Churn, retention, and expansion for platform resellers

The first year of an account is the least profitable one

If you run a book of client accounts — an agency, a systems integrator, a reseller with a vertical practice — profit is not spread evenly across the life of an account. It's back-loaded, and badly.

Year one carries everything expensive: the sales cycle you paid for before you knew whether you'd win, the discovery, the data migration, the workflow modeling, the two rounds of "actually, we do it slightly differently in the Cleveland office." Most of that is delivery labor, and delivery labor is the least leveraged thing you sell.

Year two is different. The system is configured, the client's people know how to use it, and your involvement drops to change requests and support. Similar revenue, a fraction of the cost. Year three, better still.

Two consequences follow. Churn is far more expensive than it looks: losing an account at month fourteen doesn't cost you next year's revenue, it costs you the only profitable years and leaves you holding the acquisition bill. And expansion inside an existing account is the cheapest revenue you have, because the discovery cost is already sunk — you know the operation, you have the relationship, no new sales cycle to fund.

Most partner practices know this and still run as if new logos were the growth engine. They aren't.

"The client is happy" is not a retention metric

The failure mode is almost never a client who tells you they're unhappy. Unhappy clients complain, and complaints are recoverable. The accounts you lose are the ones that went quiet.

Happiness is a lagging, self-reported, politeness-distorted signal, and the person who cancels is usually not the person you talk to. Here's what actually leads:

Usage decay in the workflows that matter. Not total logins — logins stay flat for months after a system stops being load-bearing. Watch the workflows the system was bought to run. If the approval step that used to fire forty times a week is firing eight, the work moved somewhere else, and somewhere else is usually a spreadsheet. That's the earliest honest signal you get, and it lives in state transitions and event history, not a satisfaction survey.

The champion leaving. The highest-risk event in an account's life is the departure of whoever sponsored the purchase. Their replacement inherited your system without inheriting the reasons for it, and their first instinct will be to review it. Treat a champion departure as a re-sell, starting the week you hear about it.

Support tickets going quiet rather than loud. A drop in tickets reads like success and is often the opposite — engaged users file tickets because they're pushing on the system. Silence can mean "working perfectly" or "we stopped relying on it," and the queue alone can't tell you which. Cross it against usage.

Anyone asking about export. A data-export question is rarely idle curiosity. It's due diligence, and it usually means someone is already evaluating an alternative. It's also the last moment you can act, so answer it straight and then have the real conversation.

Usage-based economics change the expansion motion

Reselling on a per-seat license, expansion means a renegotiation: more seats, a new contract line, a purchasing conversation with someone told to hold budget flat. That friction is why much expansion that should happen doesn't.

Usage-based economics change the shape. On FastYoke the apps cost nothing — the free Logistics core, CRM Suite, Inventory, Accounting, Project Tracker, Field Service, and Forms all install at $0, with Warehouse Management in early access. No per-app license, no seat count. You pay for platform usage above a free tier, either pay-as-you-go or bundled into an Enterprise Platform arrangement that adds region pinning, SSO, dedicated compute, and an uptime SLA. So you can land one department and grow inside the account without renegotiating seats: a second team starts working in the same workspace and the bill follows the work, not a headcount audit. Expansion becomes an operational decision the client makes, not a procurement event you have to survive.

Be honest about the other edge. Metered revenue moves with the client's activity in both directions. When they grow, you grow without selling anything. When they have a soft quarter, your revenue softens with theirs — and unlike a locked-in seat count, there's nowhere to hide it. Plan your cash against a book that breathes; don't model metered revenue as a subscription.

The expansion paths that actually work

Three, in rough order of reliability.

A second department. You've proven the thing in operations; sales or finance is watching. The easiest expansion available, because the hard objection — "will it work here?" — was answered next door.

A second app on the shared data model. This is where one platform beats a stack of point tools. CRM records and Field Service jobs link automatically: one system, not two products joined by webhooks somebody maintains forever. Adding the second app is configuration, not an integration project, so it fits inside a normal engagement.

An integration that makes the system load-bearing. Once the platform is the source of truth feeding the client's accounting, customer portal, or reporting, removing it stops being a software decision and becomes an operations project. That's the strongest retention move you have. What doesn't work: selling more of the same to the same users, and "strategic" expansion nobody asked for.

Quarterly reviews worth the client's time

The review that gets cancelled is the one where you present uptime numbers and ask if everything's okay. Make the hour earn itself:

• Bring usage by workflow, with the trend. Say which are growing and which have gone quiet, and ask why about the quiet ones.
• Bring the audit trail. With an append-only event log, "who changed this and when" is a query — answering that live beats a quarter of reporting.
• Name one thing you'd remove or simplify. It costs a little revenue and buys a lot of credibility.
• Confirm who the champion is, every time. Org charts move quietly.

Price expansion so success isn't punished

The fastest way to stall growth inside an account is to make the next increment of usage feel like a penalty. If your markup compounds against a growing base, the client's reward for leaning on your system is a bill that outruns the value — and the next department quietly never gets onboarded. Price it so that department is obviously worth it, and be transparent about what meters.

The underlying billing is forgiving by design: a prepaid wallet debited by usage, a low-balance alert while there's still runway, suspension only at zero, credit lifting it automatically, an immutable usage ledger behind all of it. The debit is deliberately soft — a legitimate workflow transition is never failed over a billing edge case, which matters at 4 p.m. on a Friday.

What to watch for

The formal partner program is still being shaped. FastYoke has named partners today — Pay n Go Systems as a Channel Partner in retail, point-of-sale, and food service, and iNetko as a Strategic Partner for implementation, support, sales engineering, and consulting. But published revenue-share terms, self-serve signup, and the partner-facing management surfaces are not generally available. If you're building a practice around this, the honest next step is a conversation: partners@fastyoke.io.

Per-customer isolation is a capability, not a business model. Each tenant gets its own database file — excellent for isolation and clean exits. It doesn't by itself make an account sticky. Nothing does except the work being genuinely better inside your system.

Retention work is unglamorous and gets deprioritized. It has no launch date and nothing to announce. Put it on a named person's calendar or it loses every week to whatever is on fire.

The takeaway

Pick your five largest accounts and, this quarter, pull the usage trend by workflow for each. Not logins — the workflows they bought the system to run. You'll find at least one account that looks fine and isn't, while there's still time to do something about it.

Further reading: pricing for how platform usage is billed, the marketplace for what installs on the shared data model, and channel partners for where the program stands today.