<?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 Elena Sigacheva on Medium]]></title>
        <description><![CDATA[Stories by Elena Sigacheva on Medium]]></description>
        <link>https://medium.com/@elenasigacheva?source=rss-73aeda649930------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*fvEyxMaJxVSU-ItC.</url>
            <title>Stories by Elena Sigacheva on Medium</title>
            <link>https://medium.com/@elenasigacheva?source=rss-73aeda649930------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 22:57:22 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@elenasigacheva/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[Too Fast, Too Many, Too Much]]></title>
            <link>https://medium.com/entelo/too-fast-too-many-too-much-three-common-mistakes-when-making-data-driven-decisions-ed0251dcea27?source=rss-73aeda649930------2</link>
            <guid isPermaLink="false">https://medium.com/p/ed0251dcea27</guid>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[big-data]]></category>
            <category><![CDATA[metrics]]></category>
            <category><![CDATA[analytics]]></category>
            <category><![CDATA[data-driven-decisions]]></category>
            <dc:creator><![CDATA[Elena Sigacheva]]></dc:creator>
            <pubDate>Mon, 14 Nov 2016 21:37:32 GMT</pubDate>
            <atom:updated>2016-11-16T17:54:58.034Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*HgHp6WW-Qx7DX2_afIvdTQ.jpeg" /></figure><h3>Three Common Mistakes When Making Data-Driven Decisions</h3><p>Living in the era of big data, thousands of companies adopt a data-driven approach to making decisions. In this blog post I’d like to talk about three common mistakes we often make when working with data.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*cui9pXpH5s_KKJyT3Wdfmw.jpeg" /></figure><h3>Too Fast: Jumping to a “Solution” Before Identifying the Problem</h3><p>Don’t fear spending extra time on defining the problem. The assumption that the sooner you start the project, the sooner it will be finished is not necessarily true and may cost you many hours of work if you don’t have a good understanding of the problem being solved. Regardless of whether you’re considering a new product feature, making a strategic decision regarding your company’s direction, building a machine learning model, or just writing an email to your coworker, make sure to spend extra time understating the need behind your project to avoid situations where after spending hours of hard work you realize that you wasted time solving the wrong problem.</p><p>Don’t forget that the approach to solving the problem depends directly on the type of problem you are trying to solve. For example, in the data science world, before building any model, it’s crucial to understand whether the goal of the model is to increase recall (how well your model returns specific data) or precision (out of all positive predictions, how many were correct). Understanding the goal of the project from the beginning will allow you to train your model correctly — optimizing for avoiding false positives (in case of precision) or false negatives (in case of recall).</p><p><a href="https://hbr.org/2012/09/are-you-solving-the-right-problem"><strong><em>This</em></strong></a> <em>Harvard Business Review</em> article recommends the following four steps to help you to solve the right problem:</p><p>Step 1: Establish the Need for a Solution</p><p>Step 2: Justify the Need</p><p>Step 3: Contextualize the Problem</p><p>Step 4: Write the Problem Statement</p><p>Another famous approach that is taught in MBA programs is called <a href="https://en.wikipedia.org/wiki/5_Whys"><strong><em>Five Whys</em></strong></a>, a problem-solving technique developed at Toyota that helps to get to the root cause of a problem by repeating the question “Why?” multiple times.</p><h3>Too Many: Not Distinguishing Between Important and Unimportant Metrics</h3><p>Living in the word of big data with a variety of tools available for tracking all types of user activity, you can have metrics for every user action. But do all these metrics really matter? Can you draw an actionable conclusion by looking at all these metrics?</p><p>It’s quality, not quantity that matters. Instead of tracking all possible kinds of data, try to identify KPIs (key performance indicators) that truly allow for measuring product health and improvement over time. Depending on the business model, company’s goals, and stage of the product, you can decide what kind of KPIs you need to concentrate on.</p><p>I recommend checking out Dave McClure’s <a href="http://www.slideshare.net/dmc500hats/startup-metrics-for-pirates-long-version/2-Customer_Lifecycle_5_Steps_to"><strong><em>presentation</em></strong></a> where he breaks down metrics for startups into five segments:</p><p>· Customer Acquisition</p><p>· Activation</p><p>· Retention</p><p>· Referral</p><p>· Revenue</p><p>Another well-known minimalist approach is described in the book <em>Lean Analytics</em> that talks about the idea of selecting “One Metric That Matters.”(And check out <a href="https://growthhackers.com/videos/video-lean-analytics-one-metric-that-matters-omtm/"><strong><em>this</em></strong></a> workshop from Google Ventures on the Growth Hackers website.)</p><h3>Too Much: Trying to Find Patterns When They Don’t Exist</h3><p>Although we often seek to make data-driven decisions because our intuition isn’t always right, and sometimes the data is surprising, it’s also important to apply common sense. Remember that the model of the world isn’t the world itself. There are assumptions in your model, and limitations, and sometimes these can lead you astray. Regardless of what the model is telling you, whether it is confirming your expectations or telling your something surprising, always apply your own judgment and think probabilistically about the odds of the forecast being true and how strong the signal is.</p><p>Sometimes there isn’t a “truth” to be found at all. We may not realize that sometimes we are unintentionally seeking a signal when it doesn’t exist. Understanding broader aspects of the industry, competitive landscape, factors of macroeconomic conditions, seasonality effects and other factors will allow you to set more accurate growth goals and derive intelligent conclusions from your data. For example, a decrease in certain engagement metrics may not necessarily be interpreted as a negative signal but can be just a seasonality effect or an increase in the number of customer complaints can be a healthy sign of growth given that the user base increased as well.</p><p>“Correlation doesn’t imply causation” is a phrase commonly used in statistics to emphasize that a correlation between two variables does not mean that one factor causes another. Sometimes reading a blog post is the reason people improve their data-driven decision making, and sometimes <a href="http://tylervigen.com/spurious-correlations"><strong><em>stuff just happens</em></strong></a>. :-)</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ed0251dcea27" width="1" height="1" alt=""><hr><p><a href="https://medium.com/entelo/too-fast-too-many-too-much-three-common-mistakes-when-making-data-driven-decisions-ed0251dcea27">Too Fast, Too Many, Too Much</a> was originally published in <a href="https://medium.com/entelo">Entelo Engineering</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>