<?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 Jenn Dearth on Medium]]></title>
        <description><![CDATA[Stories by Jenn Dearth on Medium]]></description>
        <link>https://medium.com/@jenndearth?source=rss-30ce0bc97f06------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*J8ExdNgm9pEUbpdhezqnAw.jpeg</url>
            <title>Stories by Jenn Dearth on Medium</title>
            <link>https://medium.com/@jenndearth?source=rss-30ce0bc97f06------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 13:37:55 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@jenndearth/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[Building AI for Financials Products]]></title>
            <link>https://medium.com/workday-engineering/building-ai-for-financials-c3a9d512816c?source=rss-30ce0bc97f06------2</link>
            <guid isPermaLink="false">https://medium.com/p/c3a9d512816c</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[user-experience]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[accessibility]]></category>
            <category><![CDATA[explainable-ai]]></category>
            <dc:creator><![CDATA[Jenn Dearth]]></dc:creator>
            <pubDate>Wed, 29 Jan 2025 03:58:23 GMT</pubDate>
            <atom:updated>2025-05-06T17:49:29.906Z</atom:updated>
            <content:encoded><![CDATA[<p><em>How high stakes requirements drive better explainability and trust</em></p><p>By <a href="https://www.linkedin.com/in/jenndearth/">Jenn Dearth</a> with contributions from <a href="https://www.linkedin.com/in/jvantomme">Jason Vantomme</a>, <a href="https://www.linkedin.com/in/brainynet/">Howard Shaw,</a> <a href="https://www.linkedin.com/in/kyle-mollinet-64b225133/">Kyle Mollinet, </a><a href="https://www.linkedin.com/in/omshankartiwari/">Shankar Tiwari</a>, <a href="https://www.linkedin.com/in/andrea-villa-ca/">Andrea Villa</a>, and <a href="https://www.linkedin.com/in/rebecca-novinger/">Rebecca Novinger</a></p><p>When I joined Workday’s Machine Learning team four years ago as a Product Lead for Financials, I quickly realized it was a complex domain packed with highly trained specialists. I found myself excitedly digging into the details to better understand the journey of our users’ experiences; how the job-to-be-done (JTBD) by an Accounts Receivable specialist differs from a Cash Manager and a Financial Planning and Analysis Manager.</p><p>I also noticed how there were high stakes for our users. An error in reconciling transactions flows onto financial statements, potentially leading to regulatory fines, restatements, or even personal liability for the individuals who signed off on those statements. Because of this, finance professionals are cautious to adopt new technology without strong assurance that it meets their needs and complies with regulatory standards.</p><p>Solving for specific, high stakes requirements of our financial users reminded me of the “Curb Cut Effect”. Don’t know what a curb cut is? Well, if you’ve ever wheeled a suitcase across a crosswalk or pushed a stroller down a city block, you’ve likely taken advantage of a curb cut. These ramped edges of sidewalks were first introduced in the United States during the 1960s, advocated by veterans and disability-rights activists, and later mandated by the Americans with Disabilities Act (ADA) in 1990. Their initial goal was to enable safer navigation for people in wheelchairs. But to everyone’s surprise, curb cuts also proved invaluable for delivery workers with handcarts, young children learning to bike, parents with strollers, and travelers hauling heavy luggage.</p><blockquote><strong><em>When we solve for a specific, high-stakes use case, we often build better solutions for all.</em></strong></blockquote><p>The “curb cut effect,” illustrates a broader principle: when we solve for a specific, high-stakes use case, we often build better solutions for all. Designing for finance users is hard, but if we do it well, everyone benefits. While not immediately obvious, delivering solutions for wheelchair users and finance power users followed a similar path. <strong>This post tracks our journey into designing AI features for Workday’s Financials product and how that has served as a forcing function to create safer, more accessible, and ultimately more trustworthy products for all types of users.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/902/0*mrZ10TJuIvjLbUKB" /></figure><p><a href="https://dceg.cancer.gov/about/diversity-inclusion/inclusivity-minute/2021/curb-cut-effect">https://dceg.cancer.gov/about/diversity-inclusion/inclusivity-minute/2021/curb-cut-effect</a></p><h3>Why Finance Demands So Much from Software… and AI</h3><p>According to Gartner, 90% of finance functions will deploy at least one AI-enabled technology solution by 2026, and by 2027, 90% of description and diagnostic analytics in finance will be fully automated. So how can we help usher this technology into financial products? What do users working in the office of the CFO need?</p><p>From day one, I noticed three aspects of Financials products that distinguish them from most other enterprise applications:</p><ol><li><strong>Complex, Intertwined Personas<br></strong>It’s rare for a single transaction to involve just one person in finance. Behind the scenes, Cash Managers and Controllers may also review or reconcile these expenses against budgets. A single purchase might touch five or six roles (or even more), each with specialized workflows and responsibilities. As we design AI features, we have to be sure to look upstream and downstream of transactions in order to make sure we’re designing a holistic experience.</li><li><strong>Unprecedented Volume<br></strong>Finance professionals often handle significantly more data than many of their peers in other functions. A recruiter might juggle a few hundred applicants at a time, which is intense, but manageable. Meanwhile, a single Accounts Payable specialist could process thousands or tens of thousands of invoices monthly; a large enterprise might have millions of journal entries per year. These massive volumes mean that any AI solution we build must focus intently on time savings and efficiency. A feature that saves a few seconds on one transaction might translate into hours of time saved when multiplied across thousands of transactions.</li><li><strong>Stringent Auditability and Traceability<br></strong>By its nature, finance is under constant scrutiny, both internal (from management or boards) and external (from regulators, auditors, and sometimes even investors). Customers purchase financial management products specifically to maintain controls over record-keeping, enforce compliance, and create an unbroken chain of documentation for every transaction. If an AI-powered system can’t provide the rationale behind a recommendation, or if it leaves a gap in the audit trail, finance professionals will rightly hesitate to adopt it.</li></ol><h3>The Tug-of-War Between Accuracy and Explainability</h3><p>It’s almost impossible to talk about transparency in AI without addressing the tension between sophisticated “black box” models and user-friendly explainability. In machine learning, complex models like deep neural networks often offer higher accuracy because they can capture more nuances in the data. Yet, it’s notoriously difficult to understand how they arrive at a particular decision.</p><p>Finance use cases force us to find creative ways to maintain both. After all, if you’re telling a Cash Manager that “the system recommends reconciling this bank statement line with three open invoices,” you’d better be able to show exactly why this decision makes sense and how confident the system is. Without clarity on the ‘why,’ the recommendation might be dismissed outright, no matter how accurate the predictions are.</p><p>Building for financial professionals demands that we peel away the layers of the model’s reasoning. It’s not enough to say “trust the AI because the model says so.” Finance teams manage money and need to see evidence, like the logic or the rules the model used, plus the potential impact of any errors. If an AI-driven action is wrong, fixing it isn’t a simple matter of “delete and move on.” They may need to create a prior period adjustment, document the reason for that adjustment, track it in the audit log, and keep supporting documentation for regulators. This overhead means that financial teams are especially reluctant to adopt black-box AI, and even less so without supporting decision explanations.</p><h3>AI for Rules Management: Shifting from Transaction-Level to Rule-Level</h3><p>All of this led us to start exploring AI for managing rules, since rules are deterministic. Instead of showing AI predictions on each <em>transaction</em> level, why don’t we use AI on the <em>rule</em> level, which can then impact multiple transactions?</p><p>This question was driven by two main points:</p><ul><li><strong>Transaction-Level Challenges</strong>: Because finance data is often high-volume, even the most accurate transaction-level predictions can generate friction. Reviewing thousands of AI recommendations one-by-one can be overwhelming and cancel out time-saving benefits.</li><li><strong>Rules Are Deterministic</strong>: Finance teams often rely on “if-then” rules, such as “If the date range is within three days and the amounts are within 5% of each other, then suggest a match.” These rules are easy to audit and can be manually adjusted if necessary.</li></ul><p>Instead of building an AI that flags anomalies transaction by transaction, we focused on building a system that suggests new rules (or modifications to existing ones) based on historical data. For example, the AI might detect that 80% of a user’s reconciliations follow a pattern involving invoice amounts and date ranges, then propose a rule to automate future matches.</p><p>This approach offers two major benefits:</p><ol><li><strong>Time Savings</strong>: A single rule can automate thousands of transactions. Instead of verifying each transaction, the finance team verifies the rule itself — once — and sees the efficiency gains ripple across the entire dataset.</li><li><strong>Explainability</strong>: Because the rule is deterministic and explainable in plain English, the user can see exactly what conditions trigger an AI-driven match. If the system is wrong, the user adjusts or deletes the rule, and the next transactions will align with that updated logic.</li></ol><h3>Building Trust Through Transparency</h3><p>For one of our use cases, the main persona is a Cash Manager who usually spends hours every day reviewing bank statement lines and matching them to transactions in Workday (e.g., “This one line from Bank of America is actually a payment for these three customer invoices.”). By having AI review the historical data and analyze which transactions were reconciled in the past, there are patterns that emerge and the AI can recommend rules that would have reconciled those transactions; in the future, users won’t have to do it manually anymore.</p><p>To help finance professionals trust AI-driven rule suggestions, we implemented several “trust-building” features:</p><p><strong>Rule Impact Visualization</strong></p><p>As we tested with Cash Managers, we realized that transparency about the rule (e.g., Match if transactions are within this date range and this amount range) would not be enough — we needed to also include information about the <strong>impact </strong>of the rule (e.g., “This rule would have potentially matched 1,000 transactions.”) We provide a quick and easy way for users to “trust but verify” that the transactions should be matched, as well as a seamless way to edit the rule after reviewing the list of historical matches. While this feature starts with the user manually reviewing every rule before accepting them, we heard feedback from users that as the system gets smarter and as their trust builds, more and more automation can be built in.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Aiu6LCED3YpiJVD8" /></figure><p><em>Example of how a user views the impact of an AI-generated rule</em></p><p><strong>Version Control and Change History</strong></p><p>Finance teams need a “paper trail” not just for transaction data but also for the logic that generated those transactions. If a colleague edits a rule, or the AI proposes a new version, the product logs exactly what changed, who changed it, and why. Over time, we can compare AI-suggested changes to those made by humans to see which approach yields more accurate results.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*PUuyNoYW_8GeQ9Gp" /></figure><p><em>Example of how a user would be able to review the edits of an AI-generated rule</em></p><p><strong>Reporting and Auditability</strong></p><p>We created a centralized dashboard that shows all current rules, their status, the user or AI entity that created them, and any modifications since inception. This is especially critical in a finance environment, where everything must be traceable. <strong>But here’s the curb cut effect: this type of dashboard is also valuable for manually written rules.</strong> Even if a rule wasn’t AI-generated, the entire team now has a consolidated view of which rules exist, how well they perform, and when they were last updated.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*2S1Q0oHy9Vez0eKf" /></figure><p><em>Example of the report view for users reviewing rules created by humans and AI</em></p><p>This trust-building layer reassures finance professionals that our AI isn’t a black box. We’re making the internal logic visible and simple to adjust, which reduces friction for teams traditionally wary of “letting the machine decide.”</p><h3>Conclusion: Finance as the Ultimate Forcing Function</h3><p>Designing AI for Financials may seem daunting, but it’s also a golden opportunity to build solutions that benefit an entire organization, department, or company. Just as curb cuts opened up city streets to a broader range of people, financial-grade AI solutions that are anchored in transparency, auditability, and high-volume scalability will create a safer, more accessible, and ultimately more trustworthy environment for all types of users.</p><p>In a world where trust in AI is still far from universal, finance stands out as the ultimate proving ground. If we can satisfy the unyielding demands of finance users by ensuring our models are both accurate and transparent, then we can bring those same qualities to every other corner of the enterprise. The next time you step off a sidewalk onto a curb cut, think of finance-driven AI. Both serve as reminders that when we elevate our standards for one group, we often raise the bar for all.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c3a9d512816c" width="1" height="1" alt=""><hr><p><a href="https://medium.com/workday-engineering/building-ai-for-financials-c3a9d512816c">Building AI for Financials Products</a> was originally published in <a href="https://medium.com/workday-engineering">Workday Technology Blog</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[3 Ways to get Design, Dev, and Product Management to Work Together Effectively]]></title>
            <link>https://medium.com/product-labs/3-ways-to-get-design-dev-and-product-management-to-work-together-effectively-ce10b769c3b5?source=rss-30ce0bc97f06------2</link>
            <guid isPermaLink="false">https://medium.com/p/ce10b769c3b5</guid>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[balanced-team]]></category>
            <category><![CDATA[lean-startup]]></category>
            <category><![CDATA[agile]]></category>
            <category><![CDATA[startup]]></category>
            <dc:creator><![CDATA[Jenn Dearth]]></dc:creator>
            <pubDate>Fri, 17 Jun 2016 17:23:11 GMT</pubDate>
            <atom:updated>2017-05-01T19:23:16.669Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/701/1*t4x8L5X-J4SAe7YhYVRvwg.png" /></figure><p>A couple weeks ago, I had the honor of speaking on a panel at Boulder Startup week with my work colleagues whom I greatly admire.</p><p>Working at Pivotal Labs has taught me the true meaning of that quote, “Culture eats strategy for breakfast.” In order for multi-disciplinary teams to work effectively together, you have to start with your relationship with each other as people. It is all about fostering a culture of profound respect and trust. At Labs, we strive to structure all our engagements with a Balanced Team.</p><h4>What is a “Balanced Team”?</h4><p>It is summarized as three things:</p><ol><li>Co-located and dedicated to one product</li><li>Sharing the vision; no silos, no handoffs</li><li>Multi-disciplinary with high value on collaboration and iterative development</li></ol><p>What this looks like at each organization can vary, but in general it means one team of people per product. There are certainly exceptions for larger products (e.g., Netflix, Spotify, Google Chrome, etc.) and I will address this later on in the post.</p><p>The team is responsible for the product, is in communication daily, and feels a deep ownership over what is created. This removes a culture of blame, and if anything goes wrong, it falls on everyone on the team.</p><p>When we say “shared ownership” this is what it looks like for each role:</p><ul><li>Designers are the “Empathizer-in-Chief,” with an obligation to the team to understand the customer at an expert level, translate high-value needs into product, and have an important authority to prioritize customer problems.</li><li>Developers “Raise High the Roofbeams,” with an obligation to the team to envision the best possible solutions based on available technology, commit to the customer outcome, and have an important authority to measure outcomes over time.</li><li>Product Managers are “The Scales of Justice” with an obligation to the team to translate vision into execution by balancing customer needs and business value against feasibility, make fast and concrete decisions despite inadequate evidence and conflicting priorities, identify and ensure that the appropriate feedback loops are implemented to continuously de-risk the product, and have an important authority to delegate decisions to the appropriate persons on the team.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/995/1*7MN0XSEV0kET3AhMfrEbaA.jpeg" /><figcaption>Credits: Janice Fraser (<a href="http://www.slideshare.net/clevergirl">http://www.slideshare.net/clevergirl</a>)</figcaption></figure><p><strong>One thing to note</strong>: If you’re a two-person startup with one person playing two roles and other person playing one, you may be asking, “What if I can’t get a third person? I’m bootstrapping here!” This is completely doable, but it will be vitally important that the person who is playing a dual role is consciously balancing both voices. Just be aware of the risk having one voice muted by the other two if someone is not there consistently advocating for that third point of view.</p><h4>What are the benefits?</h4><p>Here’s our top three at Pivotal Labs:</p><ol><li>Get to a better outcome faster</li><li>Make your feedback loops tighter so you can avoid rework</li><li>Create a sense of shared ownership, trust, common understanding and empathy to work more effectively together</li></ol><p>Building products is hard. Every day you’re making hundreds of little decisions. When you’re working alone and you have a question about what direction to go in, you’ll often make a decision based on the available information at hand, your experience, and your assumptions of what the rest of the team wants. So when you’re working together, you get the benefit of everyone’s knowledge at hand (making the experience pool bigger!) and fewer assumptions.</p><p>So, if you’re working on a UX decision, dev can steer you to a more achievable solution. PM can help you remember what the business requirements are, and <strong>everyone</strong> can help to remember “What did Patty say in that user interview?!”</p><h4>How does it fit into the daily workflow?</h4><p>While the process is a little different depending on what stage of the product lifecycle you are in, we have two grounding meetings per week where everyone is together. If you’re a startup with remote resources, try at the very least to have these two touchpoints. If you can be face-to-face, all the better:</p><ol><li>Planning: Happens at the start of week to see what we will be doing and discuss items that may need clarification.</li><li>Reflection: Often called Retros, these happen at the end of week to see how things went, how everyone is doing, and what can we do to improve things. This is a safe and creative space where the team takes an honest look at how we are performing as a team.</li></ol><p>At Labs, we sometimes kick off a project with a short “Discovery and Framing” engagement that focuses on validating and invalidating the product’s riskiest assumptions. This helps with creating the first one or two weeks worth of backlog, and increases our confidence that we are not just building the thing right, but we are also building the right thing.</p><p>One common mistake (and we are sometimes guilty of it ourselves) is not bringing Dev into the conversation early enough during the beginning of the product lifecycle. Because PM and Design are usually the ones conducting user interviews and synthesizing feedback during the initial Discovery phase, Dev might not enter the picture until halfway through the Framing (or “solutioning”) phase.</p><p>Developers are often informed about what is just now technically feasible. Having the most relevant information about what is at the cusp of technology means you can validate or invalidate implementation assumptions. <a href="http://www.svpg.com/developer-powered-innovation/">Marty Cagan</a> recently wrote about this as well.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1023/1*ki9_rem4jo2Z5940LAgv5w.jpeg" /><figcaption>Sharing the stage with Nick Wade, Ethan Andrews and Erica Birkman</figcaption></figure><h4>So what are some next steps?</h4><p>When we came together to plan for the panel, our goal was to make sure the audience left with at least one actionable step they could take the very next day. Every change begins with a small step to integrate new ideas into your world.</p><ol><li><strong>Have meaningful conversations as often as you can</strong>: Train yourself to notice when you’re making decisions on behalf of another “voice”, and make a conscious effort to engage that team member in conversation before taking any other steps.</li><li><strong>Dissent, but also give each other the benefit of the doubt</strong>: It is important to remember that dissent is invaluable — you need a dissenter, even if you don’t agree with the specific dissent itself. Dissenters open up discussion and sharpen everyone’s thinking. There’s a great story about this in the book “<a href="http://www.amazon.com/Sway-Irresistible-Pull-Irrational-Behavior/dp/0385530609">Sway</a>”, and it even impacts how the Supreme Court is set up. At the same time, always remember that you are all on the same team. Listen well to your teammates — there is tremendous value from each voice.</li><li><strong>Value group facilitation over heroism</strong>: Don’t try to be a hero. As <a href="http://www.slideshare.net/clevergirl/2015-balanced-teams-product-management-engineering">Janice Fraser</a> puts it, we need to be good participants in a team dialogue that gets the right thing to happen without drama, <em>forever</em>. It’s unnatural, and we need to be taught how to do it, but we can do it! At the end of the day, reflect on how things went right or wrong, figure out what happened, devise a new next step, do it quickly, and do it without drama. This is only possible when we stop acting like heroes, and work as a team.</li></ol><p>At the end of the panel, I heard a really interesting question about what does Balanced Team look like across platforms or with really large products? I wasn’t satisfied with the answer I had, so I did some research and found this <a href="https://www.kennorton.com/newsletter/2016-02-24-bringing-the-donuts.html">article</a> by Ken Norton. It emphasizes tactically dividing up ownership across PMs <strong>around customers, not code repositories</strong>.</p><blockquote>Envision a feature you might want to build, such as rescheduling a doctor appointment. How many PMs need to be in the room when you talk about this feature? How many deciders? Is it one patient experience PM who has the final say? Or an iOS PM, an Android PM, an account system PM, a web PM, and a backend PM… you get the point.</blockquote><p>All in all, a great time. Huge thanks to the BSW volunteers, <a href="https://medium.com/u/bfa723af31d8">janice fraser</a>, <a href="https://medium.com/u/b4647cf19b34">Tessa Reuber</a>, <a href="https://medium.com/u/69474693dba8">Erica Birkman</a>, <a href="https://medium.com/u/489d45edf66c">Ethan Andrews</a>, <a href="https://medium.com/u/f0877a794573">Nick Wade</a>, <a href="https://medium.com/u/3e5573e31081">Matthew Heidemann</a> and <a href="https://medium.com/u/c9a23f66b276">Mike Gehard</a>!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ce10b769c3b5" width="1" height="1" alt=""><hr><p><a href="https://medium.com/product-labs/3-ways-to-get-design-dev-and-product-management-to-work-together-effectively-ce10b769c3b5">3 Ways to get Design, Dev, and Product Management to Work Together Effectively</a> was originally published in <a href="https://medium.com/product-labs">Product Labs</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[A Primer to Product Management]]></title>
            <link>https://medium.com/product-labs/a-primer-to-product-management-89bcecdb3247?source=rss-30ce0bc97f06------2</link>
            <guid isPermaLink="false">https://medium.com/p/89bcecdb3247</guid>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[startup]]></category>
            <dc:creator><![CDATA[Jenn Dearth]]></dc:creator>
            <pubDate>Sun, 10 Apr 2016 20:03:36 GMT</pubDate>
            <atom:updated>2016-04-25T15:09:32.719Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*1MSwa1_0aKsq0hh5FJElWw.png" /></figure><p>I was recently asked to speak at the<a href="http://www.meetup.com/Boulder-UI-UX-Group/events/228778361/"> Boulder UI/UX Meetup</a>, and had such an awesome time preparing for it that I wanted to write about the experience. <strong><em>Quick disclaimer: </em></strong>there are so many topics in the talk that I wanted to dive a bit deeper into but was constrained on time.</p><p>Product Management has been <a href="http://www.mindtheproduct.com/2015/10/history-evolution-product-management/">around for a while</a>, but the practice really changed for the software industry after the Agile Manifesto redefined how teams should work in 2001. Since its shift into software, the role of Product Manager varies from company to company so it was fun to explore the gifts of the internet and see whether or not certain Product Management practices differed by industry. By and large, most articles aligned with my expectations, but I felt that <a href="https://medium.com/@joshelman/a-product-managers-job-63c09a43d0ec#.rus6h9c59">Josh Elman’s statement about Product Management</a> was the most succinct and applicable to my experience.</p><p>Therefore, I framed my discussion around his statement:</p><blockquote>Help your team (and company) ship the right product to your user.</blockquote><h3>Helping Your Team</h3><p>The first part of the statement, “<strong>help your team</strong>” covers the more tactical practices, such as grooming the backlog, building the roadmap, managing ops, and writing good stories. The general takeaways are:</p><ul><li><strong>Write small stories</strong>: The size of the story should be the smallest possible while still delivering value to your end user. Smaller stories falls in line with Agile’s principle of people over process. Don’t spend all your time writing stories; the stories should instead be placeholders for conversation. Smaller stories also allow for shorter feedback loops so you know you have (or haven’t) added value for your user before you spend a bunch of time building something.</li><li><strong>PMs are not the “CEO of the Product”</strong>: I firmly believe that PMs should be facilitating discussions to bring the best ideas out of the team; they are not dictating direction. The roadmap is a team effort that is best articulated with a co-located, <a href="https://blog.pivotal.io/pivotal-labs/tech-talks/balanced-team-janice-fraser">balanced team</a>.</li><li><strong>Steward of Team Happiness</strong>: As a PM, one of your top priorities is making sure the team is happy. That could be achieved through a well groomed backlog (are all the assets that need to be attached to stories there?) or it could be making sure the team is not working weekends. This can take many forms, but you want to avoid this level of frustration from your team</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/350/1*fR-M8RK_L1icm2sbU-0gKQ.gif" /><figcaption>Avoid (as much as possible) team frustration</figcaption></figure><h3>(and Company)</h3><p>By adding “<strong>and company</strong>” as another component to helping your team, it highlights the strategic element of Product Management. It is just as important for the PM to understand the team’s goals as the company’s goals. Therefore, understanding the domain, identifying trends, and taking advantage of the right opportunities all fall under the PM’s purview.</p><h3>Ship</h3><p>The second part of the statement, “<strong>ship</strong>,” covers the PM’s responsibility to ship the product. Although it’s fun to dream about all the cool features, at some point, you have to get the product out the door. There’s a great talk that <a href="https://blog.pivotal.io/labs/labs/get-comfortable-shipping-imperfect-products">Lauren Gilchrist gave about shipping imperfect products</a>. She also has a great analogy of how we should all move away from this idea of being an expert, and move towards being a scientist. There are too many unknowns in the world, and the pace in which software disrupts industries means that you can’t know it all.</p><p>Therefore, the solution is shipping early and often in order to build the evidence towards what to build next. A false hypothesis is still valuable because it teaches you that your assumptions were wrong. Only the market could have taught you this, no matter what kind of expert you are.</p><h3>The Right Product</h3><p>“<strong>The right product</strong>” section covers the topic of defining MVP, testing, and measuring. I use the <a href="http://venturebeat.com/2014/11/11/minimum-viable-products-exceptional-products-and-cupcakes/">cupcake model</a> to talk through how MVP needs to have both <em>desirability</em> and <em>completeness</em>. However, MVP is just the first step of finding the right product, and I know the term itself is quite loaded. For startups, it may be easier to define because your customer is more easily defined. But for enterprise products, there are usually a lot more stakeholders and their definitions of desirability and completeness can vary depending on their goals.</p><p>Therefore, working together with your team, create different experiments to test out your riskiest assumptions. The right test can save both time and money without having to build out a full product. The following are different types of tests, but keep in mind that they each have different applications at different steps in the validation process. For example, surveys and Net Promoter Scores are only useful if you have a very large sample size and the assumption is very quantitative.</p><ol><li>Customer discovery interview</li><li>Data Mining</li><li>Surveys</li><li>Net Promoter Score</li><li>Concierge Testing</li><li>Wizard of Oz Testing</li><li>Product Analytics / Dashboard</li><li>Smoke Test</li><li>Published and validated Research</li><li>Solution Test</li></ol><p>For how we address measuring results, I use Ash Maurya’s framework from his new book,<a href="http://golean.ly/ref/q1200241"> Scaling Lean</a>. I appreciate how he approaches the typical “pirate metrics” (Acquisition, Activation, Retention, Referral, and Revenue) in a nonlinear way. He deconstructs the term “traction” into a handful of key levers so you know what to measure and can ignore the non-actionable data.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/778/1*mbHoof-zG6BfnyDBvXYaag.png" /><figcaption>Credits: Ash Maurya, <em>Scaling Lean</em></figcaption></figure><h3>To Your User</h3><p>The final part of the statement, “<strong>to your user</strong>,” highlights how to define your user, and how do you get your users to care about your product.</p><p>When you’re in the earlier stages of a product, making sure you’re focused on one or two user personas is important. This can help to design the right user experience. However, as products evolve over time, it’s important that the business thinks about <em>potential</em> customers as well. As you conduct your user research, keep in mind that current customers probably won’t give you feedback that will make their current implementation obsolete, so you have to be alert and ask around the “why”.</p><p>For example, an instructor at <a href="http://pragmaticmarketing.com/">Pragmatic Marketing</a> once shared a story where he and his wife built their own house in the late 90s. They were both technologists, so they actually wired their entire house with CAT5 internet cables. At the time, Cisco was just starting to explore expanding from the commercial router market for residential use. When they did a user interview with him, he told them he wanted two things: Faster speed and more ports. He never would have told them (wait for it)… WIRELESS. Building wireless routers made it so that the <em>potential</em> customers became customers.</p><p>The other story I shared in this section was the <a href="http://firstround.com/review/take-on-your-competition-with-these-lessons-from-google-maps/">story told by Bret Taylor</a> and Google Maps. My two takeaways from the article were:</p><ul><li>Familiarity can be a fatal weakness</li><li>Build a lens to allow users to see a new world rather than features to help them see an old world better</li></ul><p>I thought these were particularly insightful since oftentimes as PMs, we want to use other products to compare and contrast to our own. It can be a way to affirm our decisions or think that we have a shortcut to our users. Instead, we have to be deliberate about each experience we create, and we need to think deeply about user needs. What solutions are we providing that others are not?</p><p>That concludes my talk, and huge thanks to all the wonderful folks that were cited in this post (and BJ Cooper for his mentorship and all-around awesomeness). I welcome your thoughts on how your definition of Product Management may differ!</p><p>Here is a more “<a href="https://docs.google.com/presentation/d/17BYikWNamh70TZ1Qnpjxqn8dkGgQ51OJuYIFOL0Py4Q/edit?usp=sharing">sharable</a>” version of the deck in case you are interested.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=89bcecdb3247" width="1" height="1" alt=""><hr><p><a href="https://medium.com/product-labs/a-primer-to-product-management-89bcecdb3247">A Primer to Product Management</a> was originally published in <a href="https://medium.com/product-labs">Product Labs</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>