Growth teams chase 10% wins and miss the ones that compound. A 5% lift in 3 places does more for ARR than a single redesign, and it ships in a week.
The math is unforgiving. A funnel with 5 steps at 50% conversion each ends at 3.1%. Lift each step by 5% absolute and you end at 4.5%. That is a 45% lift in end-to-end conversion, from changes that no one notices in isolation.
Most growth teams know this in theory. In practice, they keep funding the redesign.
Why teams default to big bets
3 pressures push teams toward large, visible changes.
Stakeholder narrative. A redesign is easy to present to a board. A 5% lift in the second step of activation is not. The first looks like strategy. The second looks like maintenance.
Designer incentives. Designers build portfolios on redesigns, not on changing the copy of a single button. The org rewards work that photographs well.
The novelty bias. Teams remember the launch that doubled signups. They forget the 8 quiet experiments that doubled retention over the same period. The launch is a story. The experiments are a spreadsheet.
Where the compound wins live
The funnel steps that respond to small fixes are the ones with the highest drop-off and the lowest design investment. In most B2B SaaS products, that is a short list.
The signup form. Every required field costs conversion. Every social login option recovers some of it. Every error state that does not explain itself loses users who will not try again.
The first empty state. What the user sees in the first 3 seconds after signup determines whether they stay. A blank dashboard is a leak. A pre-populated example, even a synthetic one, is a patch.
The activation prompt. The single action you most want the user to take, presented in the place they are most likely to take it. Most products bury this under a navigation pattern.
The trial-to-paid wall. The page where the user decides whether to enter a credit card. Copy on this page is usually written once and never revisited. It is also one of the highest-leverage paragraphs in the product.
The reactivation email. The message sent to a user who has not logged in for 30 days. Most teams send a generic re-engagement template. The teams that win send a message about the specific value the user got the last time they used the product.
The mechanics of small wins
Small funnel fixes share 3 properties. They are cheap to test. They are reversible. They compound with each other rather than canceling out.
A button copy change is cheap. An empty state illustration is cheap. A clearer error message is cheap. None of these require a sprint. All of them can be tested in a week with statistical significance, assuming the funnel has enough volume.
They are also reversible. If the new button copy underperforms, you switch back. The cost of a wrong call is the time spent on the test, nothing more. Compare this to a redesign, where reverting means shipping a second redesign.
And they compound. A 5% lift in signup conversion, a 5% lift in activation, and a 5% lift in trial-to-paid conversion multiply. The redesign that lifts one step by 20% often drops an adjacent step by 10%, because changing the visual language of the funnel retrains the user mid-flow. The net impact of a large change is frequently smaller than the sum of 3 small ones.
How to run the program
A compound-fix program looks boring from the outside. That is the point. The teams that run it well share a few habits.
A live funnel map. Not a slide. A dashboard that updates weekly and shows conversion by step, with absolute numbers, not just rates. The map is the backlog.
A weekly test cadence. 1 test per step per month is a reasonable pace for a team of 2. Most teams run fewer tests than this and wonder why growth is flat.
A 90-day review. Small wins are invisible in a one-week window and obvious in a 90-day window. Report on the cumulative lift, not on individual tests. This reframes the work for stakeholders who are looking for narrative.
A kill list. Changes that did not work get reverted within a week. Changes that worked stay. No nostalgia for ideas that lost.
When to redesign anyway
Redesigns are not always wrong. They are the right call when the funnel is architecturally broken, when the information hierarchy no longer matches the product, or when the brand has drifted far enough from the interface that new users cannot parse what the product is.
They are the wrong call when the funnel is directionally correct and leaking at predictable seams. In that case, a redesign replaces known leaks with unknown ones, and the team spends the next quarter rediscovering what it already knew.
The test for whether you need a redesign is simple. List the 5 highest-drop-off steps in your funnel. If you cannot name the specific friction at each step, you have a research problem, not a design problem. Fix that first.
The uncomfortable part
Compound-fix work is hard to staff. It rewards patience over ambition. It produces results that are hard to attribute to any one person. It is the opposite of what most career paths in product and growth optimize for.
This is why it works. The teams that do it are doing work that most of their competitors have structurally decided not to do.