Investing App Gamification Got Me Three Taps at a Time. The Fix Was Never More Willpower.

The three worst trades I ever made took about forty seconds each.

Not forty seconds of thinking. Forty seconds total, in bed, from waking the phone to owning the thing. The long version of what those taps cost me is here. Short one: I got lucky on a single stock in my early twenties, decided that meant I could read markets, and went hunting for the next one. It was crypto.

For years I filed that under character defect. The internet now files it under investing app gamification, which is closer, and still not quite right.

Because here is what I actually remember, and it is not confetti. No leaderboard. Nothing flashed. What there was, every time, was a list of things that had moved, sorted by how much they had moved, a buy button under each one, and an amount field that already had a number in it. I never sat down and researched Cardano. I scrolled to it.

That distinction decides what you do about it. If the problem is that the app is too fun, the fix is discipline — look less often, be more patient, grow up. I ran that experiment for a couple of years. It holds beautifully until a Tuesday evening when you are tired.

What worked had nothing to do with my personality. I changed where the money lived and which screen I opened first, and I picked the new arrangement with one question: what does this company earn in a year where I do absolutely nothing?

That question has an uncomfortable answer, including for the broker I use now.

There is no neutral screen. Somebody chose every one of them.

I do product management for a living, so let me say the least flattering true thing about my own trade.

Every screen you have ever seen in a money app was picked by a person in a meeting, and that person was measured on something. Almost certainly not on your retirement. Activation, retention, orders per active user. Nobody in that room is a villain, and I have sat in enough of them to know it never feels sinister from the inside. You ship the version that moves the number you are accountable for. Do that a few hundred times, each decision small and defensible, and you have built an app that produces orders.

So "the home screen shows your daily percentage change" is not a fact of nature. It is a choice, and somewhere there is a document explaining why it beat the alternative in a test.

Regulators have started treating it that way. IOSCO surveyed 30 securities authorities across 26 jurisdictions for its May 2025 review and catalogued ten families of digital engagement practice. One of them is simply default settings. Not the loud stuff. The ones you never opened.

The features that got people were not the ones everyone complains about

Now the part where I have to be careful, because there is a version of this article that asserts gamified apps ruin investors, and the evidence does not say that.

The cleanest test anyone has run is British. In June 2024 the UK's FCA published a randomised experiment with 9,140 participants, testing four app features against a control group that averaged 13.6 trades. Two moved the needle. Two did nothing.

Data: FCA, "Digital engagement practices – a trading apps experiment" (UK, June 2024, n=9,140).

Points and a prize draw: 12% more trades. Push notifications: 11% more trades. A trader leaderboard: nothing significant. Flashing red and green prices: nothing significant either. None of the four shifted final portfolio riskiness overall, though the FCA does say the features raised the share of risky trades by 6 to 8%, with bigger effects among people with low financial literacy, women, and 18 to 34 year olds.

Those results tell you where to spend your afternoon. The colours did nothing measurable. The thing that reached into somebody's evening uninvited did. The app does not beat you on the screen. It beats you on the lock screen.

The observational picture is bigger and messier. The FCA went through 176,159 active UK investors for a paper in April 2025 and found the median user of a high-engagement app made seven times more trades than the median user of a quiet one: 16% averaged 10 or more trades a month, against 1.2% elsewhere. Damning, until you read who these people are. Median age 34 against 40. Median portfolio £1,255 against £13,348. And the FCA says virtually all the underperformance it measured came from crypto and contracts for difference, not equity screens.

Those are not the same person on two apps; they are two different lives, and one of them has a tenth of the money to lose.

Same caution for the receipt everyone reaches for at the end of these articles. Morningstar's Mind the Gap put the US investor return gap at 1.2 percentage points a year over the decade to December 2025, and I have leaned on it myself. It is now contested: a paper in the Financial Analysts Journal in May 2026 re-ran Morningstar's own data and got 0.10% a year, twelve times smaller. I am not building a case on a number two credible groups disagree about by a factor of twelve.

What survives is modest. An 11% nudge, aimed at the one decision that compounds, repeated monthly for fifteen years. Modest is not nothing.

Investing app gamification is the symptom. The revenue line is the cause.

Every "is your trading app manipulating you" piece I read while writing this describes the psychology and then ends. Nobody turns it into a decision you can make on a Sunday. But design follows revenue, and revenue gets published. So I stopped arguing about screens and went and read earnings releases. Dullest hour behind this article. Also the only one that changes anything.

Data: Robinhood, Interactive Brokers and Charles Schwab FY2025 earnings releases (US).

In FY2025, 58.8% of Robinhood's net revenues came from customers transacting (US, own filings). In the same year Interactive Brokers took 57.4% of net revenues from net interest income and Charles Schwab took 49.1% from the same line, with another 27.2% from asset management fees (both US-listed, both their own filings).

Read that as three different bets on your behaviour. One needs you to move. Two need you to leave cash sitting where they can lend it. Not one of them is primarily paid for you quietly owning a fund.

The 2026 detail I cannot stop thinking about sits in Robinhood's Q2 numbers, reported in July: event contracts brought in $156m in a single quarter, up more than 10x year on year (US). Whatever you think of prediction markets, that is a screen finding a new thing for you to tap.

Europe deleted the incentive. The US left it in place.

The tidiest proof that this is a business-model story happened in Europe this summer, and almost nobody wrote about it.

Payment for order flow is now banned across the EU. The prohibition is Article 39a of MiFIR, in force since March 2024, with one carve-out: member states where firms were already taking PFOF could exempt them until 30 June 2026. Germany was the only country that notified ESMA it was using that carve-out. It expired. Since 1 July 2026 the ban has been uniform across the bloc.

Now the interesting bit. Börsen-Zeitung reported in November 2024 that order-flow rebates made up roughly a third of Trade Republic's operating revenues, at around three euros per forwarded order (Germany). Take away the legal right to be paid when you trade, and the neobrokers start building revenue around you holding instead: banking, interest, subscriptions. That tells you what the old screen was for better than any behavioural study could.

The US went the other way. In June 2025 the SEC withdrew its proposed rule on conflicts of interest in predictive data analytics, one of fourteen proposals it dropped. PFOF is still legal there. Europe legislated the conflict away; America is relying on you to read a document.

The ten-minute test: what does this broker earn while I do nothing?

Four questions, each with a place you can actually go and look.

Read the fee schedule backwards. Not the commission line, because that is the line they want you comparing. Start at the ones that fire while you are asleep: custody fee, inactivity fee, FX conversion on foreign assets and dividends, what interest they pay you on cash against what they earn on it, securities lending, the cost of transferring out. A provider paid by a flat custody fee or a subscription is aligned with a year of your inactivity. One paid by spread, order flow, FX and idle cash is not.

Find out which behaviour pays their rent. For a listed broker this takes two minutes: open the last annual results release and read the revenue table. Transaction, net interest, asset-based, subscription. For a private European broker, use the press releases and accept a partial answer.

Know what routing disclosure exists where you live. US readers get a quarterly Rule 606 report showing where orders went and what was paid per 100 shares. Europeans are about to get less: new technical standards adopted in April 2026 repeal the old annual top-five-venue reports in favour of internal policies from February 2028. So read the order execution policy instead, and count how many venues it names and whether one market maker dominates.

Audit the defaults on day one. Notifications off, all of them, because that is the only feature the FCA proved causally moves trade counts. Move the home screen off daily profit and loss if the app allows it. Ignore the "most traded this week" list, which is the actual mechanism in the Robinhood herding study that found roughly minus 4.7% abnormal returns over the next 20 days for the most extreme episodes (US, 2018 to 2020). Check whether the buy flow pre-fills an amount.

One honest limit before you optimise this to death. Execution quality does vary: a Journal of Finance study in October 2025 placed 85,000 simultaneous market orders through five US brokers and found round-trip costs from 0.07% to 0.46%. The same authors found payment for order flow did not explain the variation, so "my broker takes PFOF, therefore my fills are bad" is not supported. On a €500 monthly purchase, a 0.10 percentage-point difference is fifty cents. Fund charges and your contribution rate dominate that. Same reason I decided a robo-advisor's 0.25% was the expensive part of that comparison.

No broker earns nothing while you sit still. Mine included.

I have to take my own test now, and I fail the naive version of it.

My monthly world ETF purchase goes through a deliberately boring provider, and the line that pays for a provider like that is net interest. It is the same line I quoted at you earlier: 57.4% of Interactive Brokers' FY2025 net revenues, 49.1% of Charles Schwab's (US, own filings). So my sitting still is not free money at the boring end of the industry either. It is a different business model, wanting a different thing from me, and what it wants is cash left lying around.

So the criterion needs rewriting, and this is the version I use: no broker earns nothing. The question is which of your behaviours their revenue depends on.

That sounds like a smaller claim. It is not, because the two conflicts cost wildly different amounts to defuse. A broker paid on your idle cash wants a balance sitting there, and you close that conflict in one afternoon by not leaving cash at your broker. Closed, permanently, no willpower involved. A broker paid on your trades wants a decision out of you every month for thirty years. There is no afternoon that fixes that.

The idle-cash conflict is not hypothetical. In January 2025 the SEC settled with two Wells Fargo advisory firms and Merrill Lynch for $60m combined over cash-sweep programmes, noting that in rising-rate periods the yield gap between the sweep and the alternatives "at times grew to almost 4 percent" (US). I wrote separately about what that gap does to money you forgot you had. BETTER FINANCE's 2024 neobroker paper flags three more that never reach a European pricing page: securities lending of client stock, currency conversion buried in spreads, and inducements from fund providers for shelf space.

A pair of American law professors gave this the name I wish the debate used. Writing in the Yale Law Journal in 2022, Langvardt and Tierney argue that regulating interface aesthetics is the wrong instrument, because "gamification" has no clean edge and invites endless workarounds. Their alternative is to name the practice behavioural churning: using psychology to generate trading volume for the broker's profit. Judge the business model, not the confetti. Two lawyers got there before I did.

The on-ramp worked. I am not going to pretend it did not.

The strongest argument against everything above is that the frictionless app got a generation invested, and it is true.

Data: Deutsches Aktieninstitut annual shareholder survey (Germany, 2025 data published January 2026).

Deutsches Aktieninstitut counted 14.1 million people in Germany owning shares, equity funds or ETFs in 2025: roughly one in five of the population aged 14 and over, up two million in a single year and up from around 9.0 million in 2016. Around half of German investors under 40 now buy equities on a schedule, against just under a third of the over-40s. Trade Republic reported over 10 million customers across Europe in December 2025, 70% of them first-time investors.

I cannot look at those numbers and argue the app should have been harder to use. The on-ramp is the best thing that happened to European retail investing in my lifetime.

My argument is narrower. The property that made the on-ramp brilliant, that buying takes three taps, is the same property that makes year three expensive. Which of the two the company builds on next is decided by the revenue line, not the design team.

Two apps, two jobs

My setup is boring enough to describe in a paragraph, and I offer it as an illustration rather than a prescription.

One place buys. Money leaves on payday into one accumulating world ETF, automatically, with every notification turned off. I do not decide anything monthly; I deleted that decision years ago and the gap in my life where the decision used to be is the single most profitable thing I own.

One separate place shows me where we are. Opened on purpose, roughly monthly, no buy button in it. That is the tracker I build, and I owe you the disclosure I have been demanding of everyone else: MFFT is paid by subscription, so my revenue survives you opening it twelve times a year, and I am biased toward you renewing. That is an incentive, not a halo. Weigh it accordingly.

The point of splitting the jobs is not discipline. It is that the screen you check on a bad Tuesday is no longer the screen that can transact.

Where investing app gamification stops being the app's fault

I still pressed the button. Nobody pressed it for me.

The measured effects are real and modest. Eleven percent. Twelve percent. Even on high-engagement apps only 16% averaged 10 or more trades a month, which means 84% did not. The FCA says its own 2022 findings are correlational and cannot tell us whether the design caused the bad outcomes. If you want to argue that blaming the interface is learned helplessness, you have decent evidence.

I changed the surface anyway, and not because the surface is irresistible.

I will not be the same person at 9pm on a bad Tuesday in year eleven that I am on a Sunday afternoon with a spreadsheet open. That is not a character flaw I can read my way out of. An 11% nudge is easy to resist once. The cost of resisting it monthly while tired, while a baby is teething, while work is on fire, is the part nobody prices.

Changing the provider and the notification settings costs one afternoon and then costs nothing. Same argument I make about fixed costs: the cheap version is a decision you make once instead of a habit you rent.

Design explains why restraint is hard. It does not excuse the trade.

What I would go and check tonight

If you do one thing after this, do not resolve to look less.

Open your broker's last published results, or its fee schedule if it publishes none, and find out which of your behaviours pays for the building. Then open the notification settings and turn everything off. Those two take less time than the average argument about which ETF to buy, and unlike that argument they keep paying long after you forget you did them.

I would rather have learned this at twenty-five with a small amount of money than at forty with a real one. The three taps were cheap. What they bought me was the discovery that my willpower is terrible infrastructure for a thirty-year plan, and that there is a duller thing available that never gets tired.