Perspectives / Where decisions break

The illusion

You added a third PSP. Your approval rate didn't move.

1 min read

First published on LinkedIn, 14 May 2026.

You added a third PSP. Your approval rate didn't move.

Your engineering team spent six weeks on the integration. Reconciliation got more complex. Routing rules got longer. The number you were optimizing for is roughly where it was.

Here is why: more execution options with the same decision logic produces the same outcomes with more moving parts.

The data to make a better decision already exists in your stack. UK Visa above £150 approves better on one acquirer - but not on weekends, and not when the issuer is Barclays.
Mobile transactions below £50 from new cardholders behave differently again. EUR corridors with dynamic currency conversion are a third pattern entirely.

A rule reads one dimension at a time. The decision that maximizes approval on any given transaction requires reading all of them simultaneously - in real time, before execution.

The signal exists in your transaction history. The pattern is visible. The rule cannot see it.

The bottleneck was never the number of PSPs. It was the intelligence feeding the routing decision.

Start a conversation

Have a view on where decisions break?

I read every reply. If something here matches — or contradicts — what you see in your stack, tell me.

Start a conversation partners@pillar8.com

You'll hear back from Yaniv directly.