We added 32% to monthly GTV without acquiring a single additional user.
For a stretch at Transfez, every growth conversation I chaired was about the top of the funnel. More installs. Cheaper installs. New channels, new creative, new partners. That is the default gravity of a performance marketing background, and I have it worse than most people.
Meanwhile the largest pool of money in the business was sitting in users who had already downloaded the app and were, right at that moment, trying to send money and not finishing.
The default gravity of acquisition
It is worth being honest about why the conversation kept drifting upward, because I do not think it was an accident.
Acquisition is where most growth leaders grew up. It is the part of the job with the most mature tooling, the clearest vocabulary and the cleanest charts. Spend goes in, installs come out, and you can draw a line between them in a board deck without anyone asking a hard follow-up question. When something is underperforming, the reflex is to ask what you can buy more of, or buy more cheaply.
And acquisition problems are pleasant problems to have. They are solvable from your own desk. You can test a new channel this week without asking anyone for a sprint. You can rewrite creative without a design review. The feedback loop is short and the work feels like yours.
So the meetings filled up with that work. Not because anyone decided acquisition mattered most, but because it was the work most available to the people in the room.
Where the money actually was
The problem is that installs are not the business. Completed transfers are.
Between "I want to send money" and "the money left," there is a sequence of screens. Each one asks the user for something: an amount, a recipient, details, a confirmation, a payment. And at each one, some share of people who fully intended to send money stopped.
These were not cold leads. They were not people we had to convince. They had found us, installed the app, signed up, and arrived at the exact screen where we make our money, with the intention of giving it to us. And we were losing a share of them to friction we had built ourselves.
Once you see the funnel that way, the comparison is uncomfortable. Every percentage point of completion inside the product is worth the same as a percentage point more users at the top. Except one of those you pay for every month, and the other you pay for once.
So we stopped adding users for a while
We stopped adding users for a while and started rebuilding screens.
Not a redesign. Nobody got a new brand, and nobody gave a launch presentation. It was a run of small, unglamorous UI and UX experiments on the steps between "I want to send five million rupiah to my sister" and "the money left."
The wins were boring to the point of embarrassment:
- Fewer fields on screen at once. Every additional field is a small decision, and small decisions add up to a reason to put the phone down.
- The fee and the exchange rate shown before the user commits, not after. Nothing makes someone abandon a payment faster than a number they did not expect, arriving at the moment they thought they were done.
- Recipient details remembered. The second transfer should be less work than the first. For a lot of products it is exactly the same amount of work, every time, forever.
- A confirmation screen that answers "did that actually go through" without the user leaving to check their banking app. Uncertainty at the end of a money transfer is not a minor UX issue. It is the thing people remember.
None of these is an insight. If you showed the list to any product designer, they would nod and say they could have told you. That is sort of the point. The fixes were not hard to think of. They were hard to prioritise, because nobody in a growth meeting was asking for them.
Total GTV went up 32% month over month.
None of that came from marketing. It came from deleting reasons to stop.
Acquisition is the only growth lever you have to keep paying for. Everything you fix inside the product keeps paying you.
Why we prefer the lever we have to keep paying for
I think a lot of growth leaders, me firmly included, prefer acquisition because it feels like our job.
Buying users is visible work with a clean chart and a number you can defend. Fixing a form field is somebody else's ticket in somebody else's sprint. If you are measured on growth and your instinct is to own the outcome, it is natural to reach for the levers you control directly.
But there is a cost to that instinct that does not show up anywhere obvious. Every month you spend optimising the top of a leaky funnel, you are paying full price to pour users into a flow that loses a share of them before they ever transact. The better your acquisition gets, the more expensive the leak becomes, because you are losing more of the users you worked hardest to get.
There is also a compounding difference that is easy to underrate. A channel improvement holds only as long as you keep funding it. Turn off the budget and the users stop arriving. A friction fix inside the product keeps working whether you are paying attention to it or not. It applies to every user who arrives through every channel, including the ones you have not opened yet. It is the closest thing growth has to an asset rather than an expense.
Why this keeps getting skipped
If the economics are this clear, why do growth teams keep under-investing in it? In my experience, three reasons.
Ownership. The flows where the money gets lost usually belong to a product team with its own roadmap. Growth can see the drop-off but cannot ship the fix. So the problem gets raised, logged and deprioritised behind work that the product team is actually measured on.
Glamour. "We removed a field from the recipient screen" is not a sentence anyone wants to put in a quarterly review. "We launched a new channel" is. Teams gravitate toward the work that sounds like strategy, even when the boring work is worth more.
Visibility. Nobody on the leadership team uses the app the way a real customer does. They do not send money on a bad connection, on a mid-range phone, under time pressure. So the friction is invisible to exactly the people deciding where effort goes.
That last one is the easiest to fix, and the one I would start with.
Go and send some money
Before you approve another budget increase, go and complete your own core flow on a mid-range Android on a bad connection. Not on the office Wi-Fi. Not on the newest phone in the building. On the setup most of your users actually have.
Notice every moment you hesitate. Every field you were not sure how to fill. Every number that surprised you. Every point where you wondered whether it had worked.
Whatever makes you sigh is worth more than your next channel test.
The money is already inside the app. Most of the job is getting out of its way.