How to Translate a Retention Problem Into Language a Board Will Act On

A product team can walk into a board meeting with the retention problem fully mapped and walk out with nothing. The slide is there. Day one, day three, day seven, all sliding downwards in the same direction. The board nods, asks a question about the roadmap, and then spends forty minutes on the acquisition line instead.

Retention is a finance number that happens to be measured inside the product.

That happens because the problem arrives in the room dressed as a product issue. Retention sounds like something designers and developers sort out between themselves, so the board files it under work in progress and moves to the part of the meeting where money gets allocated.

The translation is learnable. We have run it on both sides of the table, and most of what follows comes out of two projects in particular. One was an auction game where the client kept asking for new features while engagement quietly fell away. The other was a genetics wellness app where the brand promise was sound and the product had been built in a way that threw the story out. Both ended up being conversations about money, and both needed the retention problem said out loud in a way a board was able to act on.

Why Boards Hear 'Retention' as a Product Problem, Not a Money Problem

Boards like levers they can pull with cash. Acquisition is the clearest one. Put more money in the top, get more installs out the bottom, and the relationship between the two is easy to see in a single line on a chart. Retention has no equivalent lever, so it reads as housekeeping.

There is a second reason, and it is more uncomfortable. Download numbers going up feel like proof that things are fine. The chart climbs, the deck looks healthy, and nobody is looking at how many of those people are still there on day three, day five or day seven. A product can grow its install base every month and be losing users faster than it wins them.

When confidence replaces evidence

On one project, between funding rounds, we found that investor confidence had been doing the job retention data should have been doing. The round had closed, so the product was treated as proven. Nobody had asked the users. The decision we took was to put in much better tracking of retention and to start asking people proactively how things were going, rather than waiting for complaints that were never coming.

Most unhappy users say nothing at all. They just stop opening the app, and silence gets read as contentment.

The Numbers a Board Already Tracks and Where Retention Sits Inside Them

You rarely need a new number. The board already watches customer acquisition cost, lifetime value, payback period, monthly revenue and burn. Retention is sitting inside three of those five, doing most of the work, and nobody has pointed at it.

Lifetime value is the obvious one. How long someone stays is half of what that figure is made of, so any movement in retention moves lifetime value directly. Payback period is the same story told over time. If people leave before they have repaid what you paid to get them, the payback period is theoretical. Burn follows, because a product that leaks users needs a bigger acquisition budget every month to stand still.

Ratios the board already believes

Plenty of boards work to a rough target of one to three on acquisition cost against lifetime value, a benchmark set out by Userpilot. That target is useful in the room for one reason. The denominator is retention. If the ratio is slipping, the fastest honest way to move it is to keep more of the people you already paid for.

Say that sentence and you have stopped talking about product quality. You are talking about the ratio they already chose to be judged on.

Turning Day-One to Day-Thirty Drop-Off Into Acquisition Cost Maths

The conversion is arithmetic, and it is worth doing on one slide. Take a month of paid installs. Say the number is 20,000. If half of those people never come back after the first session, then half of that month's acquisition budget bought a single session each. Not a customer, not a trial, one session.

Then carry it to day thirty. If one in ten of the original group is still active, your real cost per retained user is ten times the headline cost per install. That is the number worth putting in front of a board, because it is the one that is actually true.

Half of last month's acquisition spend bought one session per user, and finance can price that.

The reason this lands is that boards already know winning a new customer is the expensive option. Bain & Company put the cost of acquiring a new customer at five to twenty-five times the cost of keeping an existing one. Nobody argues with that in principle. The argument only starts when you ask for money to act on it.

Anything below fifty per cent retention after day one is borderline in our book, and that is the point where we start treating the acquisition budget as partly wasted rather than partly working.

Benchmarks That Make the Gap Visible: 77 Per Cent Versus 40 to 50 Per Cent

Two figures do most of the work when we are asked to explain why early retention is worth paying for. On average, around 77 per cent of apps lose their daily active users within the first three days of download. Even a very strong product sees something like a 40 to 50 per cent drop over those same three days.

These are the numbers we put in the room. The second one matters because it stops the conversation turning into a fantasy. Some drop-off after day one is normal and expected, and no amount of money buys it away. A board that hears a promise of ninety per cent day-three retention stops listening, and rightly so.

Where the money actually is

The distance between 77 per cent and 40 to 50 per cent is the part you can buy. That space is onboarding, early value, and whether people understand what they have in front of them in the first few minutes. It is not a small sliver either. For a product paying for installs at any scale, moving from the average towards the better end changes what every pound of acquisition spend returns.

Put both benchmark figures on the same slide as your own day-three number. The board sees where you sit without you having to say it.

What Counts as a Warning Sign and What Counts as Normal

Boards get nervous when everything is described as an emergency, and they stop listening when nothing is. So be clear about which is which. A sharp fall after the first session is ordinary. A fall that keeps the same slope at day seven and day thirty is a product that never gives anyone a reason to stay.

That third row is the one teams get wrong most often. Session length going up is only good news if people are getting value from the product and resonating with it. The same number rises when people are confused, and it rises again when a product has been gamified to hold them in place. Simon calls those the vanity metrics of engagement, and a board that has been shown them for two quarters will be slow to trust the next number you bring.

Sentence Structures That Get the Conversation Moving

The words matter more than the chart. Drop the word engagement entirely and say what is happening to the money instead. Here are the shapes that work in the room.

• "We paid for 20,000 installs last quarter and 1,900 of those people were still active at day thirty, so our real cost per kept user is around ten times what the dashboard says."
• "Our day one retention is 41 per cent. We treat anything under 50 as a warning, so we are below the line we set ourselves."
• "Every point of day-thirty retention we win takes cost out of next quarter's acquisition budget."
• "The product is growing its installs and shrinking its active base at the same time."

Each of those sentences does the same job. It takes a figure the board has no feel for and attaches it to a figure they approve every quarter.

Say the number you want

Finish with a target rather than a worry. "We want day thirty at 15 per cent by the end of Q3, up from 9" gives a board something to agree to. A sentence that ends with concern about user experience gives them something to sympathise with, and sympathy has never released a budget.

The Comparisons That Land: Cohorts, Competitors and Cost Per Retained User

A single retention percentage invites one response, which is to ask whether that is good. Comparisons answer the question before it is asked, and there are three worth preparing.

Cohorts come first. Show January's users next to June's and the conversation changes, because now you are talking about something that moved and a reason it moved. Averages flatten all of that into one unhelpful line.

A cohort beats an average, because averages hide the month when everything changed.

Competitors come second, and app store reviews are the cheapest version of this you will find. Pull the complaints on a rival listing, pull yours, and put the two side by side. Boards respond to a competitor doing better far more quickly than they respond to an internal chart.

Cost per retained user is the third and the strongest. Divide the quarter's acquisition spend by the number of people still active at day thirty. That number belongs in the finance pack, and once it is in there it gets tracked whether the product team pushes for it or not. Growth that comes out of people you already have is worth saying plainly too, with SaaS Capital finding businesses that pull over a fifth of new revenue from existing customers growing at 19 per cent against 11 per cent.

Diagnosing the Cause Before You Ask for the Budget

Walking in with a problem and no cause gets you a request for more analysis. Walking in with a cause gets you a decision. We work through it in a set order.

1. Find where in the first thirty days people actually leave, by day and by screen.
2. Work out what changed in the product in the weeks before the drop started.
3. Ask a group of lapsed users directly, rather than waiting for them to tell you.
4. Separate a problem in the core journey from a problem in the surrounding features.

Step four is where the auction game project turned. The client wanted to keep adding new features to pull people back in, and we watched the engagement rate drop anyway. So we stopped and focused on the core journey of someone arriving and playing the game, rather than all of the ancillary features bolted around it. Adding things to a weak centre makes the centre harder to reach.

The genetics wellness app had a different cause with a similar shape. Our audit found the brand promise was exactly right, a real inside look at how your body works and how it is ageing. The product had been built largely by a developer team who had stripped the storytelling out in favour of functional delivery, so people were looking at their own results and feeling nothing.

Evidence You Need in the Room: Tracking, Feedback and the Audit

Boards ask one predictable question, which is how you know. Three pieces of evidence answer it, and they are stronger together than any one of them alone.

Tracking is the first, and it has to cover retention and engagement over time rather than new users arriving. Simon puts it simply. Don't just look at new users coming into the product, look at how the retention is, look at engagement levels, ask for feedback, all of that kind of thing. On the project between funding rounds, that was exactly the decision we made, because nobody could answer a basic question about who was still there.

Feedback is the second, and it has to be solicited. People who are unhappy mostly leave without a word, so a product that only hears from the users who complain is hearing from a tiny and unrepresentative group.

What an audit adds

The audit is the third piece, and its value is that it names the cause. On the genetics wellness app, the audit was what forced the decision to put narrative back in and give people a sense of identity inside the product, instead of redesigning a brand promise that was already working.

Bring one lapsed-user quote to the meeting, in their words. A sentence from a real person does more than another chart.

Making the Ethical and Experience Case in Commercial Terms

The honest-product argument gets dismissed as soft, which is a shame, because it has the better commercial case of the two. Products that are straight with people hold onto them. Products that trick people into staying produce numbers that look fine for a quarter and then fall over.

Expectation-setting is the cheapest version of this. On a fitness product we worked on, we had to take people through a handful of orientation questions, and one of the changes we made was simply telling them how long it was going to take. Nothing technical. People went into it in a much more prepared frame of mind, and fewer of them dropped out partway through.

The listing is part of the product

The same logic sits upstream of the app itself. When an app store listing says accurately what the app does, the people who download it have already lined themselves up as the right audience. They open it knowing what it is, which takes the abandonment rate right down and means your acquisition spend is buying informed users rather than curious ones.

For a board, that is a quality-of-spend argument. You are buying the same volume and keeping more of it.

Pricing the Fix: Scoping an Ask a Board Can Approve

An open-ended request for design time gets deferred. A scoped ask with a defined output gets a yes or a no, and a no is more useful than a deferral because it tells you where you actually stand.

Set the ask against the leak rather than against last year's design budget. If the quarter's wasted acquisition spend is larger than the cost of the work, that comparison makes the decision for them.

A board approves money against the cost of the leak, so put both figures on one slide.

There is a widely cited figure here that boards tend to already know. Bain & Company put the profit effect of a five per cent lift in retention at 25 to 95 per cent. Use it once, as corroboration, and get straight back to your own numbers.

Setting Expectations on Recovery Time

This is the part teams skip, and skipping it is what makes the next funding conversation harder. Retention does not snap back the week the fix ships. People who had a poor experience have already decided something about your product, and that decision takes time to shift.

On the auction game, once we had put the core journey right, the retention went back up. It took a little while, because of the trust that had been eroded over the months of decline, and it did get back up to the mid sixties and low seventies. That lag is worth saying out loud before you start, because a board that expects movement in three weeks reads a flat month four as failure.

What to promise instead

Promise the leading signals first. Completion of onboarding, day one, then day seven, then day thirty, in that order, over roughly a quarter. Each one moves before the one after it, so the sequence gives you something honest to report at every meeting while the headline number is still catching up.

Say the word trust in the room. Boards understand that a reputation takes longer to repair than a bug.

Handling the Three Objections You Will Get

Three objections come up almost every time, and all three have short answers.

"Can't we just spend more on acquisition?" You can, and the leak scales with it. Every extra pound buys users at the same retention rate you have now, so a bigger budget on a leaking product buys a bigger leak. Show the cost per retained user at current spend and at double spend, and the answer is on the page.

"Isn't this just a redesign?" Point at the diagnosis. On the genetics wellness app, the fix was restoring the storytelling and giving people a sense of identity inside the product, and the brand promise was left exactly where it was. Naming the cause is what separates a fix from a refresh.

"How do we know it will work?" Scope the first step small enough that the question answers itself. Tracking and proactive feedback cost little and tell you where people leave. Then come back with the evidence and ask for the bigger piece. A board that has seen one honest, modest result funds the second ask far more readily than the first.

Reporting Back So the Next Conversation Is Easier

The second retention conversation is shaped entirely by how you handled the first. Report in the same format every time, with the same four figures, in the same order. Cost per retained user, day one, day seven, day thirty. Boards learn a slide they see repeatedly, and after two quarters they start asking about it themselves.

Include what moved against what you said would move. If day one lifted and day thirty stayed flat, say so plainly and say which part of the plan is still running. Burying a flat number is the fastest way to lose the credit you earned.

Keep one cohort chart on the same slide all year. The shape of three cohorts next to each other tells the story without a single sentence of explanation.

Carry the feedback through

Keep asking users how things are going, and keep bringing a handful of their words into the pack. The numbers tell the board what happened and the quotes tell them why. Over time that pairing turns retention from a product team's concern into something the whole board watches, which is where it belongs.

Conclusion

Retention becomes fundable the moment it stops being described as a product health issue and starts being described as the cost of everything else. The install spend, the payback period, the lifetime value target the board already set. All of it runs through how many people are still there on day thirty.

The pieces are straightforward enough. Know where people leave and when, know why, price the leak against the fix, set a sensible expectation about how long trust takes to rebuild, and report the same four numbers every quarter until the board starts quoting them back at you. On the auction game, that meant leaving the feature list alone and rebuilding the core journey. On the genetics wellness app, it meant putting the storytelling back in so people felt the product was speaking to them.

Neither was a redesign for its own sake. Both were decisions about where the money was going and whether it was coming back.

If you are heading into a board conversation with a retention problem and the wrong vocabulary for it, we are happy to help you shape the argument. Let's talk about your retention numbers.

Frequently Asked Questions