The Living Roadmap

Almost every roadmap is made the same way. People go into a room with what they’ve heard — from sales, from support, from the loudest customer, from the competitor who shipped something on Tuesday. They argue about it, sensibly. They come out with a list in an order.

It’s a reasonable process, and it produces a document that starts going out of date on the way out of the room. Not because the people in it were wrong. Because a roadmap is a record of one conversation, and the thing it describes is a live product with customers moving through it right now.

You already have a better source

Your customers answer the roadmap question continuously. They just don’t answer it in words. Every session is a small opinion about what your product is missing. A few ordinary shapes — illustrative, not measured:

• —Someone opens the same two options over and over and picks neither. The page doesn’t contain the thing they need in order to decide.
• —Someone searches for the same thing three times in different words. What they want exists and isn’t called what they call it.
• —A team exports your data every Monday and rebuilds it in a spreadsheet. The view they actually need isn’t in your product.
• —A phone number turns up in a notes field. There was nowhere else to put it.

None of that arrives as a complaint. Nobody files a ticket about a page that’s merely unhelpful — they leave, or they invent a workaround. Workarounds are the most useful thing a product can tell you and the least likely to be reported, because the moment someone solves a problem themselves it stops feeling like your problem.

Hesitation counts too

Often the clearest information is in what didn’t happen. The pause before a button. The form abandoned at the same field every time. The page people read twice before moving. Hesitation is a customer doing arithmetic about whether to trust you, and where they hesitate is where your product hasn’t answered something.

Analytics have been able to show this for years. Most teams could pull a funnel this afternoon and find the step where people fall away. Which is the strange part of the problem: seeing it was never the bottleneck.

Seeing it isn’t the hard part

A roadmap drifts from reality not because teams can’t see reality, but because seeing produces work, and work has to be paid for. Every observation has to be turned into an argument. The argument has to win against other arguments. The winner waits for capacity. By the time something noticed in March is built in July, it has passed through four people and lost most of its detail — and nobody involved did anything wrong.

So teams end up with an excellent system for looking and a queue for responding. The dashboard updates every hour. The product updates when someone gets to it.

A roadmap made in a room is a guess about what customers will want. A roadmap made from behavior is a record of what they already tried to do.

When noticing and acting are the same system

What changes this isn’t better dashboards. It’s the noticing and the building being one system. When the software that sees the abandoned step is also the software that can build a different step, test it, and put it live once you’ve approved it, an observation never has to survive being translated into an argument. It goes straight to a change.

That’s what a living roadmap means. Not a document updated more often — no document in the middle at all. What to do next is worked out from what’s actually happening, and the order changes when your customers change, not when the quarter does.

It also sharpens what a roadmap is for. You still decide what the product is, who it’s for, and what it will never be — the judgement calls behavior can’t make for you, because customers can only react to what already exists. What goes away is the other part: the small, obvious, evidence-backed improvements that sit in a list for a year because nobody has a spare afternoon.

Your customers have been writing your roadmap since the day you launched. The only questions are how much of it you get to read, and how long it takes you to answer.