Prioritization Is a Funding Decision

A ranked backlog answers the wrong question. It tells you the order in which work will happen and conceals what the organization believes about where value comes from, which is the only thing a prioritization exercise exists to settle. Real prioritization is a funding decision: a written argument for a small number of bets, sized with numbers you are willing to stand behind, and paid for by stopping something that people can see. If nothing visible died, you did not prioritize. You accumulated.

What a backlog hides

Stack ranking is popular because it feels like rigor and costs nothing. Everything stays on the list, merely lower, so no one has to be told that their project is over, and the organization gets to believe it has decided while retaining the option to have decided otherwise. The ranking meeting is a ritual of postponement. The tell is that the same items sit in positions eleven through thirty for years, fully alive on paper and permanently unfunded in practice, and that their owners are never quite told which.

Fund from papers

The strongest evidence I have is what happens when the decision is made properly. The multi-year data strategy I wrote for my current company was a paper, with an explicit dependency chain and explicit capability priorities, and it was funded directly out of the paper by the executive staff. There was no roadmap review and no ranking meeting, only a written argument read slowly by people who were free to disagree with it. A paper forces the argument into a form that can be wrong. If a strategy cannot survive being read without a presenter in the room, it is not ready to be funded, and the presenter's job is often to prevent the reader from noticing.

The discipline that keeps this honest is in the sizing. When we built the data platform for a new consumer lending product, the case rested on a small number of real loans from an early release and on projections, labeled as projections in every sentence they appeared in, of what the product could become over several years. The label is the point. Funding decisions survive contact with reality only when the people who made them knew which numbers were earned and which were bets, and an executive who has been surprised once by a projection dressed as an actual will discount every number you bring afterward. Credibility is the scarcest input to prioritization, and you spend it once.

The rules

Sequence by dependency rather than by appetite. A data connection, a shared model to map it into, the links to what the company already knows, the joins with first-party data, and finally the interfaces product teams build on: that chain orders the work whether or not anyone likes the order, and funding a later stage before an earlier one exists is spend rather than strategy. Kill in discovery to fund delivery; a fatal blocker found before a delivery team spins up is the cheapest source of funding in the company, and the one most organizations refuse to use because it looks like failure rather than thrift. And give every priority a heartbeat. A small set of golden metrics reviewed on a fixed rhythm is what keeps a priority a priority, since a bet nobody reviews monthly has already been quietly deprioritized by everyone except its sponsor.

Where it breaks

Paper-driven funding favors strong writers, and strong writing is not the same as strong ideas. I keep a reserve for bets that cannot yet be argued, only tested cheaply, and I have never found a principled way to size it; too small and you fund only the provable, too large and rigor becomes optional. Platform prioritization also carries an obligation that a product backlog does not. Other teams build against your roadmap, so re-prioritizing every quarter breaks trust even when each individual change is correct. Stability is itself a feature of a platform roadmap, and there are years in which the right call is to fund the second-best portfolio you can hold steady over the best one you would churn.

The open question is the reserve. What share of a portfolio should go to the unarguable, the ideas whose evidence cannot exist until someone builds a little of them? I have run it by feel, which is a polite way of saying I do not know, and I would like a better answer than feel.