The PSP RFP playbook for fintech operators

If you have ever sat through a PSP RFP, you know the pattern. Six providers send back beautifully formatted decks. Five of them score within two points of each other. The decision drifts toward whichever sales team has the best relationship with your CFO. Six months later, you discover that the winner's approval rates in your top market are five points below their competitor's.

The problem is rarely the providers. It is the RFP.

Stop asking what they support. Ask what they deliver.

Most RFPs are 80% feature checklists. "Do you support Apple Pay? 3DS2? Tokenisation? Account updater?" Every serious PSP ticks every box. The checklist tells you nothing.

Replace feature questions with performance questions tied to your traffic profile:

• "On Visa credit cards issued in [your top market], what was your median approval rate over the last 90 days for merchants in [your vertical] with a ticket size between X and Y?"
• "Show me your network token penetration for the same segment."
• "What was your median latency for an authorisation in [your region] last month, at the 95th percentile?"

If a provider cannot answer these in concrete numbers, you have learned something important.

Bring your own data

The single most powerful move in an RFP is to give every shortlisted PSP the same anonymised sample of your real traffic and ask them to score it. The variance in the responses is striking. You learn more from this exercise than from twenty pages of capability slides.

Negotiate on the right line items

The discount on the per-transaction fee is the one number that everyone fixates on and that often matters least. The items that actually move your blended cost are:

• Interchange-plus markup versus blended pricing
• Scheme fee pass-through (and the cap on it)
• Cross-border and currency conversion margin
• Chargeback fees and representment success
• Minimum monthly commitments

Build a model that shows total cost of ownership on your projected volume, not the headline rate.

Decide on operating model, not just price

Finally, ask yourself an honest question: do you want a PSP that operates as a black box, or one that exposes its routing logic and lets you tune it? Both are valid. Pretending you want the second when you really want the first leads to disappointment on both sides.

A good RFP ends with two pieces of paper: a scoring matrix and a clear narrative about why this provider, for this stage of your business, for this period. Revisit both in eighteen months.