<?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 Reinaldo Camargo on Medium]]></title>
        <description><![CDATA[Stories by Reinaldo Camargo on Medium]]></description>
        <link>https://medium.com/@reinaldocamargo?source=rss-f02cac568ab1------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*dLnjUtlTTXvakuR5lFXy_g.jpeg</url>
            <title>Stories by Reinaldo Camargo on Medium</title>
            <link>https://medium.com/@reinaldocamargo?source=rss-f02cac568ab1------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 02:24:28 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@reinaldocamargo/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[Analytic Retrospectives]]></title>
            <link>https://medium.com/@reinaldocamargo/analytic-retrospectives-bd822d6470ee?source=rss-f02cac568ab1------2</link>
            <guid isPermaLink="false">https://medium.com/p/bd822d6470ee</guid>
            <category><![CDATA[continuous-improvement]]></category>
            <category><![CDATA[retrospectives]]></category>
            <category><![CDATA[agile]]></category>
            <category><![CDATA[metrics]]></category>
            <dc:creator><![CDATA[Reinaldo Camargo]]></dc:creator>
            <pubDate>Tue, 30 Apr 2019 15:08:11 GMT</pubDate>
            <atom:updated>2019-04-30T19:01:16.152Z</atom:updated>
            <content:encoded><![CDATA[<p>A retrospective model focused in metrics, charts and flow</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/919/1*2_KSej0AeME10nnNW9ck_Q.png" /></figure><p>In Agility, we talk a lot about making experiments, scientific method, PDCA cycle and so on to achieve better results or solve problems. However, rarely I see scientific method been used on Retrospective meetings or any meeting where the goal is continuous improvement. Let me explain what I mean: most of the retrospective activities are based on perceptions and they are related with &quot;soft-skills&quot;. Rarely I saw someone (Scrum Master, Agile Coach, Product Owner, Project Manager or whatever) bringing data about some experiment that they tried a month ago (for instance). How do you tell that some action worked or not? Perception? Some person’s point of view?</p><p><strong>The problem</strong></p><p>Retrospectives meetings, in general, are based in people&#39;s perceptions and perception is deceiving. It depends on people&#39;s context/point of views not necessarily the reality.</p><p>Sometimes the techniques used in a Retrospective opens too much space for complaining and the meeting’s goal is missed: How do we improve our process to delivery more with better quality?</p><p>More vague are the words/sentences used in a technique more difficult is to find the real problems. When you leave room with vague sentences like &quot;Improvements points&quot; things like &quot;the coffee was horrible&quot;, &quot;My chair broke&quot;, &quot;The PO/SM was not available 24/7&quot; and a lot of useless excuses can emerge from it.</p><p><strong>Analytic Retrospectives</strong></p><p>I’ve been running a different approach of Retrospectives based on metrics. Those metrics are impacted by the day-to-day actions of a team or events.</p><p>When I say metrics I’m not talking about throughput or if the sprint was delivery or not. I’m talking about really building experiments to see the actions’ impacts on the flow.</p><p><strong>The model</strong></p><p>Every Retrospective meeting start showing metrics and explaining one by one to the team. Show the actual state of metrics and compare then with the prior metrics relating the actions discussed on the last Retrospective meeting. I recommend 4 metrics by default: burn-up chart (iteration length), Cumulative Flow Diagram (iteration length), Leadtime distribution (3 months length) and Flow Efficiency (3 months length). Try to link one metric with another and build a rational line of thought on the team members heads.</p><p>After this “ground work” run any technique of retrospective that you’re most familiar with.</p><p>In general all techniques has the same basic structure:</p><ul><li>What went well?</li><li>Improvement points</li><li>Actions</li></ul><p>Change all the vagues sentences to specific sentences towards those metrics showed earlier. I would recommend:</p><ul><li>What actions improved the flow/quality?</li><li>Which problems affected negatively the delivery flow?</li><li>Action/Experiments</li></ul><p>In the “Action/Experiments” section discuss with the team the actions to be done and how this actions should affect those metrics or what they expect to see of its impacts.</p><blockquote>If we can’t measure it we can’t say that it worked or not</blockquote><p>In the next Retrospective meeting, review those actions in the beginning and see if the actions made the expected impacts on the metrics. And then repeat the process.</p><p><strong>Conclusions</strong></p><p>Retrospective meetings are great to continuous improvement. Doing it with scientific method (hypothesis, experiments, successful metrics) make actions more tangible, more process centric and less prone to excuses.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bd822d6470ee" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How’s your workflow efficiency?]]></title>
            <link>https://medium.com/@reinaldocamargo/hows-your-workflow-efficiency-47f80a74b169?source=rss-f02cac568ab1------2</link>
            <guid isPermaLink="false">https://medium.com/p/47f80a74b169</guid>
            <category><![CDATA[agility]]></category>
            <category><![CDATA[workflow]]></category>
            <category><![CDATA[bottleneck]]></category>
            <category><![CDATA[agile]]></category>
            <category><![CDATA[efficiency]]></category>
            <dc:creator><![CDATA[Reinaldo Camargo]]></dc:creator>
            <pubDate>Thu, 09 Aug 2018 20:18:08 GMT</pubDate>
            <atom:updated>2018-09-01T16:40:16.268Z</atom:updated>
            <content:encoded><![CDATA[<h4>Bottlenecks, velocity and more metrics…</h4><p>I love to talk about workflow efficiency and bottlenecks. I think they’re some of the most interesting topics in Agility. Workflow optimization is the key to increase velocity and decrease the amount of time that a demand spends in process (leadtime/responsiveness/time2market).</p><p>Let’s start talking about the Cumulative Flow Diagram (CFD). The CFD shows the fingerprints left by demands on the workflow. So we can tell the history of a team just by looking at this chart. It’s my favorite chart and it’s a shame that Jira’s CFD is so ugly and not user-friendly. #gettingoutofthechest</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4GKn7ol9e41mwi9vU3-p0g.png" /><figcaption>Fig. 1: Cumulative Flow Diagram from a team</figcaption></figure><p>Reading Fig.1 left to right, we can see that some demands probably was sent directly to “Ready 2 test” state (light orange color) to be tested and some development work was started in “In progress” state (dark green color). After a while some demands was sent to “Waiting 2 deploy” (light blue color) to be deployed and after that some work was moved to “Done” (dark blue color) without testing (hahaha). Some work items were blocked (light green) later. And so on.</p><p>The work items flows throught the states ‘till Done, the last state.</p><p>Something that calls our attention is the bigger vertical space from some states. It happens when work items got stuck in a workflow state. For example, items with development work finished were moved to “Ready 2 test” were accumulating in this waiting state. As we can see, this state is growing bigger and bigger. That’s what we call bottleneck.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/345/1*TAvrmTYuIv71wf_v5mNzCQ.png" /><figcaption>Fig. 2: bottleneck</figcaption></figure><p>As you can notice as well, the development work (“In progress” state) pratically stopped latter. The team wasn’t limiting their WIP. They were just getting demands from left and passing them to the right.</p><p>Theory of Constraint from Goldratt tells that the system overall velocity is the velocity of its bottleneck/constraint. Optimizations made before the bottleneck won’t affect the velocity of the system. So we can say:</p><blockquote>“Bottleneck laughts at optmizations before it”</blockquote><p>There’s no use to make people get more and more work items from the “Selected 4 development” state. Actually, doing this will get bottleneck worse. The best approach would be to stop getting more items and put everyone to help on the items on “Ready 2 test” and the states ahead until the workflow allows the entrance of more items. And then, limiting the amount of work in progress to prevent this scenario again.</p><p>Another topic is how to calculate the workflow efficiency. The amount of time that a demand spends in process (Leadtime) is the sum of the time of doing some work (called Touch Time) and the time that this demand spends waiting for someone pulls it to do some work (called Wait Time).</p><blockquote>LD = TT + WT</blockquote><p>So, to calculate the workflow Efficiency (EF):</p><blockquote>EF = TT/LT * 100</blockquote><p>In order to calculate the TT if your team is using physical kanban board, just get the expedite card leadtime as TT. This leadtime doesn’t have any WT (or shouldn’t have). If your team uses Jira, go to the Leadtime chart (Control chart) and check only the states that some work is acctually being done. On the example of the Fig. 1, the efficiency is:</p><blockquote>EF = 1day/12days * 100 = 8%</blockquote><p>That’s a very low effiency. Work items spends 92% of time blocked or waiting some work to be done.</p><h4>Conclusion</h4><p>Limit work in progress. Team should always thinking getting things done and not just starting new ones. A bottleneck is a team’s problem and not a responsability of some specific person. Working better/more before a bottleneck is a useless optimization.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=47f80a74b169" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Make your bets to win without crystal ball]]></title>
            <link>https://medium.com/@reinaldocamargo/make-your-bets-to-win-without-crystal-ball-db8b52d9f6e9?source=rss-f02cac568ab1------2</link>
            <guid isPermaLink="false">https://medium.com/p/db8b52d9f6e9</guid>
            <category><![CDATA[agile-development]]></category>
            <category><![CDATA[metrics]]></category>
            <category><![CDATA[poker]]></category>
            <category><![CDATA[software-development]]></category>
            <dc:creator><![CDATA[Reinaldo Camargo]]></dc:creator>
            <pubDate>Fri, 13 Jul 2018 20:58:16 GMT</pubDate>
            <atom:updated>2018-07-16T20:48:43.313Z</atom:updated>
            <content:encoded><![CDATA[<h4>How to make predictions by managing risks with histogram</h4><p>I was talking about how to use histogram to make predictions about team’s deliveries couple of days ago with a friend at work. It seemed to me that he didn’t get it and I had no idea how to explain differently. After some thought I came with a Poker analogy.</p><p>I wrote about <a href="https://medium.com/@reinaldocamargo/weibull-distributions-curves-and-lead-time-metric-15b0b787b054">Weibull distributions, leadtime and histogram relations</a> but I didn’t write about how to use them to make predictions.</p><p>Make predictions about a team’s delivery it’s like a Poker game. In Poker you do some calculations based on the cards of the game and make bets according to the odds of risk. How much risk you are willing to take may change by your mood but the probability remains the same. It&#39;s all about managing risk.</p><p>Take a look at Fig. 1. It shows a leadtime chart from a team. The average time of a delivery is 17 days. Would you go “all-in” (bet all your money) that this team will deliver something between zero and 17 days?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wzn5RFWP7oQUZJQuk76n5A.png" /><figcaption>Fig. 1: Leadtime chart</figcaption></figure><p>Probably not. This is the average (Mean value) and, by default, the average shows that 50% of the cards were delivered ‘till 17 days. In &quot;Poker words&quot; I have a risk of 50% of losing money on the bet. I may be feeling lucky today and I may wanna go for it but the 50% of risk doesn’t change.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/765/1*3pLu6XtEV2COfufuWhsotw.png" /><figcaption>Fig. 2: Histogram from lead times</figcaption></figure><p>Fig. 2 shows the same data but in histogram chart. You can say for sure that this team can deliver something between zero and 110 days (I know it’s a long tail but ignore this for now). Doing that it’s like having an Ace in your hand and get the other 3 Aces using all the deck of cards. Of course you’ll get them eventually but it’s not sane (and not allowed by Poker rules :P). What I mean is: “between zero and 110 days” is really NOT what an executive/stakeholder wants to hear when he/she asks for previsibility. The variability is too wide.</p><p>How to make a good prediction without committing professional suicide? Here comes percentiles of certainty. I usually use three percentiles: 25%, 50%, 90%. I call them, in order, “I believe in miracles”, “I’m feeling lucky” and “I am Mr. Playsafe”. I’m kidding ;). It looks like I’m a pessimist person.</p><blockquote>To calculate these percentiles I need the sample space, in this case 452 items.</blockquote><blockquote>Percentile of 25: 452*0,25 =113 items</blockquote><blockquote>Percentile of 50: 452*0,5 =226 items</blockquote><blockquote>Percentile of 90: 452*0,9 = 407 items</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/765/1*Q7HP1YArzttKoM9nK0_PCg.png" /><figcaption>Fig. 3: Histogram with percentiles</figcaption></figure><p>Knowing these percentiles (Fig. 3) you can make some predictions. The Median (P50%) is showing that some feature may be develivered between zero and 12 days. But you’re assuming 50% of risk! Finally, in P90% shows that the team can deliver something between zero and 42 days. You can bet “all-win” because there’s a very good probability of getting it right. I know the range is too wide yet but it’s shorter than 110 days.</p><p>In Fig. 4 there’s a histogram from another team’s deliveries. The percentiles used was the same (25, 50 and 90).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/765/1*jeyI3b9Q-BxtHEDgHWEj-g.png" /><figcaption>Fig. 4: Histogram with percentiles from another team.</figcaption></figure><p>Here, P25% shows there’s a chance to delivery something between zero and 3 days. The Median (P50%) shows there’s a 50% of chance to delivery something between zero and 6 days. And P90% there’s a very good chance to delivery something between zero and 13 days.</p><p>Well, I hope this Poker analogy make histograms and percentiles simpler to understand. Making predictions involves risks. Know your chances of “winning” by managing risks.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=db8b52d9f6e9" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Stop guessing, start measuring]]></title>
            <link>https://medium.com/@reinaldocamargo/stop-guessing-start-measuring-223d49ec671d?source=rss-f02cac568ab1------2</link>
            <guid isPermaLink="false">https://medium.com/p/223d49ec671d</guid>
            <category><![CDATA[workflow]]></category>
            <category><![CDATA[kanban]]></category>
            <category><![CDATA[metrics]]></category>
            <dc:creator><![CDATA[Reinaldo Camargo]]></dc:creator>
            <pubDate>Tue, 03 Jul 2018 00:28:31 GMT</pubDate>
            <atom:updated>2018-07-05T12:07:04.265Z</atom:updated>
            <content:encoded><![CDATA[<h4>A pragmatic approach</h4><p>There’s an approach that Kanban adepts are calling <a href="https://www.linkedin.com/pulse/statik-systems-thinking-approach-implementing-kanban-david-anderson">STATIK</a>. This approach has some steps and seems to be a good way to start evolutionary (not <a href="https://medium.com/@reinaldocamargo/scaling-agility-is-your-company-ready-for-it-6f5d081a49d4">REvolutionary</a>) changes in a company. In STATIK there&#39;s a <strong>Step 4</strong> which analyze capability of a system by its metrics. The lack of data and how to build it will be the subject of this post.</p><p>Most companies take some conclusions based on guessing. “We need to deliver more taking less time” it’s a common request by executives. When asked for more details like “what exactly is expected by ‘deliver more’?” and “what exactly is expected by ‘taking less time’?” the answer usually has some assumption based on feelings. No evidence. No clue where to start.</p><p>Of course we could delve into what exactly executives are not satisfied with the current system but let’s focus on building metrics from work delivery.</p><blockquote>There’s no improvement without measuring. There’s no measuring without workflow. There’s no workflow without visualization. Without visualization, there’s no work.</blockquote><p>It’s like a mantra. Start with this mantra but in backwards.</p><p>It’s obvious there’s always work to do. But if it’s not visible, they exist for some and doesn’t for others. So, make work visible. Write them in a post-it or make a list in a virtual tool.</p><p>Pieces of work flows by many states/steps. Map these states and make it visible. If mapping states is getting too hard, use a simple proto-kanban (to do, doing, done). Use a physical or virtual board for it.</p><p>Then, you can measure how much work is being delivered in a timeframe and the average time spent in a work demand.</p><p>Now, there’s space for improvements. You have some clue where to start. “Deliver more” and “taking less time” have a more grounded meaning now. Experiments can be performed and its impacts are easily observed on the system.</p><p>It seems pretty straight forward but it’s amazing how many companies make their strategies based on feelings. Measuring and experimenting don’t come easily. It takes time to get used to. But when it becomes part of the company’s culture, every change is treated like an experiment. The outcome, regardless positive or negative, will make the company more robust and prepared to face external changes.</p><h4><em>References</em></h4><p>ANDERSON, David J., <a href="https://www.linkedin.com/pulse/statik-systems-thinking-approach-implementing-kanban-david-anderson">https://www.linkedin.com/pulse/statik-systems-thinking-approach-implementing-kanban-david-anderson</a>, 2016</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=223d49ec671d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[A Feedback canvas real implementation]]></title>
            <link>https://medium.com/@reinaldocamargo/a-feedback-canvas-real-implementation-6fb8e65b0253?source=rss-f02cac568ab1------2</link>
            <guid isPermaLink="false">https://medium.com/p/6fb8e65b0253</guid>
            <category><![CDATA[teamwork]]></category>
            <category><![CDATA[soft-skills]]></category>
            <category><![CDATA[feedback]]></category>
            <category><![CDATA[agility]]></category>
            <dc:creator><![CDATA[Reinaldo Camargo]]></dc:creator>
            <pubDate>Fri, 27 Apr 2018 17:34:25 GMT</pubDate>
            <atom:updated>2018-08-03T22:42:27.119Z</atom:updated>
            <content:encoded><![CDATA[<p>A study case of Feedback canvas session</p><p>Some weeks after I wrote an article here called &quot;<a href="https://medium.com/@reinaldocamargo/feedback-canvas-2-0-2f2040aa4a49">Feedback canvas</a>&quot; a team that I coach asked me for a feedback session. They wanted a 360 degrees feedback session. I changed a little bit the Feedback canvas implementation and ran the activity with them. So, I&#39;ll talk about this experiment on this post.</p><p>It is important to point out that it was not an imposed activity. All started by the need: It was the end of the quarter and they felt the need to give and receive feedback from each member of the team. I quickly explained to them the activity and they were willing to try it out. So, I scheduled two meetings with the team. The first one to show how <a href="https://medium.com/@reinaldocamargo/feedback-canvas-2-0-2f2040aa4a49">Feedback canvas</a> activity works and to make the <em>Pre-game</em> activity and second meeting to run the activity.</p><p>In the <a href="https://medium.com/@reinaldocamargo/feedback-canvas-2-0-2f2040aa4a49">Pre-game</a>, after explained how Feedback canvas works, we agreed to use 4 skills in the feedback activity of all team&#39;s members.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*cvC8rLznokIwUP6ymQdOfA.png" /><figcaption>Fig. 1: The four most voted skills by the team</figcaption></figure><p>On the second meeting we did the activity itself. I made a board for each member of the team and placed them at the walls. Each member take turns to get his/hers feedback. We started with self-evaluation, evaluation from others, strengths, improvements and, finally, actionable tasks as described <a href="https://medium.com/@reinaldocamargo/feedback-canvas-2-0-2f2040aa4a49">here</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*f8QSbFW3MRxiexY1l-dozg@2x.jpeg" /><figcaption>Fig. 2: My feedback board</figcaption></figure><p>The whole activity took two hour and thirty minutes with 6 people in the process.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tO6n6aIIXDUYTkzMYWMdHQ@2x.jpeg" /><figcaption>Fig. 3: The team members who participated on the Feedback activity. Everyone looks happy except the guy on the left =P (*)</figcaption></figure><p>After that, I submitted a survey to all participants. I wanted to evaluate the activity and the value perception over the time invested.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*T8fBkrTiFDBVG43TRt-jeg.png" /><figcaption>Fig. 4: Perception of value of the Feedback canvas activity</figcaption></figure><p><strong>Conclusions</strong></p><p>From the results shown in Fig. 4, they felt safe to share feedback and the activity promoted mutual help and professional enhancement. One thing pointed out to improve was the duration of the activity. Some of them said that it could last less than two hour and a half.</p><p>I would like to express my gratitude for the confidence of the team whose people I came to admire and learn.</p><p>* He said that he never smiles at photos LOL.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6fb8e65b0253" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Weibull distributions curves and Lead time metrics]]></title>
            <link>https://medium.com/@reinaldocamargo/weibull-distributions-curves-and-lead-time-metric-15b0b787b054?source=rss-f02cac568ab1------2</link>
            <guid isPermaLink="false">https://medium.com/p/15b0b787b054</guid>
            <category><![CDATA[agile]]></category>
            <category><![CDATA[lead-time]]></category>
            <category><![CDATA[weibull-distribution]]></category>
            <category><![CDATA[agile-metrics]]></category>
            <dc:creator><![CDATA[Reinaldo Camargo]]></dc:creator>
            <pubDate>Thu, 12 Apr 2018 22:20:18 GMT</pubDate>
            <atom:updated>2018-04-12T23:45:15.786Z</atom:updated>
            <content:encoded><![CDATA[<p>Analyzing team&#39;s performance by math and charts.</p><p>If you like gathering information from graphics about teams&#39; productivity this post is for you. Probably, you already use Throughput, Cumulative Flow or even the lead time chart. This post presents lead time and histogram graphics and how they’re related with Weibull distributions curves.</p><h4>The lead time metric</h4><p>In software development we call &quot;lead time&quot; the space of time that a demand enters the system and it gets out the system. I&#39;m not going too deep in the &quot;lead time&quot; matter now (maybe I&#39;ll dedicate a text about it later). So, the graphic below shows basically how much in days (y axis) a demand (x axis) took to be completed.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*vGzmV1M2cQkgqMnfvSONpQ.png" /><figcaption>Fig. 1: Lead time graphic from a development team</figcaption></figure><h4>The Weibull distributions curves</h4><p>The Weibull distribution, accordingly with [Wikipedia, 2018], is a continuous probability distribution.</p><p>The probability density function of a Weibull random variable is:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/292/1*ZP42oTmaNFl7M6szNT6ybA.png" /><figcaption>From <a href="https://en.wikipedia.org/wiki/Weibull_distribution">Wikipedia</a></figcaption></figure><p>Where k &gt; 0 is the shape parameter and λ &gt; 0 is the scale parameter of distribution.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Gk7urTjN0_-wdGmtstmLUA.png" /><figcaption>Fig. 2: Weibull distributions changing the shape parameter &quot;k&quot;. From <a href="https://en.wikipedia.org/wiki/Weibull_distribution">Wikipedia</a></figcaption></figure><p>Some interesting analysis here is when shape parameter kappa k &gt; 2 (green line), the distribution is near Gaussian shape and when k = 1 (red line), the shape is like a exponential function. Increasing or decreasing the sahpe parameter kappa will change the curve drastically in certain points.</p><h4>Histogram representation of lead time</h4><p>So far, so good! The last step before the analysis itself is to show the histogram representation of that lead time chart (Fig. 1).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/765/1*YHIK9dWla7l-4qLel7f_Qw.png" /><figcaption>Fig. 3: Histogram representation of lead times</figcaption></figure><h4>Comparing Weibull distributions with histogram for lead times</h4><p>We can clearly see a Weibull distribution with k &lt; 1 shape in Fig. 3. The space of time between the item which took less time (0 ~5.24 days) and the item which took more time (110 days) is too wide to make any reliable forecast. That is what we call “a long tail”. In other words: longer the tail, less predictability we have. This distribution is observed in development teams with unlimited work-in-progress (WIP), lots of blocking issues, too much wasting in workflow, high level of uncertainty, low maturity software development and so on. This distribution also appears in IT operations and IT services.</p><p>As [ANDERSON, 2016] wrote:</p><blockquote>&quot;If there is less than an even chance of something being delayed then the tail of the distribution gradually gets pulled to the right. Therefore, long duration tasks have more chance of being delayed and things which are delayed have more chance of being delayed further. This is what produces the long tail to the right.&quot;</blockquote><p>With k = 1 you still have a long tail on the right hand side and so unpredictability.</p><p>Improving flow efficiency, limiting WIP, focusing in removing blocking issues and wastes you can see the shape parameter kappa increasing. The value of k would be between 1 and 2. As we can see in Fig. 4 you still have a tail but it is shorter than k ≤ 1 or a “thin tail”. On this case you can make better predictions.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*i6caUhe9ot1GjI4-deANbw.png" /><figcaption>Fig. 4: Example of Histogram with Weibull distribution curve with k = 1.4.</figcaption></figure><p>About 2.0 &lt; k &lt; 4.0, in my opinion, is the best shape to make predictions. There’s a “very thin tail”. But in real scenarios of development [ANDERSON, 2016] “pre-emption does happen and some delay is incurred and also given that a specific item of a defined type may still have size and complexity that varies by two orders of magnitude, it is likely that we will see task durations with a distribution curve that has a shape parameter in the range 1.5 &lt; k &lt; 2.0”.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*GR-G6vhjFSgOrLauBpWESA.png" /><figcaption>Fig. 5: Example of Histogram with Weibull distribution curve with 2.0 &lt; k &lt; 4.0.</figcaption></figure><p>Rarely we&#39;ll have k=5 distribution like showed in Fig. 2 (no tail) in real scenarios of development.</p><h4>Conclusions</h4><p>We could see that lead times follows Weibull distributions curves. Of course that if you take a good look only at the lead time chart (Fig. 1) you can tell the variation is too wide, making forecasting difficult. But showing histograms and comparing it with Weibull distributions gives you a quick view of a development team&#39;s performance and its process maturity.</p><h4>References</h4><p>&quot;Weibull distribution&quot;. <a href="https://en.wikipedia.org/wiki/Weibull_distribution">Wikipedia</a> . Accessed April 2018</p><p>Anderson, David. &quot;<a href="http://anderson.leankanban.com/what-we-know-about-duration-workflows/">What We Know About Duration: Workflows</a>&quot;. 2016</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=15b0b787b054" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Product backlog prioritization technique: Kano Model]]></title>
            <link>https://medium.com/@reinaldocamargo/product-backlog-prioritization-technique-kano-model-63bc5d28a1fe?source=rss-f02cac568ab1------2</link>
            <guid isPermaLink="false">https://medium.com/p/63bc5d28a1fe</guid>
            <category><![CDATA[prioritization]]></category>
            <category><![CDATA[product-backlog]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[agile-methodology]]></category>
            <dc:creator><![CDATA[Reinaldo Camargo]]></dc:creator>
            <pubDate>Tue, 10 Apr 2018 13:56:44 GMT</pubDate>
            <atom:updated>2018-04-10T19:40:34.261Z</atom:updated>
            <content:encoded><![CDATA[<h4>No, it&#39;s not about Kano from Mortal Kombat ;) It&#39;s about making your product a &quot;must have product&quot; among users by doing killer features (without Fatality ;))!</h4><p>The Product Backlog list should be a place where everything is important. Not a bulk list of &quot;things&quot;. There&#39;s a lot of practices and techniques being discussed nowadays focused on how to deliver the right thing to the right client in the right time. To get there, you should know in earlier stages which needs should be satisfied in order to build a killer product. This post presents a technique called Kano model.</p><p>An educator, lecturer, writer and consultant in the field of quality management called <a href="https://en.wikipedia.org/wiki/Noriaki_Kano">Noriaki Kano</a> studied and classified the customer&#39;s expectations into 3 categories:</p><ol><li><em>Mandatory quality</em> (Basic expectations): Expectations on this category are considered basic that should be present in the product. Their absence will frustrate the customer but their presence will not increase their satisfaction, since the customer expects that feature to be present. As an example we can mention the water heating system in hotels. Hotels spend fortunes to keep water warm but no guest will post on Twitter or TripAdvisor &quot;OMG the water here is so warm!&quot; :P. However, if the water is not hot, or it takes too long to warm up, you can be sure that there will be complaints from all parties. No matter how much effort you spend on this you&#39;ll only be able to achieve the client&#39;s basic expectations.</li><li><em>Desired quality</em> (Satisfiers): Satisfiers are characteristics that can increase or decrease customer satisfaction. They are usually linked to performance. A good indicator from this quality is &quot;the more the better&quot;. For example the battery life of a cell phone. The longer it lasts the greater the customer satisfaction, however, the low duration will lower the satisfaction.</li><li><em>Delightful quality</em> (Delighters): Delighters are characteristics that satisfy the customer when they are present in the product, but if they are not they will not cause dissatisfaction. Most of the time it does not take much effort to deliver it to the customer, but the degree of satisfaction grows exponentially. As an example, a restaurant which serves a free portion of some finger food while you wait for your table.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/966/1*_uoR7jP_GV-GdSIa5ZVcIA.jpeg" /></figure><p>An interesting point in this model is that it predicts that all &quot;Delighters&quot; features will migrate to &quot;Basic expectations&quot;. Since the customer already expect that desired characteristic — either because you had it for a while or all your competitors already have it — this becomes a &quot;must-have&quot; feature, which means in the absence of it the customer satisfaction will drop significantly. Now you must find new &quot;Delighters&quot; features.</p><p>But now comes the question: How do I know if that characteristic is &quot;Delightful&quot;, &quot;Desired&quot; or &quot;Mandatory&quot;? One way is to build a questionary for a group of people who are interested in the product to answer 2 kinds of question for each feature to be implemented: a functional and dysfunctional one.</p><p><strong>Functional</strong></p><blockquote>How would you feel if there was a free bottle of water in your hotel room?</blockquote><blockquote>[X] I would like</blockquote><blockquote>[ ] I expect to be always like this</blockquote><blockquote>[ ] Neutral</blockquote><blockquote>[ ] I can live with that</blockquote><blockquote>[ ] I dislike</blockquote><p><strong>Dysfunctional</strong></p><blockquote>How would you feel if there WASN&#39;T a free bottle of water in your hotel room?</blockquote><blockquote>[ ] I would like</blockquote><blockquote>[X] I expect to be always like this</blockquote><blockquote>[ ] Neutral</blockquote><blockquote>[ ] I can live with that</blockquote><blockquote>[ ] I dislike</blockquote><p>And then you can use the following table to categorize the answers (functional and dysfunctional):</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*38M3B0nvTgMyrJu7UkzNxQ.png" /><figcaption>The analysis table tells you where a user would place a feature in the Kano Model based on how the functional and dysfunctional responses compare</figcaption></figure><p>On the &quot;free bottle of water&quot; example, this feature would be &quot;Delightful&quot; category. Now you should do the same thing for all the answers and group them into a table like the following:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/618/1*YYg2JrxJ3-eT2iH1NyACyw.png" /><figcaption>Example of grouped answers and features to be implemented</figcaption></figure><p>Based on that, we can see that &quot;feature A&quot; is a basic feature that customers hope to see. But it is important to remember that, since this functionality has been developed, we should not add effort to this feature as this will not increase customer satisfaction. We should only maintain it. The &quot;feature B&quot; seems to be basic for some people and &quot;the more the better&quot; to another. The &quot;feature C&quot; was classified like a delightful feature. Therefore, the prioritization would be A &gt; B &gt; C.</p><p>Lastly, in order to keep the Product Backlog prioritized by this model, it is necessary that all mandatory features be in the RoadMap (without exceeding the amount of effort required to reach the expected baseline), make the most of desired functionalities/characteristics and always try to leave a space for delightful features because they can increase the degree of customer satisfaction quickly in your favor.</p><p>References:</p><p>Eriksson, Martin. WebSite: <a href="http://www.mindtheproduct.com/2013/07/using-the-kano-model-to-prioritize-product-development/">Mind the Product</a>, 2013.</p><p>Cohn, Mike. Site <a href="http://www.infoq.com/presentations/prioritizing-your-product-backlog-mike-cohn">InfoQ</a>, 2008 — Agile Conference 2008</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=63bc5d28a1fe" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Scaling Agility: is your company ready for it?]]></title>
            <link>https://medium.com/@reinaldocamargo/scaling-agility-is-your-company-ready-for-it-6f5d081a49d4?source=rss-f02cac568ab1------2</link>
            <guid isPermaLink="false">https://medium.com/p/6f5d081a49d4</guid>
            <category><![CDATA[kanban]]></category>
            <category><![CDATA[kaizen]]></category>
            <category><![CDATA[agile]]></category>
            <category><![CDATA[scaled-agile-framework]]></category>
            <dc:creator><![CDATA[Reinaldo Camargo]]></dc:creator>
            <pubDate>Tue, 27 Mar 2018 12:33:06 GMT</pubDate>
            <atom:updated>2021-12-08T13:20:28.521Z</atom:updated>
            <content:encoded><![CDATA[<p>Nexus, SAFe, DaD, LeSS, &quot;Spotify model&quot;, etc. Lots of flavors and colors but do you know your company&#39;s taste?</p><p>This is the new hype medicine that will solve all your company&#39;s problems. Instead thinking in evolutionary terms I see most of the big tech consulting companies thinking in REvolutionary terms.</p><p>The revolutionary mindset is appealing. Sometimes you have to shake things up to produce a change (a good one we hope!). But when you&#39;re dealing with organizations it may be hard to produce a big change that will result in benefits for everyone. Basically, the complexity of an organization is too high to solve all problems with a pre-build-box framework.</p><p>Implementing a framework to scale Agility in a organization that has never heard about Agility has a good probability of failure. Not because those frameworks are bad but mainly because the Valley Of Death (lower curve of the J-Curve where changes go to die) is too deep.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/758/1*632EGtoZwu7dtU4OHzf3Qg.png" /><figcaption>Aspects of J-curve of change.</figcaption></figure><p>What I mean by &quot;the VOD is too deep&quot;? If the organization has no/few Agility background, bringing a huge change (with all the rules, ceremonies, roles, artifacts and so on) can impact negatively on productivity for too long. Time enough to frustrate people that the desired state will never be achieved. Bigger the change, bigger the VOD and bigger the time to achieve the desired state. [ANDERSON, 2016] says &quot;if executive tolerance &lt; pain of change(time and/or depth) then the change agents are fired&quot; ;).</p><p>In Kanban method there&#39;s an approach called &quot;Kaizen&quot;. Kaizen are small changes along over the time accordingly with the needs that raises from a system. In simples words, lots of &quot;J-curves&quot; instead of just one (in Kanban is called &quot;Kaikaku&quot;). The gain from this approach is that the VOD (or the depth of a change) tends to be smaller than a big change and you can see results more quickly.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/608/1*Evg0-wImG0cJ5hP3Dh6kpQ.png" /><figcaption>&quot;Kaizen&quot; changes versus &quot;Kaikaku&quot;changes</figcaption></figure><h4>Conclusion</h4><p>In my opinion, doing such a change all at once, even if your organization already have some Agility experience, causes a big impact and make results to take more time to became visible. I rather do small changes incrementally. Spending few months gathering information inside a company and propose a scaled agile framework it’s not an evolutionary approach and it&#39;s hard to control its impacts. It&#39;s a &quot;on going&quot; job. There will be always something to adjust.</p><p><em>References</em></p><p>ANDERSON, David J., <a href="http://leankanban.com/organizational-maturity-the-j-curve-effect/">Organization maturity &amp; the J-Curve effect</a>, 2016</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6f5d081a49d4" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Feedback Canvas]]></title>
            <link>https://medium.com/@reinaldocamargo/feedback-canvas-2-0-2f2040aa4a49?source=rss-f02cac568ab1------2</link>
            <guid isPermaLink="false">https://medium.com/p/2f2040aa4a49</guid>
            <category><![CDATA[feedback]]></category>
            <category><![CDATA[improvement]]></category>
            <category><![CDATA[agile]]></category>
            <dc:creator><![CDATA[Reinaldo Camargo]]></dc:creator>
            <pubDate>Fri, 23 Mar 2018 19:12:44 GMT</pubDate>
            <atom:updated>2018-08-08T18:17:38.956Z</atom:updated>
            <content:encoded><![CDATA[<p>Don&#39;t wait evaluations. Ask for feedback now.</p><p>In Agility there is a meeting where the team can evaluate what went well, things that stood in the way and what we can do differently to achieve better results. It&#39;s called Retrospective meeting. Most of the time issues on this meeting are addressed to the team and rarely to a specific member. With this in mind how can a specific member knows whether he/she is in the right path? How can one evaluate the quality of ones participation on that team? And, maybe the most important, how can he/she could improve?</p><p>Here is a quickly explanation about Feedback Canvas:</p><h3>What is it?</h3><p>According to [HADDAD, 2013] Feedback Canvas is:</p><blockquote>a tool to organize the collective feedback process and create a suitable context for everyone to collaborate with their perceptions about the work of a team member.</blockquote><p>We notice this canvas is really a interesting tool to give quickly feedback and more collaborative than &quot;one-a-one&quot; feedbacks.</p><p>It&#39;s basically a board where a group of pre-selected people leave their feedbacks towards an action/participation of a specific person. I made some modifications on the model but I&#39;ll explain later.</p><h3>How does it work?</h3><p>There&#39;s 5 steps to implement this canvas: Pre-game, Competence Rating, Strengths, Improvements and Actionable tasks.</p><p><em>Pre-game</em></p><p>Let&#39;s say there&#39;s a person who wants a feedback. Let&#39;s called her Arya ;). She invites a group of people who works or worked with her to join the Feedback Canvas session. Five is a good amount of participants of a Feedback Canvas session. In our example, Arya sends 5 invites, specifying the place, date and which ability/action/competence she wants feedback from.</p><p><em>Competence Rating</em></p><p>On the meeting day, people get together and the first task is to establish which competences are important for a good feedback. Once everyone agreed with the competences, the participants should rate them using a scale (explained below). In [HADADD, 2013] proposed model there&#39;s seven levels of rating. In our experiment we use only five. Something like here:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-GTu-8BMlsxepDD5FfnXng.png" /><figcaption>Example of Feedback Canvas</figcaption></figure><p>Each person (Arya as well) will rate all competences according to the scale. It&#39;s important that Arya uses a different color to vote from the others participants to visualize the difference of self-evaluation and how others see her work.</p><p>The images on the scale are explained here:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/560/1*zENmNBC5uNMs5X-t6dsnfw.png" /><figcaption>Dreyfus Models</figcaption></figure><p>One thing that was modified from the original process is that each person should individually write down the ratings into a post-it and deliver it to the feedback facilitator so that he/she compiles all ratings on the board. This was made to eliminate influence from others people’s ratings.</p><p><em>Strenghts</em></p><p>At this stage the participants raise positive points, always taking into account facts, so that Arya knows exactly what she is doing right.</p><p><em>Improvements</em></p><p>Here the participants raise points of improvement always taking into account events of the day-to-day.</p><p><em>Actionable tasks</em></p><p>In this last stage, participants help to create the actionable tasks which will help Arya to improve her in that competence.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*XYKzwRFcGhvD4ZMpGWZ0Bg.png" /><figcaption>Example of a Feedback canvas filled. Purple squares are from Arya and Orange squares are from the team.</figcaption></figure><h3>Conclusion</h3><p>At the end of the session, Arya will know how people see her work and wich actions she can take to become a better professional. The idea is this process can be done any day whenever someone finds feedback necessary. An important thing to point out is this model totally depends on the proactivity of each person, since it is not mandatory and the process starts from the person itself.</p><p>References:</p><p>HADDAD, Matheus. <a href="https://web.archive.org/web/20160818143520/http://gestao30.matheus.eti.br/feedback-ao-inves-de-avaliacao-de-desempenho/">Feedback ao invés de avaliação de desempenho</a>., 2014</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2f2040aa4a49" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Buddhism in Agility]]></title>
            <link>https://medium.com/@reinaldocamargo/buddhism-in-agility-a2bfd91d35ef?source=rss-f02cac568ab1------2</link>
            <guid isPermaLink="false">https://medium.com/p/a2bfd91d35ef</guid>
            <category><![CDATA[agility]]></category>
            <category><![CDATA[management]]></category>
            <category><![CDATA[my-little-pony]]></category>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[buddhism]]></category>
            <dc:creator><![CDATA[Reinaldo Camargo]]></dc:creator>
            <pubDate>Tue, 20 Mar 2018 15:00:41 GMT</pubDate>
            <atom:updated>2018-03-21T16:56:59.046Z</atom:updated>
            <content:encoded><![CDATA[<p>Is there value or appearance of value?</p><p>Talking to a friend today at work we discussed about something weird it&#39;s going on in software companies nowadays. In particular those companies trying to &quot;become Agile (capital letter as a noun?? ;))&quot;.</p><p>There&#39;s a lot of frameworks, process, tools, beliefs and all kinds of things that is sold as <a href="https://en.wikipedia.org/wiki/Panacea_(medicine)">panacea</a>. Big tech consulting companies (most of them) has a good portion of guilty here. They inflate a problem and sell Agility as a magical box with meetings, roles, Dos and Dont&#39;s and without progress measurement or progress evidence at all. Some manager would say &quot;Well, let&#39;s follow exactly what they&#39;re saying. They&#39;re big and must know what they&#39;re talking about (and we paid a lot of money as well…)&quot;. Deliberately ignoring system feedback and trying to &quot;put a big square piece in a smaller rounded hole&quot;. And when you try to cut the edges it will fail and the consultants will say &quot;Well, it&#39;s not framework/process/tool&#39;s fault. You guys who cut the edges… But hey, now you have all these meetings and syncs and roles…&quot;.</p><p>The Agility community is guilty as well. Non prepared professionals with just soft skills implements lots of activities with no tangible goals, boards to show how the employees are feeling that day and feedbacks sessions to cope with feelings (yeah, just like therapy sessions). All of this just because &quot;looks cool&quot;. This has been called here in Brazil as &quot;<a href="https://www.linkedin.com/pulse/what-did-i-mean-my-little-pony-management-rodrigo-yoshima/">Pony Management</a>&quot; (yeah, a little bit pejorative). Of course soft skills are needed but without goals and progress evidences (hard skills) you can&#39;t say for sure if it&#39;s getting better. It&#39;s easy to fall in biases and get false positives. Things like &quot;Oh, in our Work Organizational Climate Survey 100% of the programmers love to work here and definitely would recommend friends to work here&quot;. Yeah, pinball games, table tennis, xbox on TV, Nerf guns and beer at Thursdays are on! Productivity, quality software delivery and business value delivery… well, not so much… Again, I&#39;m not saying that employees must remain seated 8 hour a day in a cubicle looking for a computer screen but &quot;becoming cool&quot; without a clear purpose, without a tangible and well-shared goal you just treating employees like spoiled child. And like spoiled child they will behave.</p><p>The Buddhism lesson that we can apply here is to see through appearances. Are we delivering real value in all parts of chain? Can you tell for sure if something got better ou worse? Are you balancing soft skills and hard skills? Bad managers don&#39;t see value they see the appearance of value.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a2bfd91d35ef" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>