Why your SMBs use 5.2 providers: The hidden cost of fragmentation

Here's the story most bank boards still hear: fintechs are stealing your SMBs.

Here's what the data really says: your SMBs aren't going anywhere - they're hiring around you.

The average SMB now runs on 5.2 financial providers. You're still the primary checking account, but you're not the primary anything else. Square handles the point of sale. QuickBooks handles the invoicing. Mercury handles onboarding when they spin up a new entity. Brex handles the card. A neobank handles international payments. And then there's you - holding the deposit, and not much else.

When you look at your retention numbers, they look fine. The account is still open, the relationship manager is still on the call, and the balance hasn't moved much - but the operations have.

They didn't leave. They hired around you.

Here are a few examples to demonstrate this:

A retailer wanted to take card in their cafe. Your POS integration was 6 weeks out, so they bought a Square reader on Tuesday. They routed point-of-sale through someone else.
A services firm needed to invoice 40 clients on monthly retainers. Your treasury management module asked them to log in separately and didn't support recurring billing, so they signed up for QuickBooks. They just routed invoicing through someone else.
A founder spun up a second entity. Your business account opening took 7 days, so they opened a Mercury account in under an hour. They didn't switch banks - the original entity's primary checking is still with you. They just routed the new entity through someone else.

Multiply that across every operational moment in an SMB's week - the new vendor invoice, the payroll run, the inventory loan, the international wire, and the corporate card. You will get 5.2 financial providers on average per SMB.

Deposits follow operations

For 50 years, banking treated the deposit as the sticky thing. The primary checking account was the anchor, and everything else was a cross-sell story you could win over time. As long as the deposit stayed, the relationship was intact.

That model worked because the bank was the system of record for how the SMB ran its business. Payments were on the bank's rail. Receivables came into the bank's account. Payroll left the bank's account. The operational moments and the deposit were on the same architecture by default - and that default is now gone.

The operations have moved. Receivables now route through Stripe before they ever touch your account. Payroll runs through Gusto, and only the net debit shows up on your side. International FX runs through Wise, and the deposit converts on someone else's books. The deposit you still see is a residue - what's left after the operations happened somewhere else.

Deposits follow operations, not the other way around. They always move in that direction - just slower, in $50 and $100 increments across years, which is why most banks don't notice until the trend line has already bent.

The "build a better app" trap

The comforting framing of all of this is "we need a better app," but a great app is table stakes - every bank now has one, and every SMB owner has used at least three. The app isn't where the substitution happened - the app is the visible layer of a much deeper architecture story.

Square didn't beat you on UX; it showed up when you couldn't. QuickBooks didn't beat you on invoicing UX; it worked the day the SMB needed it, and yours didn't.

The fintechs that took your share won because they run on one stack, one data model, one workflow, and one source of truth. They built the architecture you didn't, and the slick UI on top is the consequence of that architecture underneath - not the cause.

If the bank's response to 5.2 providers is "ship a better mobile app," the bank has misread the problem. The mobile app isn't the lever. The architecture is.

The hidden cost of fragmentation

The hidden cost doesn't show up on your P&L in a single line. It shows up as a slow leak across the products that actually have margin. The deposit stays, but the yield on the relationship doesn't:

Treasury fees flatten because the SMB routes payments around you.
Lending volume softens because the SMB took the inventory loan from a fintech that decisioned in 24 hours.
FX revenue drops because the wire went through Wise.
Card spend flattens because the corporate card sits with Brex.

There's a second layer of cost on the bank's side too - the integration tax. Every fintech your SMB uses is also a vendor in your reconciliation, your fraud rules, and your compliance posture. Your team is now coordinating across more systems on their side, not fewer, and vendor sprawl on the customer's side mirrors vendor sprawl on yours.

The third layer is the AI ceiling. Every AI initiative you've put against SMB banking - onboarding automation, credit decisioning, exception handling - stalls at the same place. The systems don't share state, the data lives in too many places, and the pilot looks good in a demo before falling over in production. According to Gartner, 85% of AI projects never reach production, and that mortality rate isn't a model problem - it's the same architecture problem.

Your SMBs learned fragmentation from you

Your customers didn't fragment despite your operation - they fragmented because of it.

Your relationship managers toggle between 6+ systems to get a single view of an SMB client. Your underwriters chase documents the bank's core already holds. Your onboarding specialists re-key data between systems that were never built to talk to each other. When your operation is fragmented, your customer feels every seam - in the form of delays, repeated requests, and processes that somehow take days.

So the SMB does what any rational business does when a supplier can't meet their needs. They find someone who can, one operational moment at a time. The 5.2 providers aren't a verdict on fintech innovation. They're a verdict on your architecture.

Architecture created the 5.2 average. It can also close it

5.2 isn't a fintech inevitability. It's an architecture outcome, and an architecture outcome is also an architecture choice.

The banks already running differently aren't competing on features or rate. They're competing on a different operating model - one where the SMB doesn't have to add a fifth provider because the bank can already run that operation natively.

The Unified Frontline is what banks become when they close that gap. The 5.2-provider leak is one half of the picture - the other half is the 7-day onboarding that no AI pilot has been able to fix, and the same architecture problem drives both.