transcripts.wiki · lesson

Delta 4 — Why a Good Score Isn't the Finish Line

A survey can tell you users would miss your product — it can't tell you whether they'll ever go back to the old way, and that second question is the one that decides whether you grow.

Part 5 of Finding Product-Market Fit — a guided five-lesson path.

Past the survey

The previous lesson, The Sean Ellis Score, turned product-market fit into a number: the share of users who say they would be 'very disappointed' to lose the product, with 40% as the threshold Sean Ellis identified empirically. Clear that bar and the standard story says you have fit.

Kunal Shah, founder of CRED and Freecharge, asks the uncomfortable follow-up: what if you clear the survey and the product still does not grow on its own? A user can honestly report they would miss a product and still revert to the old way the moment the product stumbles, and never tell a friend about it either way. Shah's Delta 4 framework names the condition the survey does not measure — and argues it is the one that actually drives organic growth.

The efficiency threshold

The method is a comparison, not a survey. Rate the existing solution on a 1–10 efficiency scale. Rate the new solution on the same scale. If the new solution scores at least four points higher — a delta of four or more — three things follow automatically:

Below a Delta 4, all three properties invert: behaviour stays reversible, users complain loudly at the first failure, and there is no organic growth to speak of. This is the gap the Ellis survey leaves open — it can register real, even strong, affection for a product that has not crossed the line into irreversible habit. A product can score well on 'would you miss it?' while sitting below the efficiency threshold that makes growth take care of itself.

Two worked examples

Shah's own comparisons show the delta at work in both directions:

The suit example is the more important of the two, because it shows that adding technology does not guarantee a higher efficiency score. What counts is the user's experienced efficiency, not the convenience of the transaction alone.

The honest limit

Delta 4 is a necessary condition, not a sufficient one — and its own author says so. The scoring is subjective and varies by user segment: Uber's delta is lower for a user in a city with excellent public transport than for one without it. The ten-point scale is unmeasurable without user research, and the threshold of four is, by Shah's own admission, arbitrary — chosen to force teams to ask whether they are appreciably better, not to supply a precise cut-off.

There is a second gap the framework does not cover. Some products achieve the same irreversibility Delta 4 describes through switching costs, data lock-in, or network effects, rather than through any genuine efficiency gain. A user trapped by lock-in behaves exactly like a user who has crossed a real Delta 4 — reversion is costly either way — but nothing about the product actually improved. The framework diagnoses one route to irreversibility; it does not rule out the others.

Where the series lands

This closes the arc that opened with What Product-Market Fit Actually Is, where Eric Ries's heuristic set the terms: if you are asking whether you have PMF, you do not have it — fit is unmistakable, and retention is the ultimate proof. Everything in between has been an attempt to make that unmistakable feeling checkable before the fact: a survey threshold, a segment to target, an efficiency delta to clear. None of them replaces the others. A high Ellis score without a Delta 4 can still fail to compound; a wide Delta 4 with no retention data behind it is still a guess. Product-market fit is not a formula to solve but a state to discover, and the honest practice is to triangulate it with several imperfect instruments at once, rather than trust any single number to say the search is over.

Go deeper

The single best source on Delta 4 is Kunal Shah on Lenny's Podcastwatch it here, or read the episode.