The Anti-Craft Machine
How misaligned incentives quietly destroy product quality, culture, and customer trust.
Run this diagnostic on your B2B company:
□ Product, engineering, and design have separate metrics instead of shared accountability for customer outcomes
□ Product disrupts the roadmap to land deals, regardless of whether customers will actually use the features
□ Engineering is rewarded for hitting dates and velocity, not whether what they shipped works well
□ Design has been beaten into submission to propose “achievable” mediocrity to avoid 18-month estimates
□ Decisions get made in rooms, by committees, where no one understands how all the pieces fit together
□ Teams hand off specs between functions instead of owning problems end-to-end
The more boxes you check, the more your system selects against craftspeople.
What You’re Actually Optimizing For
Product is measured on revenue growth
New customers, expansion revenue, retention numbers. To hit those numbers, they need features that unlock new markets, win competitive deals, and check boxes in sales decks.
This creates pressure to say yes to everything sales needs to close deals. The roadmap becomes a negotiation with the loudest enterprise customer. Product managers optimize for shipping features that make sales decks more competitive, not features that make the product better.
Engineering is measured on hitting dates
Story points completed. Velocity. This creates inflated estimates as self-defense and corners cut to make deadlines. Technical debt compounds while everyone pretends it doesn’t exist.
Design has been beaten down
They propose something thoughtful and get back an 18-month estimate. They learn to self-censor. They start designing mediocre shit that’s “achievable” instead of good.
Nobody is collectively accountable for customer outcomes. Not feature adoption rates. Not time to value. Not customer satisfaction with specific workflows. Not retention tied to product quality.
Product hit their revenue target. Engineering hit their deadline. Design shipped their mocks. The fact that only 8% of customers actually use the feature is nobody’s problem specifically, so it’s nobody’s problem at all.
See if you recognize this cycle: Sales needs feature X to close a big deal. VP of Product needs to hit their revenue target. Sales essentially demands that Feature X goes on the roadmap with a tight deadline. Engineering estimates 6 months. Product says we need something in 6 weeks. Engineering cuts corners to make it happen. The feature ships barely functional. The deal closes. No one uses Feature X. Existing customers complain you still have fixed the last 20 broken features.
That’s because no one is accountable for the fact that the feature doesn’t actually solve the customer’s problem well. Six months later that customer isn’t expanding because the product is frustrating to use. But that’s a retention problem, not a product problem. Different metric. Different quarter. Nobody connects the dots.
This is the machinery. Three groups with misaligned incentives, each optimizing for their own metrics, wondering why the output feels soulless.
The craftspeople you hire see this immediately. They push back. They argue for doing things right. They get labeled “not a team player” or “too perfectionist” or “doesn’t understand the business.” They leave or learn to stop caring.
Your system selected against them. Working as designed.
What This Actually Costs You
You ship features customers asked for that they don’t actually use. Product celebrated closing the deal. Engineering hit the deadline. The feature sits at 8% adoption because it’s too complicated or doesn’t fit the workflow or solves the wrong problem.
Your best people leave. The ones who could build something great go to companies where they’re allowed to. You’re left with people who’ve accepted that mediocrity is the job. They’re not bad people. They’re rational actors responding to your incentive structure.
Your product becomes a commodity. When the software barely works and everyone knows it, you compete on price, complexity, and integrations. Your differentiation is a checkbox on a comparison chart. Your customers stay because switching costs are high, not because they like using your product.
Your technical debt makes everything slower. That 18-month estimate wasn’t crazy. It accounts for the three years of shortcuts that made shipping fast possible. Now simple changes require architectural rewrites. Your velocity drops every quarter while your backlog grows.
What Actually Needs to Change
Align incentives around shared customer outcomes
Product, engineering, and design need collective accountability for metrics like feature adoption rates, time to value, customer satisfaction with specific workflows, and retention tied to product quality.
If product can close deals with features customers don’t use and still hit their OKR, you haven’t fixed anything. If engineering can ship on time with features that don’t work well and still get rewarded, nothing changes. If design can ship pixels that customers hate and still meet expectations, craft is impossible.
Own problems end-to-end
Give teams a customer problem and let them figure out the solution. Not a spec to implement. Not a backlog of tickets. A problem to solve, with authority to make tradeoffs about scope, timeline, and approach.
This requires trust that most orgs don’t have. If you can’t trust a team to make good tradeoffs, you either hired wrong or broke them with your process.
Make someone accountable for customer outcomes
Someone whose job is to look at whether customers actually use what you built, whether they get value quickly, whether they’re satisfied with the experience. Someone who can say no when the feature will close a deal but won’t actually work well.
This person will be extremely unpopular. Protect them or they’ll leave.
Kill the handoffs
If design creates specs that engineering estimates at 18 months, your process is broken. Either designers don’t understand what’s technically feasible or engineers are padding estimates because they don’t trust the requirements.
Get them in the same room earlier. Prototype together. Kill bad ideas in hours instead of months. This requires designers who understand technical constraints and engineers who care about user experience.
Stop measuring velocity
If you’re measuring story points completed, you’re measuring output. Output is not outcomes. You can ship a thousand story points of features nobody uses.
Measure whether customers actually adopted what you built. Measure how quickly they got value from it. Measure whether they’re less frustrated than before. Measure whether retention improved because the product got better, not just because switching is hard.
Why Most Won’t Do This
Changing incentive structures is political
Product VPs need to hit revenue targets. Engineering VPs need to deliver predictably. Design leaders fought for a seat at the table by proving they could keep engineering fed with easily implementable designs.
Telling them their metrics create perverse incentives threatens their positions. They’ll resist. They’ll argue that craft doesn’t scale or customers don’t care about quality or the market moves too fast to sweat details.
They’re protecting the system that rewards them.
Quarterly pressure makes long-term thinking hard
Fixing technical debt doesn’t show up in this quarter’s metrics. Rebuilding a feature properly takes longer than shipping a mediocre version. Your board wants growth now.
The companies that commit to craft accept slower growth upfront for compounding returns later. Most boards won’t tolerate this. Most CEOs can’t sell it.
Hiring for craft is harder than hiring for credentials.
You need people who care deeply about customer outcomes and have the judgment to make good tradeoffs. Interviews don’t reveal this. Trial projects do, but they’re expensive and slow.
Most companies hire for résumés and interview performance. Then wonder why their senior engineers don’t care about whether customers actually use what they built.
The Actual Choice
You can build an organization where craft thrives. It requires structural changes to incentives, hiring, and accountability. It requires collective ownership of customer outcomes across product, engineering, and design. It requires protecting craftspeople from the pressure to ship garbage to close deals. It requires accepting that quality takes longer upfront.
Or you can keep running a feature factory while performing craft. Citing Linear in all-hands meetings. Adopting frameworks that require shared accountability you don’t give people. Hiring talented people into a system designed to frustrate them.
Most companies choose the second option. Not explicitly, but through inaction. The political cost of reorganizing is too high. The quarterly targets are too aggressive. The existing incentive structures are too entrenched.
If you checked multiple boxes on that diagnostic, you probably can’t get there from here. The machinery is working as designed. It just can’t produce software people love.
The companies that actually commit become the products people talk about. They have fans, not just users. They charge more because the difference is obvious. They keep their best people because those people can do their best work.
Everyone else competes on price and prays their switching costs stay high.
You know your incentive structure better than I do. You know whether your system rewards customer outcomes or just output. You know whether you’re willing to change it.
Act accordingly.