groups to better analyze what are the steps that drive some deeper in the funnel while preventing others from continuing: Ton presents to us a classification for this based on the following layers:<br><em>(1) Users with some interaction (stayed longer than bouncing users)<br>(2) Users with heavy interaction (source multiple pages)<br>(3) Users with clear intention to buy (added products to cart)<br>(4) Users willing to buy (proceed to checkout)<br>(5) Users that succeed in buying (successful users)<br>(6) Users that return with intent to buy more (loyal users)</em><br>A report on this part would cover from broader information such as # of users per segments, conversion and time to move from segment to segment, to more in detail aspects like what are the most important pages of the website (where decisions are made) and the specific behavior within those pages (i.e. drop outs)</li><li><strong>Validated</strong> | What insights are validated in previous experiments or analyses? Most likely we might start with previous knowledge on success and failing from past experiences.</li><li><strong>Verified</strong> | What scientific research, insights, and models are available? Digging on research literature can be of great value to gather insights that may apply to your case.</li><li><strong>Voice</strong> | What insights can be taken from voice of customer data such as surveys, feedback and service contact? Social media can be as well.</li></ul><p>Lastly, when formulating experiments hypotheses from researched information, it’s important to formulate them so they are concrete:</p><p><em>“We believe that (solution to be applied), then (desired behavioral change / outcome) will happen [among (specific target group)] because of (problem / reason behind this situation)”</em></p><p>Using a frame like this allows to condense in just one sentence the problem, the solution and the outcome expected out of the experiment.</p><h3>In a nutshell</h3><p>Ton deeps dive on data and calculations needed before, during and after any A/B testing experiment in a very lecturing way. Great learnings on how to prepare the ground for testing and later read the results of the potential business outcomes.</p><p><em>Stay tuned for week 4 post.<br>// If you want to review the previous week post, </em><a href="https://ruben-jimenez-baz.medium.com/cxl-institute-growth-marketing-minidegree-review-week-2-57d55296b2a4"><em>click here</em></a><em> //</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cdb01dc39236" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[CXL Institute — Growth Marketing Minidegree Review | Week 2]]></title>
            <link>https://ruben-jimenez-baz.medium.com/cxl-institute-growth-marketing-minidegree-review-week-2-57d55296b2a4?source=rss-193d6aad3437------2</link>
            <guid isPermaLink="false">https://medium.com/p/57d55296b2a4</guid>
            <dc:creator><![CDATA[Rubén Jiménez Baz]]></dc:creator>
            <pubDate>Sun, 28 Feb 2021 21:10:04 GMT</pubDate>
            <atom:updated>2021-02-28T21:10:04.042Z</atom:updated>
            <content:encoded><![CDATA[<h3>CXL Institute — Growth Marketing Minidegree Review | Week 2</h3><p>You don’t know what you don’t know, and you can’t improve what is not measured. That’s the reason why research and testing are the most powerful elements against the unknown, personal preferences and biases.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*sFGV6NfxG-lsW4nx" /><figcaption>Photo by <a href="https://unsplash.com/@epicantus?utm_source=medium&amp;utm_medium=referral">Daria Nepriakhina</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Research are testing are two side of the same coin when working on optimization. It is a systematic process of finding opportunities for growth and developing data-based ideas for how to build upon those opportunities. Such process is well implemented when (1) teams know what idea should be tested next, and (2) how that idea is going to be tested. If the answer to the second statement is unknown, the process is broken. If the answer to the first one is unknown, you don’t have a process at all.</p><p>A good optimization process should tell you:</p><ol><li>Where are the problems (in fact, where is the biggest problem in the funnel to start with)</li><li>What are the problems (how do they look like)</li></ol><p>3. Why are they a problem (why do they come up)</p><p>4. How those problems are turned into hypotheses that can be tested</p><p>5. Which ideas or solutions should be prioritize.</p><p>But before that, you need to start with research. You have to have a complete understanding of what matters to your users: what do they care about, which problem are they truly trying to solve, how is their customer journey from start to end. You need to know that before starting any optimization. Otherwise, you will be randomly testing in the fog.</p><p>Come up with a list of questions and see what the data tells you about it. Challenge your assumptions. Make sense out of the mess. Additionally you can complement this process through a qualitative analysis that validates some of the findings and assumptions.</p><h3>ResearchXL</h3><p>The research process followed by CXL is called ResearchXL and it is based on a 6 step analysis to gather the data needed to ensure the best possible ideas are the ones first tested (optimizing the process, and achieving early results).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dOmrAvJubH6SGmZ2lv5dQw.png" /><figcaption>ResearchXL representation by CXL</figcaption></figure><h4>1. Technical Analysis</h4><p>As obvious as it might seem, first thing is to make sure that everything runs well, discarding any problem with browsers and devices compatibility. Although trying to break it by simulating being a customer on all different scenarios is an option, by comparing performance metrics between browsers and devices you may get a quicker idea of where to start looking at.</p><p>A hot topic is loading speed. Our patience has decreased over the years and now we can see tremendous bouncing rates on websites taking longer than 2 seconds to load.</p><h4>2. Heuristics Analysis</h4><p>Once the technical part has been addressed, it is time to critically review what’s in front of us, asking ourselves questions like: is the information presented <strong>relevant</strong> for the customer? Is the message <strong>clear</strong> enough? What kinds of <strong>barriers</strong> are in between the user and his goal? What <strong>frictions</strong> do the customer encounter during the process?</p><p>In this process of assessing the pages or screens from a website or app, we should always be thinking: (1) what’s its purpose, and (2) what action do we want to trigger with it. Basically it should be all about moving the customer one more step through the funnel. The Fogg Behaviour Model can be of help here when defining the level of the barrier that we’ll be presenting to the user. Depending on their level of ability or motivation in performing that action, our user might have a hard time and conversion rates might fall dramatically.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jPc7FmS-m0jRQEjiYBe2rw.jpeg" /><figcaption><a href="https://behaviormodel.org/">Fogg Behaviour Model</a></figcaption></figure><p>Of course what we might find here might not the whole truth, and things that may annoy us don’t affect customers that much during their journey. All this information serves as a starting point to later contrast it with data.</p><h4>3. Digital Analysis</h4><p>Once we have figured out what might be wrong with every screen and page, it&#39;s time to look at the data and see what it says about our findings. In detail, we want to know: (1) where are the biggest leaks of the journey taking place, (2) in which segment are these happening (device, traffic channel, etc.), (3) what is the behavior of users on the web/app, and lastly (4) which actions correlate with higher conversions.</p><p>However one of the main problems we might have to do this analysis is the lack of data. Any website should be tracking every single action any user performs. Every click, scrolling, interaction with components, should be recorded in analytics. Otherwise there is no way to know where and when things are not happening as intended.</p><p>Data is the only way to know to know what actions are triggering the desired behavior. Since this process of optimization is all about constant experimentation, the only way to know when something has changed is by comparing data from previous and current performance. The more data points we have, the more precise we will be. Specially because is not unusual to improve one part of the funnel while negatively affecting another. But we will only be aware of that if we have multiple metrics to compare against them.</p><h4>4. Mouse Tracking and Form Analysis</h4><p>At this point, we would have a nice idea of where problems might be, but not exactly how people deals with each situation. A way to unveil this is by looking how do they behave analyzing the clicks and the scrolling on the pages, as well as their reactions to the different problems they encounter.</p><p>Heat mapping is something you can use to have a visual representation of that clicking and scrolling data (now there’s even AI services that can forecast heat maps of any website or app before publishing them). However, a greater and more vivid alternative are the services that generate anonymized screen recordings of user sessions. That way we can see exactly what is that thing people struggle with on each page.</p><p>This applies as well to forms. You should analyze your form and measure which fields are causing mayor dropping and see if there’s any way you can get rid of that field or ask for it at a different moment in time (remember the Fogg Behaviour Model). We all have experience with forms that asks for too much information (at least more than we are ready to give away at that point in the relationship).</p><p>This happens because companies are only thinking about extracting as much data as possible from us, without considering that it might backfire them. Rule of thumb would be to only ask for the necessary information at each step of the process (i.e. the address at the moment of purchasing, but not at registration) to reduce the friction.</p><h4>5. Qualitative Surveys</h4><p>So far our research has been around quantitative data analysis. However, data only tells you what, where, and how something is happening, but we are missing the why. Qualitative analysis is all about completing the picture by providing human context, voice, and the reasoning thinking behind.</p><p>The best way to gather this data is from the people who have recently gone through the process you are looking at (whether is signing up, purchased something, left the cart with stuff on it). By surveying them (opened questions) or reaching them by phone, you might gather high quality information of what is preventing or incentivizing your users to take action.</p><h4>6. User Testing</h4><p>Lastly we can recruit people to tell us out loud their reasoning during the processes of tasks that we want to explore, so they can highlight us what difficulties have they experience during the process (and most likely other users would experience as well. Yet keep in mind these users, as they are economically incentivized, might not be as cruel as customers would behave. So always trust more what they do rather to what they say.</p><h3>PXL Prioritization Framework by CXL</h3><p>Once done this process, we would have our own custom list of problems (and not an infinite list of hacks that you googled). This research step is the most important part in the optimization process. It sets the foundations and conditions everything that comes after.</p><p>Having our list of problems defined, we would then build the hypotheses behind each of them. The better the hypothesis, the better the results (and better hypothesis comes from better insights from the research).</p><p>Once defined the hypotheses, it is time to design the experiments/solutions that may be tested within the website/app to see if they improve the results on the existing problems.</p><p>Testing phase follows the scientific method of comparing results on behavior between the current state and the proposed solution (A/B testing). During a previously defined amount of time, both options are displayed equally to different groups of users to measure, which one performs better.</p><p>After the test concludes, all the data gathered is analyzed and compared to extract learnings and conclusions. If done correctly, at this stage we should have a better understanding of the situation, leading to new hypothesis for improvement. Then process goes over and over again.</p><p>As not all problems are of the same nature, require the same level of effort to be tested, or simply they don’t even need to be tested. This is the case for problems where data could not be gathered because there were no tags implemented (‘Instrument’ category), or when things are too obvious and it’s simply a matter of fixing it — like missing a purchasing button on a page (‘Just do it’ category).</p><p>Most problems will fall within the ‘Test/Hypothesize’ category as we may now what the problem is but not what is the best solution to fix it. It can also be the case that we don’t know whether something is really a problem or not for the users, and need further investigation on it, which would be the last category.</p><h4>Ranking Tests</h4><p>At this point all our problems would have a category, context, and action to be performed/tested. However, as not all actions can be implemented all at once we need to rank them. Commonly known prioritization frameworks are ICE (Impact, Confidence, Ease) and PIE (Potential, Importance, Ease), which gives a score from 1 to 10 to each element of the problem, and based on the result, they get ranked.</p><p>Although some people have defined different frames to score these elements and avoid as much as possible subjectivity on the decision (<a href="https://medium.com/@nimay/inside-product-introduction-to-feature-priority-using-ice-impact-confidence-ease-and-gist-5180434e5b15">see Nimay Parekh article</a>), Peep Laja shares the alternative used in CXL: the PXL prioritization framework. Its proposal to remove the subjectivity as much as possible comes from rating binary (mainly, 1 or 0) the following questions about the test hypothesis to be ranked:</p><p>· Is it above the fold?</p><p>· Noticeable within 5 sec? (2 or 0)</p><p>· Adding or removing an element? (2 or 0)</p><p>· Designed to increase user motivation?</p><p>· Running on high traffic page(s)?</p><p>· Addressing an issue discovered via qualitative feedback (surveys, polls, interviews)?</p><p>· Addressing insights found via digital analytics?</p><p>· Supported by mouse tracking heat maps or eye tracking?</p><p>· Ease of implementation (&lt; 4h = 3 | &lt; 8h = 2 | &lt; 2 days = 1 | More = 0)</p><p>Regardless of the method, the key takeaway here is that there must exist a consistent way of ranking the hypotheses. Additionally, we could further cluster the test hypotheses based on the stage of funnel being affected. Since our focus should be one stage at a time when improving any website or app, such category makes it easier to find the right tests.</p><h4>How long tests should be run?</h4><p>Test should be run until enough evidence (positive or negative) is reached. That is achieved based on 3 criteria:</p><p>· Enough sample size</p><p>· Multiple business cycles</p><p>· Statistical significance reached</p><p>Calculating the sample size is an easy thing and there exist multiple calculation tools for this, but it is the first thing to do. It will let us know if it makes sense at all to run any test in our case, based on whether we can achieve enough sample size or not. If we don’t meet the bar, the best we can do is to implement all the best possible solution for each of the problems we have.</p><p>Regarding business cycles, well basically is about being sure you have completed all your different scenarios within the defined timeframe. As a takeaway, 4 weeks would work for most websites/apps (never go beyond that timeframe), and a mark check at 2 weeks would be ideal (it is very unlikely that the way customers behave on Tuesday is equal to how they do on Sunday; or at the beginning and end of the month due to paycheck).</p><p>Lastly, some metrics anyone should be tracking to validate that the team is operating on an efficient testing optimization process are:</p><p>· Number of variants tested</p><p>· Win rate</p><p>· Average uplift per successful experiment</p><h3>In a nutshell</h3><p>Defining a well optimization process, and follow it methodologically and rigorously is key to the success of any business. In some cases, you may not have enough data to give statistically significant results on experiments. However good research, and good analysis of current data is at hand of everyone. Let’s be data driven in our decisions.</p><p><em>Stay tuned for week 3 post. <br>// If you want to review the previous week post, </em><a href="https://ruben-jimenez-baz.medium.com/cxl-institute-growth-marketing-minidegree-review-week-1-e0c02947928c"><em>click here</em></a><em> //</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=57d55296b2a4" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[CXL Institute — Growth Marketing Minidegree Review | Week 1]]></title>
            <link>https://ruben-jimenez-baz.medium.com/cxl-institute-growth-marketing-minidegree-review-week-1-e0c02947928c?source=rss-193d6aad3437------2</link>
            <guid isPermaLink="false">https://medium.com/p/e0c02947928c</guid>
            <dc:creator><![CDATA[Rubén Jiménez Baz]]></dc:creator>
            <pubDate>Sun, 21 Feb 2021 22:36:59 GMT</pubDate>
            <atom:updated>2021-02-22T09:05:03.935Z</atom:updated>
            <content:encoded><![CDATA[<h3><strong>CXL Institute — Growth Marketing Minidegree Review | Week 1</strong></h3><p>In a crowded market of well-defined solutions, just making noise is no longer enough to earn the attention of potential customers. Alternative paths must then be found.</p><p><strong><em>Full disclosure:</em></strong><em> over the next weeks I’ll share an in-depth review of my learning experience on the Growth Marketing Minidegree by CXL Institute in a series of posts. If you’re interested in knowing more about it, stay tuned to my Medium page.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*M-EOIUAQc4zaXY8B" /><figcaption>Photo by <a href="https://unsplash.com/@austinchan?utm_source=medium&amp;utm_medium=referral">Austin Chan</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>As of 2021, it’s highly possible that you’ve encounter the term “Growth Hacking” or “Growth Marketing” multiple times in posts, videos, or even have read some of the books covering these hyped topics. Well know are the examples of (at their time) startups like Dropbox, Airbnb or PayPal on how they’ve made their way through the market with unique actions. Those actions led to rates of adoption of their products that call the attention of companies and marketers alike.</p><p>Although both terms — Growth Hacking and Growth Marketing — are often used indistinctly, they are not exactly the same thing. Truth is they share similarities on how growth is driven, results are measured, and a constant improvement approach is applied. That’s why an increasing adoption of Growth Marketing is taking place within more established companies, as it is more aligned with their scale and brand value. But, what is it about exactly and how is it applied?</p><h4><strong>Welcome to the Growth Marketing Minidegree</strong></h4><p>To answer these questions, CXL Institute has designed the Growth Marketing Minidegree, an online self-paced course that aims to cover the main methodologies and processes used by top notch growth marketers. By the end of the +110 hours of material structured along 7 modules, you would be better equipped to accelerate your business through a new growth mindset.</p><h4><strong>What you’ll learn in Week 1</strong></h4><p>First course starts with an introductory track on Growth Marketing foundations, covered on four lessons: (1) <em>‘Growth mindset: growth vs traditional marketing’</em>, (2) <em>’Building a growth process’</em>, (3) <em>’User-centric marketing’</em>, and (4) <em>’Identifying and amplifying growth channels’</em>.</p><h4><strong>Growth Mindset: growth vs. traditional marketing</strong></h4><p>John McBride begins by explaining how growth marketing differs from traditional brand marketing. This can me summarized in two key elements:</p><ol><li>Focus. While brand marketing is focused on the top of the funnel (awareness and acquisition) , growth marketing covers it entirely to the bottom (activation, retention, revenue, and referral) looking for opportunities on all of its levels.</li><li>Experimentation. Instead of one-shot actions in hope of nailing the message, Growth marketing is driven by constant experimentation and learning to improve the interactions with customers all along the funnel.</li></ol><p>This has a lot in common with the Lean Startup methodology, where the process starts by challenging the hypotheses and biases around what we think customers want, to help us fine tune the products/proposals based on the learnings and understanding gathered in the process. Growth Marketing is not a one silver bullet action mindset, but rather a process of continuous experimentation over a series of actions that compile together into a greater outcome.</p><h4><strong>Building a Growth Process</strong></h4><p>John McBride continues its lecture by presenting the different stages any team faces when building out a growth process. As you can imagine, it starts by mapping out where the company stands along its customer journey. Then it is time to decide what quarterly goals are going to be covered to start planning in more detail the specifics of the execution roadmap. Having that in place, it will be a matter of start building experiments, analyze results, adjust them, and so on and so forth.</p><p>One take away about this is the importance of focusing on one step of the funnel each time, rather than tackling all stages at once. This has to be with how learnings on each stage compiles together from experiment to experiment, allowing for a greater speed of experimentation. In this process it is key that each stage has a specific success metric on which to compare the results of the experiment.</p><p>As experiments will have different size, impact, and complexity, it is important that any team has a clear prioritization criteria. A popular framework is the ICE framework, where the “I” stands for impact, “C” for confidence, and “E” for effort. By applying a score to each element, the experiments are prioritized based on a (somewhat) objective criteria.</p><h4><strong>User-centric Marketing</strong></h4><p>Paul Boag covers how the digital ecosystem has change the way people and businesses relates and interact with each other, and how that has impact on the marketing funnel. On this scenario the strategy should be focused on people, rather than products, and how to make them happier customers. In the process we should always keep in mind what is the job-to-be-done of each person at each stage, so every interaction with them delivers the greatest value (instead of annoying them).</p><p>Although there’s already tons of data within companies to start with for a better understanding of their customers at low cost, research does not have to be expensive. With a handful of interviews (as low as 6) and well designed activities, any company could rapidly gather enough learning to pivot proposals before going into a broader panel. Most importantly, this process should continue and be monitored after the campaign has gone alive.</p><p>This is one of the key messages for marketers. A change in their mindset form jumping from one campaign to another, in favour of adjusting each as they run based on performance. Doing so allows teams to gather learnings along the whole process instead of just at the end of it, increasing the advantage and knowledge of their customer base.</p><h4><strong>Identifying and amplifying growth channels</strong></h4><p>Lastly, Sophia Eng leads a course on tactical approach to growth channels (SEO, social and display ads, email marketing, content marketing…) by providing useful takeaways. From how to choose the right channel to focus on, or how to track each key performance indicator (KPI), one learns how the energy should be distributed at each stage of the process instead of trying to optimize all channels at once.</p><p>Myths surrounding SEO or the well known vanity metrics on social media, are other insightful takeaways from this lesson as a recap for any digital marketer taking this course.</p><h4><strong>In a nutshell</strong></h4><p>A handful of interesting takeaways for the 1st week related to the Growth Marketing mindset, and its possibilities, while ensuring any gap on the basics of digital marketing is covered so anyone can be ready for more advanced materials awaiting on the coming weeks.</p><p><em>(stay tuned for week 2 post)</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e0c02947928c" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>