<?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[BacklogMiner - Medium]]></title>
        <description><![CDATA[BacklogMiner is a product management tool that can uncover themes in your data. Themes become projects. Projects fill your roadmap. We explore these concepts in detail here. - Medium]]></description>
        <link>https://medium.com/backlogminer?source=rss----456211cc11c7---4</link>
        <image>
            <url>https://cdn-images-1.medium.com/proxy/1*TGH72Nnw24QL3iV9IOm4VA.png</url>
            <title>BacklogMiner - Medium</title>
            <link>https://medium.com/backlogminer?source=rss----456211cc11c7---4</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 07:30:25 GMT</lastBuildDate>
        <atom:link href="https://medium.com/feed/backlogminer" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[The Deductive Thinker’s Path to Product Roadmapping]]></title>
            <link>https://medium.com/backlogminer/the-deductive-thinkers-path-to-product-roadmapping-cdffd3f6c3e1?source=rss----456211cc11c7---4</link>
            <guid isPermaLink="false">https://medium.com/p/cdffd3f6c3e1</guid>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[product-roadmap]]></category>
            <category><![CDATA[saas]]></category>
            <category><![CDATA[roadmaps]]></category>
            <category><![CDATA[prioritization]]></category>
            <dc:creator><![CDATA[Jason Wolfgang]]></dc:creator>
            <pubDate>Fri, 01 Jun 2018 00:40:20 GMT</pubDate>
            <atom:updated>2019-04-15T12:51:22.470Z</atom:updated>
            <content:encoded><![CDATA[<p>Let’s play a quick game. Match the following requests to the following personas:</p><h3>Requests</h3><ol><li>Can we share our product roadmap for the rest of the year with my prospect?</li><li>I need a two-year roadmap for our investors by tomorrow. I’m sure you’ve got something lying around.</li><li>Our quarterly review with our technology partner is next week, and they want to see which features on our roadmap are relevant to them. Essentially, they want assurance that we’re spending an ample percentage of our technology team on our growing partnership.</li><li>Can you fill out this section of RFP questions pertaining to our product roadmap?</li></ol><h3>Personas</h3><ol><li>CFO</li><li>Sales Engineer &amp; Solution Consultant</li><li>Sales Director</li><li>VP of Partnerships &amp; Business Development</li></ol><p>At the end of the day, the answer to “who is asking the question” matters less than the fact they are all asking the same person this question. You, the product manager. Which leads to another question: “How can we generate roadmaps?”</p><h3>The Ingredients for Roadmap Creation</h3><p>Most of the time, in my experience, requests for the roadmap imply that there’s a timeline element involved. Executives, sales reps, and client services want to know WHEN a feature of interest will be delivered, just like we’re interested when their deal will close or their client activation needs to go live.</p><p>Conversely, the tech team is very leery of giving out dates. They don’t want to be the team that lets the company down, and they know a lot of things can happen to make the dates on the roadmap untenable or irrelevant.</p><p>To navigate these dynamics when building a product roadmap, I have needed following items:</p><ol><li>A list of projects with hourly estimates (insert story points or whatever estimate measurement you use)</li><li>The number of developers within my organization</li><li>The number of weekly hours those developers can devote toward building the list of projects</li><li>The priority order of the projects</li></ol><p>I am a deductive thinker. I like Sudoku puzzles. I like Sherlock Holmes. I like solving problems with as much context as possible. Maybe that’s why I enjoy product management so much. There are a lot of constants defined in the product roadmap puzzle: time, team members, estimates, and priority order.</p><h3>Time</h3><p>There are only so many hours of quality coding that can happen in a week. There are certainly a finite amount of hours within a week. As an example, let’s say we have a 70 percent ratio, leading to 28 hours of coding a week.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*Dx6xPPDmNLHDhNlCDu9B7Q.gif" /></figure><h3>Team members</h3><p>You only have so many developers at your company. You could hire more if the business opportunity is worth it. For our example, we’ll start simple with one developer. We’ll show the impacts of adding developers and teams later.</p><h3>Estimates</h3><p>You won’t be building these projects, so you’ll need to clearly define the projects for the tech team to build them. You’ll need their input on the size of these projects. The project can be built out in phases. A project can be simplified. When it comes down to it, the product and the tech team must be able to defend the scope and size of the project’s estimate together.</p><p>To get a clear opportunity cost picture, the estimates don’t have to be precise. Consider grouping projects into shirt sizes to get a quick picture: S (8 hours; one day), M (40 hours; one week), L (160 hours; one month), XL (500 hours; one quarter), XXL (1000; more than two quarters). Just doing this will give you more clarity than you could imagine (or perhaps want, but we’ll get to that in a second).</p><p>Here’s a list of projects with estimates.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*tCtsc993rsMmRFwflIO4-w.gif" /></figure><h3>Priority order</h3><p>This sparks discussion. This is where stakeholder personalities come out. It’s helpful to make this dynamic and generate multiple roadmap scenarios to walk stakeholders through.</p><p>Let’s shuffle the priority order around and generate a few roadmap scenarios.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*sEth7a5ovnWap9OJ7NeV2g.gif" /></figure><h3>Strategically Grouping Projects Together</h3><p>In my opinion, any time the tech team thinks a project will be more than two quarters’ worth of effort, that project needs to be broken down into smaller projects.</p><p>Of course, grouping the resulting projects by a project category is still very important. This way, you can see how much investment a strategic initiative (in our example, project categories or swimlanes) will require.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*wFZPz5D36SxKNVNK7Xn4jw.gif" /><figcaption>Creating Project Categories to group similar projects on the <a href="http://backlogminer.com/">BacklogMiner</a> timeline</figcaption></figure><p>This gives you an opportunity to compare the project’s roadmap rating (<a href="https://medium.com/backlogminer/prioritize-your-roadmap-with-saas-metrics-2e0b6d5844df">check out this post on prioritization</a>) with the project’s required development and time investment.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*jloHWjbrq7yCD-fv6pGT4A.gif" /></figure><h3>Adding Team Members and Multiple Teams</h3><p>When you take a theoretical roadmap toward execution, things always get messy. Sometimes, you have to lock in a start date or an end date for one or many projects. The CEO throws a fist on the table: “This will be done by the end of the quarter!” The VP of Services asks, “What if we put more people on it?”</p><p>Here’s how I’ve used BacklogMiner to accommodate feedback like this within a roadmap scenario created on the fly during a high-pressure meeting.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*DyWjf0jh2MImc9xhL0Wt3w.gif" /></figure><p>Sometimes, you need multiple teams. Not all teams have the same skill sets. What if you need to bake a QA team into your roadmap? What if one team needs to get work done before another team? Admittedly, this starts getting into Jira-land in terms of tool sets. I have found that in certain organization structures, even a high level roadmap proposal requires this type of planning.</p><p>Here’s an example of adding teams and team-specific estimates to the mix:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*HIZi3vO7AYfTmPjklu_pKg.gif" /></figure><p>When working in multi-team mode, it’s also helpful to swap out the roadmap swimlanes from project categories to teams:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*VxE71TuitmIEWhUVmKB4DQ.gif" /></figure><p>This view reveals which teams routinely bottleneck strategically important projects. Once the executive team can see this, they have more information to address this situation from a headcount or process perspective.</p><p>Either way, it’s better to visually convey the problem rather than make decisions in the dark.</p><h3>Like It or Not: The Deduced Roadmap</h3><p>So, there you have it; we’ve backed our way into a long-term and short-term roadmap using the constants we have at our disposal and tweaking the priority order. Given all the projects we want to do and the resources we have, the roadmap’s duration is fixed regardless of the priority order.</p><p>Which projects give us the best chance of hitting our financial goals in the next six months? These type of questions and opportunity-cost scenarios can lead to some of the greatest product strategy conversations in your career.</p><p>This is also practical to maintain over time. Combine the power of text-mining product feedback into manually curated projects with forecasting roadmap, and I think you’ll find product management to be more clear and efficient.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cdffd3f6c3e1" width="1" height="1" alt=""><hr><p><a href="https://medium.com/backlogminer/the-deductive-thinkers-path-to-product-roadmapping-cdffd3f6c3e1">The Deductive Thinker’s Path to Product Roadmapping</a> was originally published in <a href="https://medium.com/backlogminer">BacklogMiner</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Text mine your product backlog. I think I get it, but….what?]]></title>
            <link>https://medium.com/backlogminer/text-mine-your-product-backlog-i-think-i-get-it-but-what-841b974eddc4?source=rss----456211cc11c7---4</link>
            <guid isPermaLink="false">https://medium.com/p/841b974eddc4</guid>
            <category><![CDATA[text-mining]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[product-roadmap]]></category>
            <category><![CDATA[product-backlog]]></category>
            <category><![CDATA[product-management]]></category>
            <dc:creator><![CDATA[Jason Wolfgang]]></dc:creator>
            <pubDate>Sun, 18 Mar 2018 18:12:34 GMT</pubDate>
            <atom:updated>2019-04-15T12:51:04.624Z</atom:updated>
            <content:encoded><![CDATA[<p>We product managers have all been there. The VP of Services sends you a spreadsheet. It’s from the big enterprise customer your growing SaaS company closed last quarter. The install isn’t going well. Expectations went sideways in the sales process. The Services executive’s clear path to success is to document all the client’s feedback (from a meeting at which you, the product manager, were not present) and commits to getting a timeline for implementing all their feedback into the product. Each row represents a request, and there are multiple columns categorizing and prioritizing each request.</p><p>Then the payoff question from the Services executive to you: “When will this be done? The go-live is in a month and the client isn’t happy.”</p><p>You muddle through the discussion, then take the spreadsheet and throw it into the folder with the VP of Sales’ spreadsheet, the CTO’s spreadsheet and the CEO’s spreadsheet. How do you create a product strategy out of that?</p><h3>I get the problem, but…text mining?</h3><p>What is text mining? From <a href="https://en.wikipedia.org/wiki/Text_mining">Wikipedia</a>, “Text mining, also referred to as text data mining, roughly equivalent to text analytics, is the process of deriving high-quality information from text. …</p><p>The overarching goal is, essentially, to turn text into data for analysis, via application of <a href="https://en.wikipedia.org/wiki/Natural_language_processing">natural language processing</a> (NLP) and analytical methods.”</p><p>A specific text mining concept using natural language processing that is very useful for product managers is <a href="https://en.wikipedia.org/wiki/Document_clustering">document clustering</a>. Within the context of this blog post and <a href="https://backlogminer.com/">BacklogMiner</a>, a document is a row in one of the spreadsheets mentioned above. It is an ad-hoc feature request, defect or product idea that came to you via email, meetings, support ticket, Jira, Pivotal Tracker, Aha.io, ProdPad, Roadmunk, ProductBoard… the list goes on.</p><p>The objective of <a href="https://en.wikipedia.org/wiki/Document_clustering">document clustering</a> is to group each document within your entire corpus of feedback (perhaps known as your product backlog, but perhaps it is really a superset of your “groomed backlogs”) and group them with other documents that contain similar words. Within <a href="https://backlogminer.com/">BacklogMiner</a>, these groups are called “bundles.” The GIF below demonstrates a spreadsheet being imported into <a href="https://backlogminer.com/">BacklogMiner</a> and organized into bundles via text mining:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*bcG-P3QyjiWQuavHbs2CDA.gif" /><figcaption>From spreadsheet to text mined bundles. <a href="https://www.youtube.com/watch?v=ltDrjQ-PVvU&amp;t=1s">CSV Import Details</a></figcaption></figure><h3>Interesting. How does that work again?</h3><p>It’s an intriguing idea right? You don’t have to pore through thousands of individual ideas/requests/defects/backlogs/documents anymore trying to bundle them together into a roadmap that your development resources can support. The document clustering routine bundles your feedback corpus for you, then you organize the bundles into development projects:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*d9VB_6h3CfSNyDB1mkeZ8g.gif" /><figcaption>Adding requests from a bundle into a project</figcaption></figure><p>The bundles actually become more useful when you roll up your <a href="https://medium.com/backlogminer/prioritize-your-roadmap-with-saas-metrics-2e0b6d5844df">prioritization metrics (more on that here)</a> from each idea/request/defect/document to the bundle. The bundles are grouped together and weighted based on their grouped priority.</p><p>So the nasty bug that everyone forgot about is linked to the big feature request from the new customer and you can knock both of them out in the same sprint cycle now that you’ve seen the defect and the feature request together in the same bundle. Thanks, text mining! We could get used to having you around.</p><h3>Bundling Maintenance Over Time</h3><p>Of course, the feedback pipeline never stops draining into your inbox, task list, slack channel or calendar. A bundle’s value doesn’t last long. As soon as your feedback corpus changes, you’ll need to re-bundle everything. Preferably without duplicate bundles, because that would defeat the efficiencies we’re trying to gain.</p><p>This is also why manual backlog grooming is an exercise in futility. As soon as you’re done, a new idea is thrown in.</p><p>For the bundling concept to overcome this duplication, we need to define the “active feedback corpus.” In other words, “Which tickets are still open?” or “Which tickets are not yet associated with a development project or release?”</p><p>We can define which tickets/requests/ideas are open or active like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*Rg_aloRxKDehlC0NnmpwPA.gif" /><figcaption>Keep the active feedback corpus clean by archiving old (or irrelevant) requests and assigning promising ones to projects containing similar requests</figcaption></figure><p>So the remaining tickets make up our active feedback corpus. Anytime something new is submitted, a subject is changed or a description amended; the bundling process needs rerun. Preferably automatically.</p><p>So, how do we prevent duplicate bundles if we’re running this routine so often? Once the bundling is done, the top five most common words (in descending order) are concatenated to give the bundle a name. Not only does this give us a quick summary of what content the bundle contains, it also is a unique identifier to deduplicate with. Observe:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*cZpdROt1N2pUKdvjMD_94Q.gif" /><figcaption>Bundle name as unique identifier</figcaption></figure><h3>This is a good time to chat about stop words</h3><p>How do you keep articles and other common words from overrunning your text mining efforts? <em>Stop words</em> are words you want removed from the text mining corpus before the document clustering routine runs.</p><p>Out of the box, BacklogMiner comes with a default list of common stop words, but adding words to it could improve your results. Especially if you find the same words throwing off your bundles repeatedly. Here’s a quick scroll through some of the default stop words, then we’ll add a custom stop word:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*o9fJi7qDwcwpsMwS4CSdNA.gif" /><figcaption>Stop words</figcaption></figure><h3>Bundles are a great new dimension for you to roll up your metrics</h3><p>Remember those <a href="https://medium.com/backlogminer/prioritize-your-roadmap-with-saas-metrics-2e0b6d5844df?source=linkShare-57feb9d20bf9-1521168549">prioritization formulas? I talked about them in a previous blog post</a>. It bears repeating that the bundle gives you a new way to roll up your prioritization metrics. And a whole new way to look at your backlog.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*T6-te3876PzVK8sU4ejq7A.gif" /><figcaption>Bundle rollups</figcaption></figure><p>Curating raw feedback into refined development projects is hard. Text mining raw feedback into bundles continuously over time creates a very handy stepping stone, helping you to always know what is going on in your backlog and how it relates to your roadmap.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=841b974eddc4" width="1" height="1" alt=""><hr><p><a href="https://medium.com/backlogminer/text-mine-your-product-backlog-i-think-i-get-it-but-what-841b974eddc4">Text mine your product backlog. I think I get it, but….what?</a> was originally published in <a href="https://medium.com/backlogminer">BacklogMiner</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Prioritize Your Roadmap with Financial Metrics]]></title>
            <link>https://medium.com/backlogminer/prioritize-your-roadmap-with-saas-metrics-2e0b6d5844df?source=rss----456211cc11c7---4</link>
            <guid isPermaLink="false">https://medium.com/p/2e0b6d5844df</guid>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[finance]]></category>
            <category><![CDATA[prioritization]]></category>
            <category><![CDATA[saas]]></category>
            <category><![CDATA[product-roadmap]]></category>
            <dc:creator><![CDATA[Jason Wolfgang]]></dc:creator>
            <pubDate>Thu, 15 Feb 2018 03:31:59 GMT</pubDate>
            <atom:updated>2019-04-15T12:50:27.013Z</atom:updated>
            <content:encoded><![CDATA[<p>Product backlogs contain a variety of problems: This customer’s sales team uses an old mobile operating system, and our mobile app doesn’t work on it; our dev team is three releases behind on a key framework used in our tech stack; it takes our services team too long to implement specific modules of the product; our sales team can’t make pipeline targets because we’re missing these three features.</p><p>How do we objectively prioritize business problems like this? Heck, how can we even objectively compare such diverse problems against each other? Assuming each of these issues needs to be solved with the same development team, shouldn’t we at least try to reconcile their value against a consistent criteria?</p><p>Note: If you have multiple development teams for different products or enough developers to handle concurrent development projects, I still contend this dynamic applies to you. You still probably have more backlog for a given product or development team than you can get done in a calendar year. If you don’t, then you need to get out of the office and start talking to your customers or tagging along with sales reps to understand your prospects.</p><p>(By the way, great Twitter thread by <a href="https://twitter.com/johncutlefish">John Cutler</a> on the <a href="https://twitter.com/johncutlefish/status/959210921779380224">topic of multiple backlogs here</a>).</p><p>(Constructive discussion with <a href="https://twitter.com/simplybastow?lang=en">Jana Bastow</a> on <a href="https://www.mindtheproduct.com/2018/02/product-warrior-podcast-janna-bastow-talks-challenges-established-products/">engaging your clients directly here</a>. NIHTO = Nothing Important happens in the office)</p><h3>To the Income Statement</h3><p>The only way I’ve had success turning the fruit salad of a product backlog into an apples-to-apples opportunity cost comparison is by mapping everything to the company’s <a href="https://www.investopedia.com/terms/i/incomestatement.asp">income statement</a>. Here’s a quick Business Finance 101 review from an unqualified source (me):</p><ul><li>Revenue</li><li>- <a href="https://www.investopedia.com/terms/c/cogs.asp">Cost of Goods Sold</a></li><li>= <a href="https://www.investopedia.com/terms/g/grossmargin.asp">Gross Margin</a></li><li>- <a href="https://www.investopedia.com/terms/o/operating_expense.asp">Operating Expenses</a></li><li>= <a href="https://www.investopedia.com/terms/e/ebitda.asp">EBITDA</a></li><li>- Interest, Taxes, Depreciation, Amortization</li><li>= <a href="https://www.investopedia.com/terms/f/freecashflow.asp">Free Cash Flow</a></li></ul><p>So, each backlog request’s business case should be able to answer some of these questions: How does this problem help us acquire new revenue? How does this feature help us protect existing revenue? How does this feature help us reduce hard or soft costs? An example of a hard cost would be reducing the Amazon Web Services bill. An example of a soft cost would be making the product easier to activate so the services team can double its project load without increasing headcount.</p><h3>The Prioritization Formula</h3><p>A wealth of product management prioritization frameworks exist. <a href="https://foldingburritos.com/product-prioritization-techniques/?utm_content=buffer99620&amp;utm_medium=social&amp;utm_source=twitter.com&amp;utm_campaign=buffer">Here’s a nice primer on many of them</a>. Steve Johnson has a <a href="https://www.under10playbook.com/articles/prioritization/">concise article on quick prioritization here</a> (I’ve been a student of Steve Johnson’s for many years. Attended his pragmatic marketing seminars early in my product management career).</p><p>Here’s a useful paraphrase I heard from Steve at Pragmatic’s Product Management in an Agile Development Environment session ten years ago: “Let’s argue about the prioritization formula, not the prioritization of today’s specific issues.” I believe the intent behind this is to make prioritization more objective and less emotional. Which is hard, because it gets emotional fast when the VP of Services has features blocking go-live dates for key clients.</p><blockquote>“Let’s argue about the prioritization formula, not the prioritization of today’s specific issues.” — <a href="https://twitter.com/sjohnson717?lang=en">Steve Johnson</a>, paraphrased by myself</blockquote><p>So, here’s the prioritization formula I’ve used for a long time and baked into BacklogMiner:</p><p>Roadmap Rating = (Severity * (Current Revenue + New Revenue + Cost Savings) + Executive Override)</p><p>It’s simple. It covers new sales and existing accounts. It considers your company’s efficiencies. It makes room for the executive idea out of nowhere or at least shows the cost of taking that idea to priority number one. Let’s dig into it.</p><h3>Definition of terms</h3><p>Let’s break down precisely what each term in that prioritization formula means, shall we?</p><blockquote>Severity</blockquote><p>Severity is a numeric representation of the objective priority of a request (a request can be an idea, feature, or defect). Consider severity at, say, five different levels:</p><ul><li>Severity 5 (Show stopper): Minimum purchasing requirements not met</li><li>Severity 4 (Broken): Money is being lost by our company or the customer</li><li>Severity 3 (Has Workaround): Primary business goal difficult to achieve</li><li>Severity 2 (Has Workaround): Non-primary business goal difficult to achieve</li><li>Severity 1 (Low): Product request from a non-targeted vertical market</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*kVS5uIYCGmFgh8r45YBZMw.gif" /></figure><blockquote>Current Revenue</blockquote><p>This metric comes from mapping a client’s account to each of their requests. The metric can take many forms:</p><ul><li><strong>Account Vote:</strong> Each account related to Feature X gets one vote.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*1iShceTBxQRUFxssrd0h8w.gif" /><figcaption>15 Rating = 5 Severity * (2 Account Votes + 1 Rating Weight)</figcaption></figure><ul><li><strong>User Count: </strong>Accounts are weighted based on how many users they have. So if Company A has 10 users and Company B has 100 users and they both vote for Feature X which has a Severity 5 rating, the Roadmap Rating would be calculated 5 * ((100+10) + New Revenue + Cost Savings). Here’s a walk-through of this example in the following gif:</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*XyZ5ymZo34XXaXo2E7zLsg.gif" /><figcaption>555 Rating = 5 Severity * ((10 + 100 Users) + 1 Rating Weight)</figcaption></figure><ul><li><strong>ARR (</strong><a href="https://www.zuora.com/billing-topics/annual-recurring-revenue/"><strong>Annual Recurring Revenue</strong></a><strong>):</strong> The math is the exact same as the user count scenario, however, we’re using a client’s <a href="https://www.zuora.com/billing-topics/annual-recurring-revenue/">ARR</a> as the metric. This is useful if your pricing is not always based on number of users. For example, a 5-user account that sends out 1 million automated emails a month is now equal with a 1,000-user account that does everything manually and calls support a lot. Here’s a walk through of this concept:</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*w8V1SxT0TWw7DnxRSY9hjQ.gif" /><figcaption>810,005 Rating = 5 Severity * ((12,000 + 150,000 <a href="https://www.zuora.com/billing-topics/annual-recurring-revenue/">ARR</a>) + 1 Rating Weight)</figcaption></figure><p>Note: it is interesting to run both the User Count model and the <a href="https://www.zuora.com/billing-topics/annual-recurring-revenue/">ARR</a> model and compare the two. You can learn a lot about your customer base, for example, who has a lot of users and is paying less, or who is a large client you never hear from.</p><blockquote>New Revenue</blockquote><p>This metric is very similar mathematically to Current Revenue, however, the user count or <a href="https://www.inturact.com/blog/acv-annual-contract-value-vs-arr-annual-recurring-revenue">annual contract value</a> (ACV) metrics are coming from your sales pipeline rather than your billing system. If your company is in growth mode, you may consider weighting new revenue as more valuable than existing revenue. That of course depends on your business model and your market.</p><blockquote>Cost Savings</blockquote><p>I’ve often referred to this as <em>internal efficiency</em> throughout my career. Here’s a scenario: the CIO raises their hand in a meeting and says, “All these features are nice, but installing them into production will be expensive in our current deployment model. We need to refactor our environment.” Or the VP of Services interjects, “Installing this module for a client takes one FTE (Full Time Equivalent) a full month. If we close five more deals like this next quarter, we won’t have enough staff to keep up.”</p><p>This has been the toughest metric to reign in (save for the Executive Override). If you’re using account votes or user counts for your revenue metrics, it is tricky to translate the “hours saved” or “future headcount not required” into account or user equivalents. If you use the actual <a href="https://www.zuora.com/billing-topics/annual-recurring-revenue/">ARR</a> or <a href="https://www.inturact.com/blog/acv-annual-contract-value-vs-arr-annual-recurring-revenue">ACV</a> in your revenue metrics, the hard and soft cost savings benefit of internal only requests are at least on the same measurement scale. The dollar.</p><p>So, if Feature Y offers a $10,000 cost savings, we can compare it with the <a href="https://www.zuora.com/billing-topics/annual-recurring-revenue/">ARR</a> version of Feature X (from the example above) like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*Nlh97vUx-msNLbTwa5EDBQ.gif" /><figcaption>Feature X: 810,005 Rating = 5 Severity * ((12,000 + 150,000 <a href="https://www.zuora.com/billing-topics/annual-recurring-revenue/">ARR</a>) + 1 Cost Savings Rating Weight) | Feature Y: 50,000 Rating = 5 Severity * (10,000 Cost Savings Rating Weight)</figcaption></figure><p>Clarifying the relative importance of <a href="https://www.investopedia.com/terms/g/grossmargin.asp">Gross Margin</a> versus <a href="https://www.investopedia.com/terms/e/ebitda.asp">EBITDA</a> versus Free Cashflow in your company can also help create general guidelines for prioritizing these type of requests. You also need to know how the cost savings being proposed helps reach the profitability goals your company cares about.</p><p>Typically, your services team will be client billable resources; meaning their salaries come out of the <a href="https://www.investopedia.com/terms/c/cogs.asp">COGS</a> line item on the income statement. Vendor payouts (like your AWS bill) also come out of <a href="https://www.investopedia.com/terms/c/cogs.asp">COGS</a>. Reducing <a href="https://www.investopedia.com/terms/c/cogs.asp">COGS</a> will increase your <a href="https://www.investopedia.com/terms/g/grossmargin.asp">Gross Margin</a>. Development and Support typically come out of <a href="https://www.investopedia.com/terms/o/operating_expense.asp">Operating Expenses</a>, which have no bearing on <a href="https://www.investopedia.com/terms/g/grossmargin.asp">Gross Margin</a>. So, if your company is less focused on <a href="https://www.investopedia.com/terms/e/ebitda.asp">EBITDA</a>, the business cases for development and support efficiencies may need to be complemented with risk assessment in addition to hard and soft cost savings analysis.</p><h3>So Now What?</h3><p>Once all these paradigms are in place, you can see the roadmap prioritization impact of growing the size of an account:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*Wqk6MXIUn3JJKFtFTD82eQ.gif" /></figure><p>Or closing a new opportunity that needs a feature you’ve already contemplated in your backlog:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*T5zyK8rNahWGgf6iRpY3Rg.gif" /></figure><p>Of course, organizing backlog data with consistent quantitative analysis doesn’t grant a license to ignore anecdotal, qualitative analysis. Again, let me insert a plug for <a href="https://twitter.com/simplybastow?lang=en">Jana Bastow</a>’s interview on <a href="https://www.mindtheproduct.com/2018/02/product-warrior-podcast-janna-bastow-talks-challenges-established-products/">Product Warrior podcast here</a>.</p><p>When the CEO has an idea or we perceive a new industry trend, we must make room on our roadmaps for innovation when appropriate. Using our quantitative analysis to measure the opportunity cost of the innovation is valuable. So, we can then ask questions like: “Do we have a chance to make $2,000,000 in new revenue off this innovation in the first year after launch? Because that is the value of the projects we’re not doing so we can innovate.” Perhaps questions like these can help us deductively write business cases for these hard to measure projects.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2e0b6d5844df" width="1" height="1" alt=""><hr><p><a href="https://medium.com/backlogminer/prioritize-your-roadmap-with-saas-metrics-2e0b6d5844df">Prioritize Your Roadmap with Financial Metrics</a> was originally published in <a href="https://medium.com/backlogminer">BacklogMiner</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why I built BacklogMiner]]></title>
            <link>https://medium.com/backlogminer/why-i-built-backlogminer-bfb18a5520b5?source=rss----456211cc11c7---4</link>
            <guid isPermaLink="false">https://medium.com/p/bfb18a5520b5</guid>
            <category><![CDATA[product-backlog]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[product-roadmap]]></category>
            <category><![CDATA[text-mining]]></category>
            <dc:creator><![CDATA[Jason Wolfgang]]></dc:creator>
            <pubDate>Sat, 20 Jan 2018 19:55:59 GMT</pubDate>
            <atom:updated>2019-04-15T12:49:06.304Z</atom:updated>
            <content:encoded><![CDATA[<p>My name is Jason Wolfgang. I’ve been a software product manager for over ten years and I’ve built a tool for people like me called <a href="https://backlogminer.com/">BacklogMiner</a>. Okay, I really built the tool for myself, because I had a few specific problems I wanted to solve. I was good on roadmap visualization tools. I was good on idea management tools. But there were still some gaps for me from an idea-to-roadmap-to-release workflow.</p><p>Here’s the story of why I built <a href="https://backlogminer.com/">BacklogMiner</a>.</p><h3><strong>Identifying roadblocks for product managers</strong></h3><p>The majority of my career has been with medium-size software companies, and with each, I’ve seen revenue, customer, and employee growth. With each, the corpus of product feedback got unwieldy in similar ways:</p><ul><li>It was tough to map each customer to the feedback they gave us.</li><li>It was tough to measure the value of their feedback.</li><li>It was tough to bundle their feedback into projects.</li><li>It was tough to prioritize the projects on to a roadmap timeline.</li><li>It was tough to measure the opportunity cost of the “benched” projects versus the projects prioritized on the roadmap.</li></ul><p>These were my problems to solve. It is the product manager’s job to bundle product feedback. We have to define projects that align with the company’s customer retention and expansion goals:</p><ul><li>Goals for new business</li><li>Goals for entering new markets</li><li>Goals for cutting costs and gaining internal efficiencies</li><li>Goals for keeping tech debt at an acceptable level.</li></ul><p>It is the product manager’s job to recommend roadmap scenarios that map to the company’s financial goals. When debates inevitably occur over roadmap mix and priorities, it’s essential to be able to clearly explain the opportunity cost of each prioritized project versus a non-prioritized project. No one is better positioned in the company to do this than the product manager.</p><h3>Organizing the feedback haystack</h3><p>I didn’t have a simple way to pull all this together. The easiest of these for me to measure was customer feedback, but that wasn’t without its challenges. The customer list was in the CRM. The product feedback from customers could be in a variety of places: support cases, email, spreadsheets, development tickets. The revenue from each client was in the finance system.</p><p>Going through feedback and mapping it to the customer list (and mapping each client’s revenue impact) was tedious work. However, it began to sift the important needles to the top of the feedback haystack. Like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*zMN0EKL9OvO0JHncU-avnA.gif" /></figure><p>Unfortunately, the haystack was thousands of product requests. During 2013, a thought crept into my mind: “Could there be a way to use technology to group requests with similar words to form feedback bundles?” My original concept was to go through 200 bundles containing similar requests (all mapped to the requesting customer along with their revenue) rather than 2,000 requests in descending order by customer revenue.</p><p>After two years of working with open-source text mining tools like khcoder, I decided to learn R (an open source development language that is great for text mining) and bake them into a web application tailored for a product manager (okay, tailored for me). There were a lot of crazy questions to answer, like “What is the right value for k?” (k-means analysis joke) and “How do you make R play nice with data that is constantly changing?” However, after all the dust settled, I was able to get a list of bundled requests like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*0eQRsxK1-hAO1-MfI72tbQ.gif" /></figure><p>The next step was to group all these feedback bundles into projects. Because even though the text mining was really cool, good project definition still needed human intervention. But much less of it! In a matter of hours, I could clean up the messy backlog into neatly organized projects.</p><h3>Providing realistic timelines</h3><p>Once the projects were organized, I kept finding myself in meetings with execs asking questions like:</p><ul><li>“If we bumped this project up on the priority list, what else would get delayed?”</li><li>“What if we put another developer on that project?”</li><li>“How many developers do we need to hire to get everything done on the timeline we want?”</li></ul><p>This led to me to build out a timeline forecast feature. I was stunned such a feature didn’t exist in any product management tool. Perhaps because managing development projects to specific dates is a function of the VP of Engineering, which means dev tools (like Jira) or project management tools (like Basecamp) would have such a feature as opposed to product management tools.</p><p>However, in my experience, the VP of Engineering was well-versed on projects in progress (<em>what is my 90-day plan?</em>) and less aware of which project should come next (<em>what is my 12-month roadmap?</em>).</p><p>This constantly put me in an awkward situation between the CXO asking me, “When will this important project that hasn’t started yet be done?” and the VP of Engineering telling me, “Don’t worry about dates, we don’t have resources anyway.”</p><p>So, I built the timeline-forecasting feature, which gave me the ability to answer CXO questions like these on the fly:</p><ul><li>When could the project be done with another developer?</li><li>What happens if we bump that project up three spots on the priority list?</li></ul><p>The opportunity cost of shifting priorities is obvious, as is the effect of adding fictional developers to visualize a brainstorming scenario. When you find a scenario you like, push it out to your roadmapping tool to make it pretty, knowing that it isn’t just another roadmap with no data to back it up.</p><p>So, those are the problems that motivated me to build <a href="https://backlogminer.com/">BacklogMiner</a>. I’m not the only one with these problems, right?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*HqAEwsNk2dLsWuJIX7OMFA.gif" /><figcaption>Timeline forecasting in <a href="https://backlogminer.com/">BacklogMine</a>r</figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bfb18a5520b5" width="1" height="1" alt=""><hr><p><a href="https://medium.com/backlogminer/why-i-built-backlogminer-bfb18a5520b5">Why I built BacklogMiner</a> was originally published in <a href="https://medium.com/backlogminer">BacklogMiner</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>