Put ten mid-market consumer brands side by side, as I do for a living, and the technology slide is close to interchangeable. Shopify Plus. Klaviyo. Meta and Google for acquisition. A reviews widget, a subscription app, a returns portal, a CDP that was expensive and is mostly a segmentation tool. The logos rearrange slightly by category. The architecture is the same building.
This is worth saying plainly: that sameness is not a failure. It is what a maturing market looks like. A decade ago, stack choices were genuine bets, and backing the wrong platform could cost you two years and a rebuild. The bets have settled. The defaults are good, everyone has them, and the app stores have turned yesterday’s competitive edge into this year’s checkbox. Tools commoditise. They always have.
The uncomfortable arithmetic
The consequence follows on its own: if everyone owns the same tools, the tools cannot be why anyone wins. Yet an enormous amount of roadmap energy still behaves as if they can. Best-practice playbooks are applied to best-practice stacks, and the result is the inboxes you and I both scroll: the same welcome flow, the same abandonment sequence, the same ten-per-cent pop-up, brands distinguishable mainly by their hex codes. Sameness compounds when everyone optimises toward the same benchmarks with the same instruments.
AI turns the volume up on all of it. The same models are now embedded in the same apps, producing the same competent output at far greater speed. This is the same shift desktop publishing and home recording brought to their crafts: access went up, and the distance between good and great went up faster, because the ceiling moved for people who knew exactly what they were doing with the new tools. Amplification, not democratisation. Feed a generic brand through the AI features of a standard stack and you get the industry average, delivered instantly.
Where the difference enters
So the differentiator has to live in what feeds the stack, because it can no longer live in the stack. The proposition the tools are asked to express. The customer knowledge that makes segmentation mean something. The taste that decides what the welcome flow should feel like rather than convert at. The numbers the business actually trusts. Two brands on identical technology will diverge exactly as far as those inputs diverge, and the tools will faithfully amplify whichever brand knows itself better.
The practical move is to stop auditing the stack for advantage and start auditing what you pour into it. The stack review takes a day and changes little. The input review is harder and changes everything downstream, because every tool in the building multiplies it.
What is best practice, after all, if everybody does it? Just average practice.