[{"data":1,"prerenderedAt":240},["ShallowReactive",2],{"writing-article-\u002Fwriting\u002Ftools-dont-make-the-product":3,"writing-index":163},{"id":4,"title":5,"body":6,"cover":148,"coverAlt":148,"date":149,"description":12,"draft":150,"extension":151,"featured":150,"meta":152,"navigation":153,"path":154,"readingTime":155,"seo":156,"stem":157,"summary":158,"tags":159,"__hash__":162},"writing\u002Fwriting\u002Ftools-dont-make-the-product.md","Tools Don’t Make the Product",{"type":7,"value":8,"toc":138},"minimark",[9,13,16,21,24,27,30,34,37,40,43,46,50,53,82,85,89,92,95,98,102,105,122,125,129,132,135],[10,11,12],"p",{},"Every few months a team decides that the problem is the tool. The backlog is messy, the handoffs are unclear, the research is scattered. A new platform arrives with a polished demo and a promise of alignment. Six weeks later, the team has migrated, the backlog is still messy, and the handoffs are now unclear in a more expensive system.",[10,14,15],{},"The tool was not the problem. It was the place where the problem became visible.",[17,18,20],"h2",{"id":19},"tools-encode-decisions","Tools encode decisions",[10,22,23],{},"A tool is not neutral. It encodes a set of assumptions about how work should happen.",[10,25,26],{},"A ticket tracker assumes that work can be decomposed into items, that items have states, and that progress can be measured by movement between those states. A document tool assumes that knowledge is best captured in pages. A research repository assumes that insights should be tagged, clustered and reused. A roadmap tool assumes that the future can be arranged in rows and columns.",[10,28,29],{},"Those assumptions are decisions. When you adopt the tool, you adopt the decisions, often without discussing them. If the team’s actual workflow disagrees with the tool’s assumptions, the tool wins by default, because it is harder to change than a habit.",[17,31,33],{"id":32},"diagnose-the-bottleneck-first","Diagnose the bottleneck first",[10,35,36],{},"Before switching tools, name the bottleneck in concrete terms.",[10,38,39],{},"Is the problem that decisions are slow? That might be an ownership issue, not a software issue. Is it that context is lost between teams? That might be a documentation habit, not a platform. Is it that nobody knows what is being worked on? That might be a cadence problem.",[10,41,42],{},"Each of those has a tool-shaped solution and a process-shaped solution. The process-shaped solution is usually cheaper and more durable.",[10,44,45],{},"A useful test: run the workflow manually for one cycle. Use a shared document, a spreadsheet, a whiteboard. If the workflow works, the tool is an optimisation. If it does not, the tool will hide the failure rather than fix it.",[17,47,49],{"id":48},"switching-costs-are-real","Switching costs are real",[10,51,52],{},"Tool migrations are usually estimated in licences and training. The larger costs are less visible:",[54,55,56,64,70,76],"ul",{},[57,58,59,63],"li",{},[60,61,62],"strong",{},"History."," The old tool contains decisions, context and conversations that will not transfer cleanly.",[57,65,66,69],{},[60,67,68],{},"Attention."," A migration consumes the team’s capacity for weeks, often during the quarter when the actual product work was supposed to happen.",[57,71,72,75],{},[60,73,74],{},"Integrations."," Every tool is a node in a graph. Changing one node changes the edges, and some of those edges are held by other teams.",[57,77,78,81],{},[60,79,80],{},"Habits."," People have built workarounds that encode local knowledge. A new tool removes the workarounds without replacing the knowledge.",[10,83,84],{},"None of this means tools should never change. It means the decision should be priced honestly.",[17,86,88],{"id":87},"ai-tools-are-not-exempt","AI tools are not exempt",[10,90,91],{},"The current wave of AI tools is particularly prone to this pattern, because the demo is so compelling. A tool that can draft a spec, summarise a meeting or generate a prototype looks like it solves the production problem.",[10,93,94],{},"But most teams do not have a production problem. They have an agreement problem: what are we building, for whom, and why. AI can accelerate the production of artifacts that encode a decision nobody has actually made. The result is more output, faster, with the same underlying ambiguity.",[10,96,97],{},"The useful question is not “what can this tool do?” but “which decision does it make cheaper, and does that decision need to be cheaper?” If the answer is no, the tool is a distraction with a subscription.",[17,99,101],{"id":100},"what-to-evaluate-instead","What to evaluate instead",[10,103,104],{},"When a tool is genuinely needed, evaluate it against the bottleneck rather than the feature list.",[54,106,107,110,113,116,119],{},[57,108,109],{},"Does it make the slow decision faster?",[57,111,112],{},"Does it make the missing information visible?",[57,114,115],{},"Does it reduce the number of handoffs?",[57,117,118],{},"Does it survive contact with the team’s actual habits?",[57,120,121],{},"Can it be removed without losing the work?",[10,123,124],{},"That last question is underrated. Tools that make export easy are more trustworthy than tools that make leaving expensive.",[17,126,128],{"id":127},"the-order-of-operations","The order of operations",[10,130,131],{},"The reliable order is: clarify the process, choose the constraint, then choose the tool. Not the other way around.",[10,133,134],{},"A team that knows how it decides, what it needs to see and who owns what can work with almost any tool. A team that does not know those things will reproduce its confusion in every new platform it adopts.",[10,136,137],{},"The tool is the shape you pour the process into. If the process is undefined, the tool does not create a shape. It just makes the spill look organised.",{"title":139,"searchDepth":140,"depth":140,"links":141},"",2,[142,143,144,145,146,147],{"id":19,"depth":140,"text":20},{"id":32,"depth":140,"text":33},{"id":48,"depth":140,"text":49},{"id":87,"depth":140,"text":88},{"id":100,"depth":140,"text":101},{"id":127,"depth":140,"text":128},null,"2025-09-09",false,"md",{},true,"\u002Fwriting\u002Ftools-dont-make-the-product",3,{"title":5,"description":12},"writing\u002Ftools-dont-make-the-product","A new tool cannot fix an undefined workflow. It usually makes the disagreement faster, more visible and more expensive.",[160,161],"tools","practice","A9wUUTtFPjgKo4CCaurboHcA-_UbeJ_jqUlWxVIfypQ",[164,173,181,188,194,200,206,213,219,226,232,234],{"path":165,"title":166,"summary":167,"date":168,"tags":169,"featured":153,"readingTime":172},"\u002Fwriting\u002Fwhat-ai-changes-about-product-work","What AI Changes About Product Work, and What It Doesn’t","The cost of producing a prototype, a spec or a synthesis has collapsed. The cost of knowing what is worth building has not. That gap is where product judgement now lives.","2026-11-04",[170,171],"ai","product-thinking",4,{"path":174,"title":175,"summary":176,"date":177,"tags":178,"featured":150,"readingTime":155},"\u002Fwriting\u002Froadmaps-are-arguments","Roadmaps Are Arguments, Not Schedules","A roadmap is a set of claims about why these problems, in this order, with these trade-offs. When it becomes a delivery calendar, the reasoning disappears and the promises take over.","2026-10-14",[179,180],"product-strategy","decision-making",{"path":182,"title":183,"summary":184,"date":185,"tags":186,"featured":150,"readingTime":172},"\u002Fwriting\u002Fdiscovery-is-a-budget","Discovery Is a Budget, Not a Phase","Treating discovery as a stage before delivery assumes uncertainty is front-loaded. It isn’t. Discovery works better as a recurring spend that every cycle pays into.","2026-09-23",[187,179],"discovery",{"path":189,"title":190,"summary":191,"date":192,"tags":193,"featured":150,"readingTime":172},"\u002Fwriting\u002Fcost-of-a-decision-you-didnt-make","The Cost of a Decision You Didn’t Make","Deferred decisions feel safe because nothing visible happens. But indecision charges interest, in rework, lost context, and options that quietly expire.","2026-08-19",[180,171],{"path":195,"title":196,"summary":197,"date":198,"tags":199,"featured":150,"readingTime":155},"\u002Fwriting\u002Fthe-feature-was-never-the-problem","The Feature Was Never the Problem","A feature request is a compressed problem statement with the reasoning removed. If you ship the request without unpacking it, you often solve the smallest part of the problem.","2026-06-30",[171,187],{"path":201,"title":202,"summary":203,"date":204,"tags":205,"featured":150,"readingTime":155},"\u002Fwriting\u002Fprioritisation-is-about-evidence","Prioritisation Is a Story About Evidence","Frameworks like RICE do not make decisions. They make assumptions comparable. The quality of a prioritisation is the quality of the evidence and the honesty about what gets displaced.","2026-05-12",[179,180],{"path":207,"title":208,"summary":209,"date":210,"tags":211,"featured":150,"readingTime":172},"\u002Fwriting\u002Fstakeholders-are-not-obstacles","Stakeholders Are Not Obstacles","“Stakeholder management” often frames other people as friction. But they hold information the product team lacks. Treat them as co-owners of the constraint, not gatekeepers.","2026-03-24",[161,212],"leadership",{"path":214,"title":215,"summary":216,"date":217,"tags":218,"featured":150,"readingTime":172},"\u002Fwriting\u002Fthe-quiet-power-of-constraint","The Quiet Power of a Good Constraint","Constraints are usually treated as obstacles. A good one narrows the search space, forces a real choice, and makes the work more honest.","2026-02-10",[171,179],{"path":220,"title":221,"summary":222,"date":223,"tags":224,"featured":150,"readingTime":172},"\u002Fwriting\u002Fwhen-research-stops-being-useful","When Research Stops Being Useful","Research becomes decorative when it is disconnected from a decision. More interviews will not fix a team that has not agreed on what it is trying to learn.","2025-12-02",[225,187],"research",{"path":227,"title":228,"summary":229,"date":230,"tags":231,"featured":150,"readingTime":172},"\u002Fwriting\u002Fthe-product-review","The Product Review That Actually Changes the Product","Most product reviews are status meetings with slides. A useful one changes a decision, and that requires a different format, a different question set and a named owner.","2025-10-21",[161,180],{"path":154,"title":5,"summary":158,"date":149,"tags":233,"featured":150,"readingTime":155},[160,161],{"path":235,"title":236,"summary":237,"date":238,"tags":239,"featured":150,"readingTime":172},"\u002Fwriting\u002Foutcomes-are-a-habit","Outcomes Are a Habit, Not a Metric","Teams say they are outcome-driven but manage by output because output is legible. Outcomes require a cadence of looking at behaviour and the discipline to change course.","2025-07-15",[179,161],1795010564451]