Apps that are pleasant to open are expensive to own

Most explanations of app design and investor behaviour stop at the point where it starts to matter. This one carries on.

Engagement is the product goal

Consumer software is built and measured around engagement, because engagement is what a product team can influence and observe. For most applications this alignment is harmless, since using a mapping app more often does not damage the journey.

For an investment account, the behaviours that generate engagement are largely the behaviours that cost money. Checking more often, trading more often and reacting faster all raise engagement metrics and lower expected outcomes. This is a structural conflict rather than a conspiracy, and it exists whether or not anybody intends it to.

The specific design features

Push notifications about price movements convert a decision you would never have made into one you are now considering. Displaying gain and loss against purchase cost on the home screen makes anchoring the default frame for every single visit. Showing the day's change in a large font makes the least meaningful timeframe the most prominent thing on the page.

Put simply, lists of most-traded holdings import the crowd directly into a decision that should be about your own written plan. Each of these features is individually defensible, and collectively they all point in exactly one direction.

Frictionlessness cuts both ways

Removing friction from contributions is genuinely good, because the barrier there is preventing a beneficial action. Removing friction from trading is different, because the barrier there is the only thing between an impulse and a transaction. A well-designed arrangement makes one easy and the other slightly awkward, which almost no consumer product does by default.

The useful part is this: the commission-free era removed the last explicit cost of trading without removing any of the other costs involved. Spreads, timing and reporting obligations all continue regardless of what a platform charges for the deal itself.

What you can change

Turn off every notification that is not about a failed payment or an administrative deadline you have to meet. Remove the app from your phone and use a browser on a computer, which adds enough friction to filter most impulses.

Where the interface allows it, change the default display from daily change to total balance or to contributions made. Log in on a schedule rather than when something prompts you, which breaks the link between news and checking.

None of this is difficult, and all of it is more effective than simply resolving to be more disciplined.

Choosing a provider on this basis

Platforms differ substantially in how hard they push activity, and the difference is obvious within a week of using one. A provider whose interface foregrounds contributions and treats trading as unremarkable is built for a different customer than one organised around a live ticker. This is a legitimate selection criterion alongside charges, and it is almost never mentioned in any comparison table.

The cheapest platform that encourages you to trade stops being the cheapest once the trading has been counted. For most people the interface will influence outcomes more than a small difference in the annual fee ever could.

The honest caveat

Nobody has good numbers on what any specific design feature costs a specific investor, and the research that exists is contested. What is better established is that higher trading frequency is associated with worse outcomes across large samples of accounts. The link between interface design and trading frequency is plausible and partly demonstrated rather than proven in any strong sense.

Acting on it costs almost nothing, which makes it a reasonable precaution even under genuine uncertainty. That asymmetry is the whole argument: turning off notifications has no downside anybody has ever been able to measure.

The takeaway

Delete the app, keep the standing order. The friction belongs on the side that loses money.

The version you keep doing is the version that works.