<?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 Ruairi McGuinness on Medium]]></title>
        <description><![CDATA[Stories by Ruairi McGuinness on Medium]]></description>
        <link>https://medium.com/@ruairi.mcguinness?source=rss-a024fd07b593------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*92YifzMbU-3cH_oyw_SniA.jpeg</url>
            <title>Stories by Ruairi McGuinness on Medium</title>
            <link>https://medium.com/@ruairi.mcguinness?source=rss-a024fd07b593------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 17:55:30 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@ruairi.mcguinness/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[Agents Climb Hills, But They Don’t Pick Them]]></title>
            <link>https://medium.com/@ruairi.mcguinness/agents-climb-hills-but-they-dont-pick-them-edfa654358b9?source=rss-a024fd07b593------2</link>
            <guid isPermaLink="false">https://medium.com/p/edfa654358b9</guid>
            <category><![CDATA[agentic-ai]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[product]]></category>
            <category><![CDATA[experimentation]]></category>
            <dc:creator><![CDATA[Ruairi McGuinness]]></dc:creator>
            <pubDate>Thu, 22 Oct 2026 15:46:07 GMT</pubDate>
            <atom:updated>2026-10-22T15:46:07.797Z</atom:updated>
            <content:encoded><![CDATA[<p><em>“We can’t run three versions, we don’t even have time to build one.”</em> That sentence has killed more good ideas than any competitor has.</p><p>I’ve said it. I’ve had it said to me by engineering leads who were being completely honest. It wasn’t an excuse. It was simple maths. Every version cost real time, real people and real opportunity cost. So teams aligned on one version, built it as well as they could and called the agreement strategy…but it was an educated guess on a roadmap slide.</p><p>That constraint is gone.</p><p>Building three versions of something now costs less than building one did three years ago. That changes more than delivery speed, it changes what a product person is actually there for.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-yTGRGujxWxWFynk7697nA.jpeg" /><figcaption>Image courtesy of Sadat Alam Protik via Unsplash</figcaption></figure><h4><strong>The old job was betting once and being right</strong></h4><p>When experimentation was expensive, product management organised itself around one moment: <strong>the decision</strong>. Research, a few interviews, a business case, some sleepless nights and then you picked. You simply couldn’t climb three hills to find out which one had the moat.</p><p>So our craft grew up around de-risking the bet before it got made. Frameworks, discovery tactics, scoring and validation gates before a single line of code. All of it exists because the bet was expensive and you only got one go.</p><p>Not long ago, I built a validation gate into a strategy for exactly this reason. We wanted to commit several million pounds to a platform build on a hypothesis nobody had actually tested. Before anyone touched the budget, we ran the thing by hand. Humans simulated what the system would eventually do for a small cohort of users, and we watched whether it moved the numbers we cared about. If it passed we jump in and build but if it failed we got to learn some important lessons and save valuable time and millions in costs.</p><p>That was the right move for that world.</p><h4><strong>What replaces the single bet</strong></h4><p>Nobody selling AI tools will say this plainly: an agent writing code faster for your product doesn’t make you better at knowing what to build. It will optimise whatever you point it at, tirelessly, to the top of that hill but it will never stop and ask if you picked the wrong one. I don’t expect a better model to fix that. Picking and climbing are different kinds of work.</p><p>What changes is how often you get to pick.</p><blockquote>Instead of one gate before a big commitment, you now get a loop. Ship three versions. Let them run. Read what happened. Choose the winner, then ship the next three versions based on that winner (not on a hypothesis you wrote three months ago before anyone had touched the product). Rinse and repeat.</blockquote><p>That sounds like it shrinks the job, but it does the opposite. Someone has to decide what the next test proves, what counts as signal and what is noise and when a result means <em>get off the hill and pick another</em>. That isn’t an engineering call and it was never really a research one. It’s a product call, now made weekly instead of once a quarter.</p><p>And signal doesn’t need a big sample. Run three versions for two weeks and you won’t get statistical significance, but you will find out which variant people stop complaining about, which variant they work around and which variant nobody mentions. That’s still a read, as long as you decided before you started what you were looking for.</p><p>The stakes per decision drop but the number of decisions goes up exponentially. At every branch the question is the same: <em>what does this result mean, who does it threaten and what’s the honest read versus the comfortable one.</em></p><h4><strong>You won’t see the payoff from the bottom of the hill</strong></h4><p>Partway up, you often don’t know what you’re actually building. Plenty of product moats that look like deliberate long-term advantage weren’t planned that way. People simply used them a lot and the moat turned up later. The network effect, the proprietary data, the switching cost; usually a byproduct of small decisions made in motion, not a thesis someone had on day one.</p><p>So picking faster doesn’t mean seeing the whole hill. It means finding out sooner that you were on the right one, and finding out cheaply that you were on the wrong one.</p><h4><strong>What actually changed</strong></h4><p>Not the models getting smarter, but something far more mundane. The cost of finding out you’re wrong dropped far enough that <em>“we don’t even have time to build one version”</em> stopped being said.</p><p>Product and design sit nearer the centre of this than they used to, because someone still has to say what the next test should prove and be right about it often enough that the loop learns something. The agents will climb any hill you point them at, every time. Choosing which hill is the part nobody has successfully automated.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=edfa654358b9" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Nobody’s Born With Product Sense]]></title>
            <link>https://medium.productcoalition.com/nobodys-born-with-product-sense-924024112f69?source=rss-a024fd07b593------2</link>
            <guid isPermaLink="false">https://medium.com/p/924024112f69</guid>
            <category><![CDATA[strategy]]></category>
            <category><![CDATA[mentoring]]></category>
            <category><![CDATA[product]]></category>
            <category><![CDATA[product-management]]></category>
            <dc:creator><![CDATA[Ruairi McGuinness]]></dc:creator>
            <pubDate>Sun, 27 Sep 2026 04:11:37 GMT</pubDate>
            <atom:updated>2026-09-27T05:17:24.988Z</atom:updated>
            <content:encoded><![CDATA[<h4>If product sense is built rather than born, how do you build it faster, in yourself and in the people around you?</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_g6aLNQlFzg8nF2CajpDAA.png" /></figure><p>I’m lucky to know several people who went to art school and work in advertising, including my wife. All of them are among the most interesting people I know, but I noticed a long time ago that what unifies them is how quickly they form a judgement about what they experience. No hedging, no <em>“let me think about this”</em>, just an immediate, confident read.</p><p>I used to think that was about a certain personality type, but I’ve since realised how much of it comes from reps: repeated opportunities to make a judgement and learn what happened next. Years in a fast-moving industry help you develop an eye for what works. The speed isn’t necessarily recklessness; it can be the accumulated weight of having made that call a thousand times before. I’ve noticed the same thing happening in myself, but over a longer and stranger route.</p><p>In product work, that judgement helps you recognise what matters to users, what might be causing a problem and which decisions deserve attention. It gives you somewhere to begin. It doesn’t remove the need to check.</p><h3>Where the Reps Came From</h3><p>Two and a half decades of starting up and running businesses taught me to read a room, a customer, a supplier or a team member faster than any qualification could have. Slow reps, but real stakes.</p><p>Product consultancy compressed that learning into something much faster. As a new in-house product hire, you may get a few months to find your feet. A consulting engagement rarely gives you that luxury. You’re expected to arrive, diagnose the problem and start saying <em>“I’ve seen this before; here’s what I think is actually going on”</em> within a week or two, not a quarter. You need a view fast, even if it takes weeks more to earn the right to act on it.</p><p>Then there was building teams. At one product agency, I managed people, wrote the career tracks that defined what each product level was expected to do, and helped make hiring decisions. Hiring combines a relatively quick judgement with a long feedback loop: you form a read on someone in an hour, but understanding how well you judged their abilities takes much longer.</p><p>Writing those tracks forced the team to agree on what good actually looked like at each level. That is harder than it sounds when so much of what separates a good PM from a great one is judgement: which questions they ask, which trade-offs they recognise and when they decide there is enough evidence to act.</p><p>But accumulating experience is only part of the work. You also have to keep examining what it has taught you.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fr8ZjNGGRi1bqkm0WoBwrw.png" /></figure><h3>Experience Needs to Be Questioned</h3><p>There’s an old line about having strong opinions, loosely held. I think it’s the whole job once you unpack it properly.</p><p>The reps help you reach a fast, confident read. That’s the strong opinion part: having a clear position that other people can examine and use. An executive facing a decision needs your best assessment, along with an honest explanation of what remains uncertain.</p><p>But strength on its own curdles fast. I’ve watched experienced product people stop checking their own read and start defending it instead, mistaking familiarity or agreement in the room for correctness. That’s when good instinct turns into stubbornness.</p><p>The real discipline is holding both at once. Take a position, then deliberately look for reasons you might be wrong before asking anyone else to act on it. People early in their careers may be open to being wrong but lack the experience to form a clear view. Experienced people can have the opposite problem: a ready opinion and little willingness to reconsider it. The useful combination is being quick to a position and still willing to have it overturned.</p><blockquote>One way to practise is to write down your initial view, what you are assuming and what evidence would change your mind. Afterwards, return to it. What did you spot correctly? What did you miss? Without that step, it’s easy to remember the calls you got right and explain away the ones you didn’t.</blockquote><h3>Putting a Read to Work</h3><p>A while back, working through a diagnosis on a client engagement, I had a strong read within the first week or two on what was going wrong and why. The business had been through a period of fast growth, and every time it hit a problem, the answer was a new standalone solution. Each one was launched, celebrated and counted as a win.</p><p>My read was that no single solution explained the problem. Each made sense on its own, but together they had become a stack of separate tools, each addressing a narrow need. Nobody in a product leadership role was looking across the whole experience, and every launch added one more product that users had to remember, check and log into.</p><p>The business was measuring success through what it added, but nobody was measuring what those additions were like for the people using them. It felt obvious, which is exactly what made me suspicious of it.</p><p>Instead of immediately building my case, I spent the next few weeks trying to break it. Was the problem really how the tools fitted together, or were some simply badly delivered projects? Did the process maps I’d been handed match what happened in real life, or did they show how the work was supposed to happen? I checked my read against the people doing the actual work. And I kept coming back to one uncomfortable question: if this was so obvious to me, why had nobody said it out loud?</p><blockquote>After those checks, I judged that I had enough support for the diagnosis to put it forward. That didn’t make it proven. It made it a stronger basis for a decision than my first impression alone.</blockquote><p>That’s the discipline in practice. The fast read gets you to a position quickly, but the work of testing it earns you the right to ask others to act on it. Skipping that work is how confident people end up defending conclusions they never properly examined.</p><h3>From Diagnosis to Shared Responsibility</h3><p>The case itself was simple. The guiding principle became reduction. Every new request, feature or tool had to answer one question before anything else: does this reduce the load on the people who have to use it, or add to it? If it added more steps, more switching between tools or more things to remember, it needed a very good reason to exist.</p><p>I built the same question into the AI agents I was writing for the client, making it part of how the team assessed new ideas rather than leaving it only in a strategy document.</p><p>When I put the diagnosis to the leadership team, nobody argued. I took that as recognition of the problem, but agreement in a room is only a starting point. It doesn’t tell you whether the proposed changes will help users or whether anyone will follow through.</p><p>What happened next is the part I don’t read about often enough in articles like this. I handed the strategy to someone inside the business. I’m hopeful that it will take hold, but I suspect it might quietly fade. A handover can put a document in someone’s hands without giving them clear responsibility, time or authority to carry it forward.</p><p>I can describe the reasoning behind the strategy. I can’t yet claim that it produced a better experience for users. That would require seeing what changed and whether their work became easier.</p><blockquote>Product artefacts get handed over and filed. People who’ve learned to see the problem for themselves take that ability with them. That is why developing their judgement matters as much as sharing my own.</blockquote><h3>Making Judgement Visible</h3><p>If product sense is built through reps, then the most useful thing I can do for someone early in their career isn’t hand them my conclusions. It’s deliberately give them opportunities to build their own.</p><p>With the product people I’ve mentored, the approach that landed most effectively was the five-minute debrief. Straight after a discovery call or stakeholder conversation, we’d jump on a quick call and walk back through what had just happened. It wasn’t a lecture, and it didn’t start with me talking. I’d open with two questions: <em>“How do you think that call went?”</em> and <em>“What did you notice me do, and why do you think I did it?”</em> Their answers showed me what they were already seeing and what they were missing.</p><p>Then came the explanation: <em>“You might not have noticed, but I framed the problem more broadly there so the stakeholder had room to think it through, rather than simply respond to my suggested answer.”</em> Small choices like that can be invisible in the moment, but they become learnable when you explain them straight afterwards. Do it enough times and the person you’re mentoring starts spotting them before you point them out.</p><p>The debrief can also go beyond how the conversation was handled. What did we learn about the problem? Did we hear evidence of a need, or a preference for a particular solution? What would we need to check before changing our plans? Those questions connect the conversation to the decisions that follow.</p><p>If you manage junior product people, try it after your next stakeholder call. Five minutes, a couple of questions, then explain what you did and why. Give them room to question your interpretation, too. The point is to help them develop judgement, including the confidence to disagree with yours.</p><blockquote>I don’t pride myself on building people who trust my judgement. I pride myself on building people who don’t need to.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Ya8XlJa1zS0WUXlyQgYDPQ.png" /></figure><h3>Building a Path for Others</h3><p>Here’s a pattern I’ve seen across multiple clients: organisations expect better product thinking without deliberately developing it. Project Managers move into Product Owner roles without support in deciding what to prioritise and why. Product Owners become Product Managers without opportunities to develop strategic thinking. Business Analysts are left to work out unfamiliar responsibilities because nobody has built them a path.</p><blockquote>Those people bring useful skills. The problem is assuming that a change of title supplies everything else they need.</blockquote><p>Training helps, but development also needs to happen in the work itself. People need opportunities to make decisions, explain their reasoning and receive feedback on what happened. As their responsibilities grow, those opportunities need to grow with them, from assessing an individual request to weighing priorities across a whole product.</p><p>Without that support, organisations leave the development of judgement to chance, then wonder why the same mistakes keep returning.</p><h3>What I Want to Leave</h3><p>Success, for me, means leaving a client’s product team in better shape than I found it, with more useful ways of working and sharper thinking. People who understand not just what to decide but how they got there are better positioned to keep making good calls long after I’ve moved on to the next problem.</p><p>That’s the whole thesis, really. Product sense isn’t a trait some people are born with and others aren’t. It’s built, through reps, under pressure and checked constantly against the likelihood of being wrong. My job is not to be the person in the room with the fastest correct answer, it’s to make sure I’m not the only one who can find it.</p><p><em>Many thanks to </em><a href="https://medium.com/@tremisskeete"><em>Tremis Skeete</em></a><em>, Executive Editor of Product Coalition, for his invaluable contributions to this story. A special shout-out to founder </em><a href="https://jaystansell.medium.com/"><em>Jay Stansell</em></a><em> for creating an environment that enhances product management education.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=924024112f69" width="1" height="1" alt=""><hr><p><a href="https://medium.productcoalition.com/nobodys-born-with-product-sense-924024112f69">Nobody’s Born With Product Sense</a> was originally published in <a href="https://medium.productcoalition.com/">Product Coalition</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What Three Seconds of Music Taught Me About Product]]></title>
            <link>https://medium.productcoalition.com/what-three-seconds-of-music-taught-me-about-product-7f519512a1dd?source=rss-a024fd07b593------2</link>
            <guid isPermaLink="false">https://medium.com/p/7f519512a1dd</guid>
            <category><![CDATA[creativity]]></category>
            <category><![CDATA[strategy]]></category>
            <category><![CDATA[product-thinking]]></category>
            <category><![CDATA[lateral-thinking]]></category>
            <category><![CDATA[product-management]]></category>
            <dc:creator><![CDATA[Ruairi McGuinness]]></dc:creator>
            <pubDate>Mon, 17 Aug 2026 00:23:07 GMT</pubDate>
            <atom:updated>2026-08-17T00:23:07.509Z</atom:updated>
            <content:encoded><![CDATA[<h4>What if the best way to make better product decisions is to give yourself fewer options?</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*N--HyZjHAHkwianp5tthHg.png" /></figure><p>About ten years ago, an ad agency commissioned me to write music for a global campaign for one of its clients. The process ran for two or three months, and in the end my piece wasn’t used. That’s fine, that’s the territory.</p><blockquote>What remains with me wasn’t the rejection. It was what the process did to my thinking.</blockquote><p>Writing for a 30-second ad, with 15 and 10-second versions, means obsessing over density: how much can happen in any given second. I’d spend an entire session on three or four seconds of audio until I couldn’t see the wood for the trees.</p><p>And eventually I realised the problem wasn’t that I didn’t have enough options. It was that I had far too many.</p><h3>Infinite Choice Becomes the Problem</h3><p>In modern music production, every synth and every sound anyone has ever recorded is available to you instantly.</p><p>That sounds freeing. It isn’t always. Sometimes it’s a trap.</p><p>So I built a wall. An actual set of rules, and ones I still use ten years later. I’ll get to exactly what they are in a moment, but the important thing is that the constraint didn’t shrink the work.</p><p>It made the work faster and more honest. More importantly, it stopped me disappearing into a single synth patch for six hours and calling that progress. Treating constraint as a generative tool rather than a limiting one takes a mental leap at first. It requires a healthy dose of glass-half-full thinking: looking at what you’re stuck with and deciding to work inside it rather than constantly wishing the boundary wasn’t there.</p><p>Better still, sometimes you can build the constraint yourself. Make the hard calls harder and earlier. Descope the product before you’re forced to. Reduce the available moves deliberately. Every wall you build pushes you into a decision you might otherwise keep avoiding.</p><p>I think this matters more now than it used to. AI has made “anything is possible” a reasonable starting assumption for a lot of product conversations, and it has shifted stakeholder expectations about what teams should be able to produce.</p><blockquote>But possibility isn’t the same thing as direction. When almost anything can be built, the harder question becomes what deserves to be built at all. And without some kind of constraint, it’s easy to drift toward the obvious solution, the familiar pattern, or whatever the model already knows how to produce.</blockquote><p>Just because you can build something doesn’t mean you should.</p><p>Constraint is one way of stopping “what’s possible” from collapsing into “what’s easy and predictable.”</p><h3>Twenty Cards, One Rule</h3><p>I started noticing the same pattern outside music. In a previous role as Product Lead, I was lucky to report to a genuinely excellent CPO. Part of the remit she gave me was running a weekly knowledge-share for the Product/UI/UX team.</p><p>One Friday I curated around 20 cards from Brian Eno and Peter Schmidt’s Oblique Strategies deck, chosen deliberately for their relevance to product or design work. Each person pulled one blind and had to think out loud:</p><p><em>How does this apply to what I’m actually working on right now?</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*XhsUKKGy3PnHFVAp5Cj_Hw.jpeg" /><figcaption>Oblique Strategies © 1975 Brian Eno and Peter Schmidt</figcaption></figure><p>The UX folks in the room had done exercises like this before at art school, so it was a familiar muscle for them. For most of product, it wasn’t. But it landed anyway.</p><p>The cards were interesting, but what mattered was the mechanism. Instead of asking people to stare at an open problem and generate more possibilities, we gave them one arbitrary direction and asked them to think inside it.</p><p>People leaned into it, which is why I later introduced Edward de Bono’s Six Thinking Hats to the same team. We eventually used Six Thinking Hats in a real discovery process for an in-house product we went on to build.</p><p>Different method, same basic instinct:<em> Give a team a narrower set of moves and see what happens inside those walls.</em></p><h3>The Rules That Stop Me Second-Guessing</h3><p>Which brings me back to the walls I built for music.</p><p>No piece with more than eight tracks.</p><p>No layered drums.</p><p>No layered bass or chords.</p><p>And four fixed stages, always in the same order: creation, where nothing is a bad idea yet; curation, where the material gets cut down to what actually sits together; mixing, where nothing new gets created and I can only shape what’s already there; and mastering, where every three to five tracks get prepped as a release.</p><p>This wasn’t an Eno card. It was <em>mine.</em></p><p>And building it wasn’t really about making music harder. It was about removing my own ability to keep second-guessing every decision. The rules meant that at some point I had to stop asking what else I <em>could</em> add and start deciding whether what was already there was good enough.</p><p>Ten years on, it’s still the only way I make music.</p><h3>When the Constraint Becomes the System</h3><p>Eventually I noticed that I keep building systems that force this kind of thinking.</p><p>So I built one for product thinking directly:<em> Ephemeral Wisdom.</em></p><p>Refresh and you get an insight. It fades. Pull the counterpoint, if you want it, and an equally valid opposing view appears. Then you have to decide where you actually stand.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kNIbgjGGe5VCEG4cSJHWKA.jpeg" /><figcaption>Ephemèral Wisdom © 2026 Ruairi McGuinness</figcaption></figure><p>I use it most mornings. Two or three cards before anything else, just sitting with the friction of arguing against myself for a minute or two.</p><p>Last week, while I was halfway through drafting this piece, it served up:</p><p><em>“Embrace constraints as creative fuel.”</em></p><p>Against:</p><p><em>“Remove constraints to enable innovation.”</em></p><p>That was almost annoyingly well timed. But it also raises the obvious question. If constraints can improve thinking, and removing constraints can improve thinking too, how do you know which kind you’re dealing with?</p><h3>So What Separates a Productive Constraint From a Stupid One?</h3><p>A useful constraint removes options without removing the thing you’re actually trying to learn. That’s the difference.</p><p>“Test this in five shops next week” is useful because it forces decisions about scope, audience and what absolutely needs to exist, while preserving the underlying question you’re trying to answer.</p><p>“Build it in two days” might be useful.</p><p>Or it might just guarantee a bad test.</p><p>The constraint itself isn’t clever. What matters is what it forces you to confront.</p><p>The best constraints tend to do three things: they make a decision unavoidable, they leave enough room for good judgement inside the boundary, and they don’t quietly invalidate the outcome.</p><p>A hard deadline can create focus. But if it means you’re testing something so compromised that nobody would reasonably use it, you haven’t really learned anything.</p><p>Limiting a design palette can create coherence. Limiting it to the wrong things because somebody happens to like blue is just taste pretending to be discipline.</p><p>A good constraint reduces optionality, not thought. The eight-track rule I use in music doesn’t tell me what to write. It stops me solving every problem by adding another layer.</p><p>The three-slide rule I use for stakeholder decks doesn’t tell me what to say. It forces me to decide what matters enough to survive the cut.</p><p>That’s the test I keep coming back to in product work too:</p><p><em>Does this wall force a better decision, or does it simply make the room smaller?</em></p><p>If it’s the latter, knock it down.</p><h3>Bring the Wall Back to Product</h3><p>The tricky thing is that work rarely gives you enough room to fail cheaply or quietly while you learn this habit from scratch.</p><p>So look somewhere else first. Think about whatever hobby or habit you disappear into on a Saturday. Somewhere in there, you’ve probably already built a wall for yourself on purpose without ever calling it one.</p><p>A runner chooses a distance. A photographer chooses a lens. A cook decides to make dinner from whatever happens to be left in the fridge. The limitation becomes part of the work.</p><p>Notice where you already do this well, then bring the same discipline deliberately back into your product thinking.</p><p>In discovery, don’t always wait for someone to hand you a limitation. Go looking for one yourself.</p><p>On an innovation team I worked with recently, a plan was spinning across a dozen possible directions at once. The question that cut through all of it wasn’t:</p><p><em>“What could this become?”</em></p><p>It was:</p><p><em>“What’s the least we can do to test this in five shops next week?”</em></p><p>That one simple constraint had more impact than another five workshops would have.</p><p>Not every wall is one you get to choose, either. On an ecommerce platform I worked on, incoming FTC regulation meant we had to let customers cancel subscriptions in a single click: no retention flow and no (<em>ahem</em>) friendly delay. Nobody on the team asked for that constraint.</p><p>But once it existed, the job was fundamentally the same as with any self-imposed wall: stop spending energy wishing the boundary wasn’t there and figure out what good looks like inside it.</p><p>Sometimes the regulator builds the wall for you. Sometimes the budget does. Sometimes it’s technology, time, headcount or the five shops you can actually reach next week.</p><p>The skill is the same either way. Give design a limited palette. Give UX copy a hard word count and watch what that does to the thinking. Give a discovery effort a smaller audience, a shorter window or a narrower question. Give yourself clear and difficult limitations.</p><p>For stakeholder decks, I’ve long aimed for three slides, three bullets a slide, three words a bullet.</p><p>I’ve never actually managed it. But that almost doesn’t matter. It’s still the best forcing mechanism I know for making myself say the important thing plainly instead of hiding behind more words.</p><p>You didn’t need four years at art school to learn to think this way. You just need a rule strong enough to force a decision, enough judgement to know whether the rule is helping, and the willingness to notice that you’re probably already better at working inside walls than you think.</p><p><em>Many thanks to </em><a href="https://medium.com/@tremisskeete"><em>Tremis Skeete</em></a><em>, Executive Editor of Product Coalition, for his invaluable contributions to this story. A special shout-out to founder </em><a href="https://jaystansell.medium.com/"><em>Jay Stansell</em></a><em> for creating an environment that enhances product management education.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7f519512a1dd" width="1" height="1" alt=""><hr><p><a href="https://medium.productcoalition.com/what-three-seconds-of-music-taught-me-about-product-7f519512a1dd">What Three Seconds of Music Taught Me About Product</a> was originally published in <a href="https://medium.productcoalition.com/">Product Coalition</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What Watching People Taught Me About Product Development]]></title>
            <link>https://medium.productcoalition.com/what-watching-people-taught-me-about-product-development-79660fe5ce12?source=rss-a024fd07b593------2</link>
            <guid isPermaLink="false">https://medium.com/p/79660fe5ce12</guid>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[product-sense]]></category>
            <category><![CDATA[product]]></category>
            <category><![CDATA[entrepreneurship]]></category>
            <category><![CDATA[data]]></category>
            <dc:creator><![CDATA[Ruairi McGuinness]]></dc:creator>
            <pubDate>Sat, 11 Jul 2026 19:45:26 GMT</pubDate>
            <atom:updated>2026-07-12T18:32:30.985Z</atom:updated>
            <content:encoded><![CDATA[<h4>And what running coffee shops taught me about interpreting data.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*e9PCy-v5ovKADCmThUjS8w.png" /></figure><p>I’ve led product teams and brought digital products to market across the US, Europe and Middle East and built product strategy for £2bn businesses, but so much of my product sense comes from owning a small chain of coffee bars.</p><p>The first Coolaboola Coffee site opened in 2006 and the following year made The Independent’s UK’s Top 50 Coffee Shops.</p><p>Within 48 hours, we knew we had found something people wanted. The right product, right price, right part of the city, the right moment just as third wave coffee was cresting, so we surfed it.</p><h3>The Bet That Looked Right on Paper</h3><p>Our second Coolaboola site was Central Station in Newcastle. On paper it was an obvious bet. The official footfall shared by the landlord was almost ten times our first site and the demographics looked great. I spent time watching the footfall patterns on site, but I didn’t want to acknowledge what my gut was telling me. Because the data is always correct, right? I placed a bet with our retained profits and our reputation.</p><p>What I failed to openly acknowledge was a nuance that no footfall figure will ever surface. In public transport there’s a critical difference between being a destination and being a point of departure. When people are moving through a space to catch a train, they think about time completely differently. They’re not in a <em>treat-yourself</em> mindset, they’re in an<em> afraid of missing the connection</em> mindset. We weren’t selling coffee to people arriving somewhere, we were selling it to people leaving, and they didn’t want to risk a delay by ordering a coffee at that exact location.</p><p>I can’t blame anyone else for that (although I do still question the validity of those footfall figures). It was my business, my call and my read of the data and I got them wrong.</p><p>Within two weeks of opening I knew we weren’t going to hit product market fit at that location. What we could do was make enough to cover costs, ensure the site washed its own face, and move on to site number three. Which is what we did.</p><p>There’s a real difference between that (figuring out how to keep the lights on after a bet doesn’t land) and running a post-mortem on a 2.03.17 app release. Both are legitimate forms of learning, but only one requires you to open up again at 06:00 the next morning regardless of how you feel about it. I’m not criticising solid product work here, but it’s an observation about what gets burned in when the consequences are that real.</p><p>That experience taught me something I’ve never unlearned. Data tells you what happened, but it rarely tells you ‘<em>why’…a</em>nd the ‘<em>why’</em> is almost always a person doing something the system didn’t expect.</p><h3>When the Numbers Were Right but the Story Was Wrong</h3><p>Years later I was working on a customer-facing retail tech product. The data was telling a strange story; user adoption varied wildly across shop locations with some performing significantly better than others at peak times. The variance was too wide for me to ignore but the numbers across multiple sources all tallied up. I went to one of the locations, ordered some food and watched what unfolded. Before I left I had a conversation with staff about the new tech solution; <em>“it’s a nightmare…”,</em> someone told me within minutes.</p><p>The nightmare wasn’t the technology solution though, it was the operations; they weren’t setup to succeed with the additional order throughput so staff had found a way to disable the system during their busiest periods because it was generating additional sales and therefore putting more pressure on the team, not less.</p><p>The data had been accurate, but the story it was telling was completely wrong.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Z9ISaMJ4oQztUoA3NNUpRg.png" /><figcaption>Source: Coolaboola Coffee</figcaption></figure><h3>The Photocopy of a Photocopy</h3><p>A couple of years back I worked with an early-stage startup, who had an excellent customer service agent who read every review, triaged every support ticket and passed summaries up to the leadership team. They weren’t ignoring their users, they just had made assumptions on what the users <em>actuall</em>y needed.</p><p>Every human handoff degraded the signal; specific complaints became general themes and emotional language got edited out. By the time user feedback reached a product decision it had been summarised twice, paraphrased once and was competing with fourteen other priorities. They were reading a photocopy of a photocopy.</p><p>We fixed it by connecting the support system and app store reviews into a single pipeline, and set up ML tooling to detect patterns across hundreds of data points with the original language preserved. Suddenly we could see not just what was being complained about but how often, in what language and alongside what other issues. The data was always there, it was the signal that was dying in translation.</p><p>Sometimes the problem isn’t that you’re not watching people closely enough, it’s that you’ve got systems that don’t let good signal survive the journey to a decision. But that can be fixed once you understand what gets lost without it.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qXPav9ZPg1zCZq-fx2o_dg.png" /></figure><h3>Sometimes You Have to Leave the Dashboard</h3><p>I didn’t learn these instincts at a product conference or from a framework. I learned them standing in a train station in 2007, watching people walk past something I’d built for them and realising that the numbers had not lied to me. I had simply expected them to answer a question they could not answer on their own.</p><p>The tools available to product managers are becoming more powerful, and we are all changing how we work. We can build and test products faster than ever, but the judgement required to understand what the data is really showing us has not changed. If anything, moving faster makes that judgement even more important.</p><p>When I see an unexplained variation in the data, I now resist creating an explanation from the dashboard alone. I go to the environment where the behaviour is happening, watch how people interact with the product and speak to those closest to the work. I then compare what I observe with the assumptions behind the numbers.</p><blockquote>The goal is not to replace data with anecdotes or allow one conversation to outweigh a broader pattern. It is to uncover the context needed to interpret the data correctly.</blockquote><p>Sometimes that context is a customer worried about missing a train. Sometimes it is a member of staff disabling a system because the operation cannot handle the demand it creates. The dashboard can show you that something is happening, but it cannot always explain why.</p><p>Before you draw too many conclusions from the data, find a way to watch what is actually happening. The numbers will still be there when you return, but you will have a deeper and more useful understanding of what they mean.</p><p><em>Many thanks to </em><a href="https://medium.com/@tremisskeete"><em>Tremis Skeete</em></a><em>, Executive Editor of Product Coalition, for his invaluable contributions to this story. A special shout-out to founder </em><a href="https://jaystansell.medium.com/"><em>Jay Stansell</em></a><em> for creating an environment that enhances product management education.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=79660fe5ce12" width="1" height="1" alt=""><hr><p><a href="https://medium.productcoalition.com/what-watching-people-taught-me-about-product-development-79660fe5ce12">What Watching People Taught Me About Product Development</a> was originally published in <a href="https://medium.productcoalition.com/">Product Coalition</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Burn The Maps]]></title>
            <link>https://medium.com/@ruairi.mcguinness/burn-the-maps-61f15e1a642b?source=rss-a024fd07b593------2</link>
            <guid isPermaLink="false">https://medium.com/p/61f15e1a642b</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[strategy]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[product]]></category>
            <dc:creator><![CDATA[Ruairi McGuinness]]></dc:creator>
            <pubDate>Fri, 15 May 2026 16:15:50 GMT</pubDate>
            <atom:updated>2026-09-24T12:43:33.595Z</atom:updated>
            <content:encoded><![CDATA[<p>You’ve done things right. You’ve hired smart, experienced, credible product people, but somewhere between the offer letter and the third sprint review they disappeared into the machine.</p><p>They’re running ceremonies, grooming backlogs, writing tickets, chasing dependencies. Doing exactly what the system asked them to do. Just not what you hired them for.</p><blockquote>The honest diagnosis: most organisations don’t actually know what they hired for. They know they “need a PM”, so they pattern-match to the last one they had. But the last one was primarily doing admin work and that’s probably why they left.</blockquote><p>I’ve watched senior PMs spend three months refining acceptance criteria for features they didn’t choose and didn’t even agree with, in organisations that called this “product strategy.” Nobody set out for it to be this way and everyone is responding rationally to the system they’re in. That’s what makes it hard to see.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*iQYUer5yE7HJmUl_WveOlA.png" /><figcaption>The map we were given</figcaption></figure><h3><strong>How we got the map wrong</strong></h3><p>Back in the day, agile consultants needed a way to explain who owned the work. The Scrum Guide gave them “Product Owner”, a role inside a delivery team, accountable for the backlog, the priorities, the acceptance criteria.</p><blockquote>Then something strange happened. In the UK especially, “Product Owner” escaped the Scrum Guide and became a job title. Organisations hired for it, built career ladders around it….and of course spent lots of money certifying people into it. Somewhere in that process, the distinction between product administration and product management quietly disappeared.</blockquote><p>The result is the organisation you probably recognise; the various ceremonies are run with regularity, roadmaps ship close to schedule, etc but nobody can quite tell you why the needle isn’t moving. The most experienced product person in the room has spent the last three months refining acceptance criteria. Strategy lives in a slide deck, at most presented once a quarter but referenced never.</p><p>The map got mistaken for the territory, and we’ve been navigating by it ever since.</p><h3><strong>AI didn’t fix it. It exposed it.</strong></h3><p>Then AI arrived and everyone celebrated. Faster story writing. Automated release notes. Sprint summaries and executive updates generated in seconds. The backlog groomed before the coffee gets cold. Product teams applauded; <em>“…finally something to take the admin off our plates!”</em></p><p>Take a moment to think about this though. If AI is comfortably absorbing most of what your senior PMs are doing day-to-day, that tells you a lot about what they are doing. And it isn’t flattering.</p><p>There’s a phrase doing the rounds: “<em>AI won’t take your job, but the PM using AI will”.</em> It’s a good line, but the wrong target. The PM who thrives isn’t the one who automates their admin fastest. It’s the one who always knew what to do when the admin was gone. The one whose value was never in the coordination layer to begin with.</p><blockquote>This is the uncomfortable question for every product leader reading this: when you remove the coordination layer from your senior product people, what are they actually doing? Not in theory…in your organisation, this week.</blockquote><h3><strong>What they should be doing</strong></h3><p>Talking to users. With enough frequency that user reality shapes product thinking rather than occasionally interrupting it.</p><p>Identifying opportunities the business can’t yet see. Stress-testing the strategy. Asking the questions others can’t ask because they’re too close, too invested, too aware of whose idea it was originally.</p><p>Experimenting. Reframing. Looking at the problem from the angle nobody has tried yet.</p><blockquote>When I introduced AI product ops tooling to a recent client team, I set one measure of success that had nothing to do with velocity or output: one protected day a week for every PM to do the work they were actually hired for…to talk to users, to think strategically and to get out of the machine.</blockquote><p>That’s a different kind of ambition. And it’s only possible if you’re honest about what the machine has been consuming.</p><p>The best product mentor I had taught me that the sharpest product work happens in the space you protect for it, so create the right environment &amp; conditions for good product thinking to emerge and that you can’t mandate insight, you can only make space for it for surface.</p><p>That’s what experienced product people are actually for.</p><h3><strong>The uncomfortable idea</strong></h3><p>So here’s where it gets harder…</p><p>If AI can handle the execution layer, the Product Owner role in the original Scrum sense (the person who owns the backlog, the priorities, the acceptance criteria) can return to where it always belonged…the subject matter expert. The CRM manager, the marketing lead, the category manager. The person with genuine domain knowledge and actual skin in the game.</p><p>The execution layer doesn’t disappear, it evolves. The junior PM of an AI-native team isn’t spending their week writing stories from scratch, synthesising research, or chasing backlog hygiene. They’re the human in the loop for all of it. They know what good looks like. They direct the AI, sense-check the output, make the judgment calls no tool can make. The grunt work gets done, but the thinking stays human.</p><blockquote>Which of course frees the senior product person to operate at the layer they were always supposed to occupy. Not Product Owner. Not even Senior PM in the traditional sense, but as an experienced strategic operator who moves between problems, asks the uncomfortable questions, creates the conditions for good product thinking to emerge and has enough distance from the day-to-day to say out loud what nobody on the inside can.</blockquote><h3><strong>Burn the maps</strong></h3><p>This isn’t a radical idea. I’ve witnessed enough broken versions of the alternative to know something needs to change. The organisations that figure this out first won’t just have more efficient product teams, they’ll finally have product teams consistently doing powerful product work.</p><p>You already know something isn’t working. You’ve hired good people and watched them disappear into the machine. You’ve shipped multiple features that didn’t move the needle. You’ve run the ceremonies and wondered why nothing felt strategic. The framework told you what product looked like, but it was never true to what product management actually is.</p><p>So burn the maps. Not because the old way was wrong, but because you finally have the conditions to chart your own.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=61f15e1a642b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Broken Product Situations Keep Finding Me]]></title>
            <link>https://medium.productcoalition.com/broken-product-situations-keep-finding-me-0bced8c88567?source=rss-a024fd07b593------2</link>
            <guid isPermaLink="false">https://medium.com/p/0bced8c88567</guid>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[product]]></category>
            <category><![CDATA[consulting]]></category>
            <category><![CDATA[organizational-culture]]></category>
            <dc:creator><![CDATA[Ruairi McGuinness]]></dc:creator>
            <pubDate>Wed, 01 Apr 2026 10:00:17 GMT</pubDate>
            <atom:updated>2026-04-08T13:44:58.778Z</atom:updated>
            <content:encoded><![CDATA[<p>I’ve started calling them <em>rescue missions</em>, but the truth is far less cinematic. I’m just the person willing to say out loud what everyone in the room already knows but hasn’t found the language or courage for yet. That’s not heroism, it’s just fulfilling my role. And after doing it more times than I can count, I’ve got reasonably good at spotting the hidden rescue mission…sometimes before the engagement has even started.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*QcBEji7ufOba-yOU4PAd7A.png" /><figcaption>Everything is familiar, but not yet in focus</figcaption></figure><h4><strong>You feel it before you can name it</strong></h4><p>It never starts with a smoking gun, it starts with a gut feeling.</p><p>A comment that doesn’t quite land right. A stakeholder who asks no follow-up questions (not because they understand, but because they don’t want to admit they don’t). A silence at exactly the wrong moment or body language that says <em>I hope nobody asks me about that</em>.</p><p>On their own, these are small moments and easily explained away. But they stack.</p><blockquote>Once you’ve been in enough of these situations, you learn to notice the stacking rather than the individual moments. That’s the skill I think; it’s not pattern recognition exactly, more like accumulation awareness. Knowing when small signals are adding up to something.</blockquote><p>The hard signals come later, once people relax into the conversation. Wild assumptions stated as facts. The throwaway <em>“it’s a small job…”</em> from someone who clearly doesn’t understand the complexity. Contradictory comments between stakeholders who don’t realise they’re contradicting each other.</p><p>That last one is particularly telling…when two people in the same organisation describe the same problem in fundamentally different ways, you’re not just looking at misalignment you’re looking at an organisation that hasn’t done the diagnosis yet and is hoping you won’t notice.</p><p>It’s not always immediate. In one engagement, the full picture didn’t emerge until a second executive stakeholder entered the conversation six weeks in, and suddenly the product vision I’d been briefed on didn’t match the one they were now communicating. That wasn’t inexperience on my part. It was incomplete information, and you can’t stack signals that haven’t appeared yet.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Lkj9tsfjxWZqrUiWCt_xeA.jpeg" /><figcaption>it’s not pattern recognition exactly, more like accumulation awareness</figcaption></figure><p>But I can usually feel it in the first discovery call. Sometimes earlier. I’ve known before the engagement has even started from signals in the sales process.</p><p>Which raises an obvious question: <em>if you can see what’s coming, why say yes to the engagement?</em></p><h4><strong>So why do I keep saying yes?</strong></h4><p>Partly it’s ego. I’ll own that. There’s something in me that doesn’t want to be the person who looked at a hard situation and walked away. That’s not entirely rational and it’s not entirely healthy, but it’s honest.</p><p>But there’s a more practical calculation underneath, one that’s specific to agency and consultancy life. In that world, you’re always aware of the bench: the gap between engagements where you’re available but not billing. I’ve never been on the bench. Not once across more years than I care to count. I think it’s partly because I’ve never said no to a difficult situation and difficult situations have a way of remembering that.</p><p>There’s also a logic to it that I’ve only recently been able to articulate clearly. When I sense a broken situation and consider declining, the question I ask myself is: what happens if I don’t take this?</p><p>The work doesn’t disappear. The broken situation doesn’t fix itself. Someone else walks in…possibly someone who won’t spot what I spotted, won’t do the diagnosis, won’t immediately push back on the wild assumptions stated as facts. Maybe that’s someone I’ve mentored. At least if I say yes, I’m in direct control of the scenario.</p><p>The last reason is the simplest. I find them genuinely interesting.</p><blockquote>Broken situations are diagnostic puzzles. Greenfield work is exciting in a different way (the blank canvas, the world of possibility) but there’s a particular satisfaction to walking into a mess, understanding it clearly and creating the conditions for it to get better. I’d rather do that than optimise something that’s already working and if I’m honest, consultants rarely get invited into those situations anyway.</blockquote><p>So I say yes. Eyes open, scope adjusted, price buffered, expectations managed. But <em>yes</em>.</p><h4><strong>The first thing I do is nothing</strong></h4><p>That sounds counterintuitive. You’ve been brought in because something is broken. There’s usually pressure…from the client, from your own agency, sometimes from your own ego to demonstrate value quickly. To show you’re earning your place. But often, the best way to make a broken situation worse is to start moving things around before you understand what’s actually broken.</p><p>Quick wins matter, but not for the reason you might think. Early momentum isn’t about proving yourself, it’s about building enough trust that people will listen when you say something uncomfortable.</p><p>So I diagnose first. Not formally (I’m not walking around with a clipboard) but I’m running a continuous loop in my head. Soaking it all in. The Five Whys. What’s the presenting problem and what’s underneath it? Is this a product problem, a process problem, or a team problem? Those are different situations requiring fundamentally different responses and conflating them is how well-intentioned interventions make things worse.</p><p>Most of the time the immediate answer is systems: broken processes, absent frameworks, accumulated technical or organisational debt that nobody has named yet. That’s actually the easy diagnosis, because systems can be fixed without anyone losing face.</p><blockquote>What you can do is create the conditions for clarity. Ask the questions nobody has asked. Name what you’re seeing without assigning blame. And then (this is the part that took me longest to learn) accept that your job is not to be the hero of someone else’s story. Your job is to leave the team better equipped to do the product thinking after your time on the project is over and you’ve moved on.</blockquote><p>A broken situation will always find more broken things to show you. Knowing when to stop looking is part of the job.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*GKWIOwMWVqKEWbtOeqi_0A.jpeg" /><figcaption>What you can do is create the conditions for clarity</figcaption></figure><h4><strong>The one thing I took the longest to acknowledge out loud</strong></h4><p>Here’s what I couldn’t have told you after my first broken situation, or my third, or my fifth.</p><blockquote>All problems are people problems</blockquote><p>Not in a cynical way. The systems are genuinely broken, the processes are genuinely absent, the technical debt is genuinely real. But underneath every broken product situation I’ve ever walked into, there have been people (smart, well-intentioned people) who just got stuck. Stuck in the wrong framing, the wrong dynamic, the wrong assumption about what the problem actually was.</p><p>I encountered a project once where the evidence of this was everywhere from day one. Significant budget committed. Work already underway. And no problem statement. And I don’t mean a weak one, I’m talking about zero. None. The kind of situation where you look around the room and realise that everyone has been too uncertain of their own judgment to say what needed saying.</p><p>That’s not a reason to blame them. It’s a reason to understand &amp; support them.</p><p>Broken situations keep finding me. I’ve stopped being surprised by that fact. What still surprises me, every time, is how much clarity is possible once someone is willing to hold a mirror up and say what everyone in the room already knows but are afraid to say out loud.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0bced8c88567" width="1" height="1" alt=""><hr><p><a href="https://medium.productcoalition.com/broken-product-situations-keep-finding-me-0bced8c88567">Broken Product Situations Keep Finding Me</a> was originally published in <a href="https://medium.productcoalition.com/">Product Coalition</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Platform Nobody Wanted: What I Learned Building in Stealth Mode]]></title>
            <link>https://medium.com/@ruairi.mcguinness/the-platform-nobody-wanted-what-i-learned-building-in-stealth-mode-d20f7de3bd69?source=rss-a024fd07b593------2</link>
            <guid isPermaLink="false">https://medium.com/p/d20f7de3bd69</guid>
            <category><![CDATA[business-strategy]]></category>
            <category><![CDATA[stealth]]></category>
            <category><![CDATA[product]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[product-development]]></category>
            <dc:creator><![CDATA[Ruairi McGuinness]]></dc:creator>
            <pubDate>Mon, 16 Mar 2026 17:21:43 GMT</pubDate>
            <atom:updated>2026-03-16T17:26:17.264Z</atom:updated>
            <content:encoded><![CDATA[<p>Years ago, I spent nine months building something nobody wanted, and it remains the most educational failure of my career.</p><p>This is about a B2B platform for a well-funded unicorn. Proper budget, a talented team, a tight timeline but with executive sponsorship. Everything you’d want as a product manager. Except I couldn’t talk to users. Two senior executives were building fundamentally different products in their heads. And we launched just as the world was shutting down.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*531xGeDpsnGR5jlBTw9saQ.jpeg" /><figcaption>Stealth mode</figcaption></figure><h3><strong>Building Blind</strong></h3><p>When you’re building a B2B product in stealth mode, you’re making extremely expensive guesses about problems hoping you’re right and praying nobody asks how you know. I had one advantage: years running a multi-site coffee retail business. I knew what managing stock felt like, what annoyed me about wholesalers and how ordering worked when you’re running a busy operation. My domain expertise became the proxy for user research because it was the only source of truth available to us; which sounds sensible until you realise I was now building based on what annoyed Ru in 2010 rather than what restaurants struggle with now.</p><p>The dangerous thing about domain expertise without validation? It feels authoritative. When someone questions an assumption, you can say “I ran a catering business for fourteen years, trust me.” And they do..because they’ve got nothing else.</p><h3><strong>Two Products</strong></h3><p>Two senior stakeholders: both intelligent, committed and likeable. <em>Both with entirely different visions for the product.</em></p><p>Executive A: This platform will disrupt food wholesale in Europe. This is strategic and this matters.</p><p>Executive B: This is a value-add for restaurant acquisition and makes our core offering stickier.</p><p>This misalignment never got resolved and we built in the space between these two visions. Features got prioritised based on which executive was in the room that week and strategic direction shifted based on who’d last spoken to the CPO.</p><h3><strong>The Assumption Spiral</strong></h3><blockquote>The insidious part about building in stealth mode is that your assumptions compound. You assume users want feature A. You build feature A. You then build feature B because it logically extends feature A. Then you build feature C because users of feature B will obviously need it. Three months later you’ve built an entire architecture based on that first untested assumption.</blockquote><p>Here’s what made it worse: the deadline wasn’t negotiable. Tied to broader business objectives we had to ship by a specific date regardless of whether what we were shipping was ready.</p><p>Healthy prioritisation requires a North Star. A clear statement of what you’re trying to achieve that you can test every decision against. We had no North Star because we had no agreement on the problem. Every prioritisation discussion became a negotiation between competing visions rather than strategic alignment towards a shared goal. We ended up building a political compromise on a deadline.</p><h3><strong>March 2020</strong></h3><p>After nine months of building blind, navigating executive politics and brutal prioritisation, we finally emerged from stealth mode. Then COVID hit with full force. Every restaurant was suddenly fighting for survival in a world of lockdowns, revenue collapse, redundancies and operational chaos.</p><p>Our platform promised to save restaurants a bit of time on wholesale ordering and maybe reduce costs slightly through better order management. These benefits might have been moderately interesting six weeks earlier. Now they were spectacularly irrelevant.</p><p>We’d built something for restaurants operating in normal conditions. COVID had eliminated normal conditions entirely.</p><p>And neither side trusted it anyway. Wholesalers were suspicious because it gave away business intelligence…transparency sounds good until you’re the one being made transparent. Restaurants distrusted it because it was another platform to learn, another login to manage, another system promising to make life easier but coming from a food delivery business, there was inherent scepticism about motives.</p><p>We’d built something that solved a problem nobody agreed existed, using assumptions we couldn’t validate, for users who didn’t want it, launching at the worst possible moment in modern restaurant history.</p><h3><strong>What Sticks With Me</strong></h3><p>I learned more working on this product than any success could have taught me, and the lessons aren’t complicated!</p><blockquote>Stealth mode prevented any validation loop: we just kept building assumption on top of assumption</blockquote><p>By the time you emerge and show anyone what you’ve built, you’re twelve assumptions deep with no idea which ones were wrong. The competitive intelligence you think you’re protecting is almost never as valuable as the user insight you’re sacrificing.</p><blockquote><em>Executive misalignment meant we were building a compromise, not a product</em></blockquote><p>You cannot build a successful product when senior stakeholders fundamentally disagree on what problem you’re solving. If you’re brought into a project where this misalignment exists, your first job isn’t building features. It’s forcing the uncomfortable conversation that makes people choose.</p><blockquote>My domain expertise felt like an asset until it became a liability</blockquote><p>It made me confident about assumptions that should have remained hypotheses. It gave me a false sense of knowing my users because I used to be my users. Domain expertise should inform your hypotheses, not replace user research.</p><blockquote>Brutal prioritisation without strategy is just triage</blockquote><p>RICE scoring only works when you know what you’re optimising for. When your goal is “ship something by deadline X that doesn’t embarrass us”, you’re not doing product management — you’re into damage limitation.</p><blockquote>External context matters more than execution</blockquote><p>We could have built the right thing for the right people with the right features and still failed because the world changed. At the end of the day, so much about launching a product can come down to luck.</p><p>Sometimes the most educational experiences are the ones where everything goes wrong for interesting reasons. I spent nine months building something nobody wanted, but as a result on each product I’ve worked on thereafter can spot those patterns forming before they become problems.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d20f7de3bd69" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>