The brief is never the brief
A request lands: “Build a winback flow.” It sounds complete. It isn’t. Hidden inside it are a dozen decisions someone has already made by accident — who counts as lost, how long is too long, what we’re willing to give away, what success even looks like. Build straight from that sentence and you’re not executing a plan; you’re hard-coding someone’s unexamined assumptions and calling it a campaign.
The expensive failures I’ve seen almost never failed in the build. They failed in the brief — a flow that worked exactly as specified, for a goal nobody had actually agreed on. The hour of questions before the build is the cheapest hour in the whole project. Skipping it doesn’t save time; it moves the cost to the end, where it’s ten times larger.
The five questions that decide it
Before I touch a canvas, I want five answers. Who exactly — the precise behavioral definition of the audience, not “lapsed users” but “ordered 3+ times, no order in 45 days.” Why now — what event or threshold makes this the moment to speak. What’s the one action — the single thing we want them to do, not the four we’d like. What does success look like — the number that moves, measured against a holdout, decided before launch. What’s it allowed to cost — in margin, in goodwill, in the right to send again.
None of these are technical. All of them are the kind of thing that, left unasked, quietly sets itself to the wrong default. Most of the value I add to a flow is added in this conversation, before a line of it exists.
A campaign is only as good as the question it answers. Most briefs never name the question.
Questions are senior work
There’s a quiet pressure, especially under deadline, to look decisive — to take the brief, nod, and start building. Asking feels like friction. It isn’t. The junior move is to build what you were told; the senior move is to find out what was actually meant. I’ve run this across a global platform sending roughly two million communications a month, and the pattern held everywhere: the teams that shipped fastest were the ones that questioned hardest at the start.
Good questions also do something a spec can’t — they surface the disagreement that was always there. Half the time, asking “what does success look like?” reveals that two stakeholders wanted two different campaigns. Better to find that out over coffee than during launch.
Run the brief through the five
Next time a campaign request lands, don’t open the tool. Open a doc and answer the five out loud: who exactly, why now, what one action, what success, what cost. Where you can’t answer, you’ve found the real work — go get the answer before you build, not after. Send the filled-in version back to whoever asked; watch how often it changes the request.
It feels slower for an afternoon and is faster for a quarter. The flow you build from five clear answers usually beats the one you build from a confident sentence — and you’ll have stopped, somewhere in the asking, at least one campaign that should never have shipped.