Data & Analytics · August 2026 · 5 min read
AI can help one person build at the speed of a team. It can't tell you whether anyone cares — but five signals can.
This is the short version. Read the full article on Medium →
AI has made it easy for one person to build at the speed of a team. What it hasn't done is make anyone want what you built. When production races ahead of demand, you rack up “demand debt” — a pile of finished work with no proven buyers. Being data-driven is how you stay honest about that, but it means measuring what your next decision needs, not filling a dashboard.
Automation sped up production; it didn't speed up customer discovery. Build too far ahead of demand and you accumulate demand debt — the selling you deferred, now due with interest. The fix isn't to build faster; it's to alternate between building and selling before either gets expensive.
Measuring everything is a way of avoiding decisions. Track only the signals that change what you'll do next. Framed as a hypothesis to test rather than a number to admire, measurement sharpens decisions and makes you quicker to pivot.
Five “R”s track a customer's progress: Reach — the right people find your offer through a repeatable channel; Resonance — qualified visitors take a meaningful action; Revenue — they pay, and you note how much effort each sale took; Realisation — they actually reach the outcome you promised; and Return — they buy again, renew, or refer. A weak link shows up as a specific signal, not a vague “sales are slow.”
Not all friction is bad — some of it qualifies buyers. The goal is to remove unnecessary uncertainty: the discovery, comprehension, trust and decision gaps that stop the right person from acting. Fix what's actually blocking progress, and leave the rest.
Before automating, sell by hand. Talking to early customers directly gives you richer evidence than any dashboard, and shows you what's worth automating later. Let AI take the repetitive work around your judgment — without automating the judgment itself.
Decide your maximum downside up front — money, time, and the threshold that ends an experiment — then run it. Keep each bet small enough that the market can correct you before you're too attached to be corrected.
Once a week, glance at the five signals and pick the single experiment that would reduce your biggest uncertainty. Success isn't a prettier dashboard; it's the shortest honest path between an assumption and the evidence that settles it.
Want the full walkthrough, with the worked example and the reasoning behind each step?
Read on Medium →Get the next guide and tool the day it lands.
Subscribe — free