Ask what a CRM costs and you get a per-seat number: some dollars per user per month. It is the number on every pricing page, the number in every comparison, and among serious products it is the number that varies least. It is also, routinely, under half of what a business actually spends in year one.
This is not usually a scam. It is a mismatch: vendors price the software, and businesses buy an outcome that includes setup, migration, messaging, training, and the hours somebody spends making it all fit. Budget for the outcome and you avoid the two bad endings — the tool that was cheap and never got used, and the invoice in month four that nobody expected.
The seven lines of a real budget
| Cost line | Typical shape | Where it surprises people |
|---|---|---|
| Subscription | Per user, per month | Quoted price often assumes annual prepayment |
| Implementation / onboarding | One-off fee, sometimes mandatory | Can exceed three months of subscription |
| Data migration | Per-record or fixed project | Priced by how messy your data is, which nobody knows yet |
| Messaging and usage | Per SMS, per call minute, per email tier | Scales with success — busy season costs more |
| Integrations | Per connector, or a higher plan tier | The accounting sync is usually on the tier above the one you were quoted |
| Training time | Your own wages | Hours × everyone who touches it, not just the buyer |
| Internal admin | A few hours a month, forever | Real, ongoing, and never in the vendor's estimate |
Three of those seven are your own labour rather than an invoice, which is exactly why they get left out. They are not smaller for being invisible.
How per-seat pricing bites a service business specifically
Per-seat models were designed for sales teams where every seat is a quota-carrying rep. A service business has a different shape: two or three office people who live in the system all day, and eight technicians who need it for ten minutes at a time.
Paying a full seat for a technician who logs three notes a day makes the economics look absurd — and it creates a genuinely bad incentive, where businesses buy fewer seats and have the office re-key everything the field said. That is the exact failure the CRM was bought to prevent.
The seasonal line nobody forecasts
Messaging costs scale with the thing you were trying to increase. Missed-call text-back, appointment reminders, and quote follow-up sequences are all per-message, and they fire hardest in your busiest month, which is also the month with the most missed calls.
This is not a reason to avoid them — the return on a recovered job dwarfs a few cents of SMS. It is a reason to model your peak month rather than your average one, so the July invoice is a plan rather than a surprise.
Sizing the payback in ten minutes
You do not need a business case. You need four numbers you already know: monthly leads, close rate, average job value, and monthly quotes that currently get no follow-up. Then be deliberately conservative:
- Recovered missed calls. Unanswered calls per month × 20% reply × your close rate × average ticket.
- Recovered quotes. Unfollowed quotes per month × 10% recovery × average ticket.
- Faster invoicing. Days saved on average payment × monthly revenue ÷ 30 — this one is cash-flow rather than profit, but it is the one owners feel first.
- Hours returned. Admin hours per week × your loaded hourly cost × 4.3.
For a business doing 60 leads a month at a 25% close and a $1,400 ticket, the first two lines alone tend to land in the low thousands per month. Against a fully-loaded CRM cost of a few hundred, the decision is usually not close — which is why the useful question is rarely "can we afford it" and almost always "will the team actually use it".
Where the money is genuinely wasted
- Seats for people who left. The most common overpayment in small-business software, and the easiest to fix. Audit the seat list every quarter.
- A tier bought for one feature. If you jumped a tier for a single integration, price the integration separately — sometimes a $20 connector replaces a $200 tier jump.
- Migrating history nobody will read. Paying per record to import six years of dead contacts is the purest form of waste in this budget. Move active customers and open quotes; archive the rest as a CSV.
- Paying for two systems in parallel, indefinitely. A two-week overlap is prudent. Six months means the old system won and you are funding both sides of the argument.
What to get in writing before you sign
Four things, all of which a straightforward vendor will send without hesitation: the total first-year number including setup and migration; the per-message and per-minute rates; the price at renewal, not just today; and what the data export contains if you leave. If any of the four takes more than one email to obtain, you have learned something about year two.
Common questions
How much does a CRM cost for a small service business?
The subscription is typically the smaller half of the first-year cost. Budget for seven lines — subscription, implementation, data migration, messaging and usage, integrations, training hours, and ongoing internal admin — and get the total first-year figure in writing before comparing monthly prices.
Why is per-seat CRM pricing awkward for field teams?
It was designed for sales teams where every seat works in the system all day. A service business has a few heavy office users and many technicians who need ten minutes each. Paying full seats for light users pushes businesses to buy too few, which forces the office to re-key field information — the exact problem the CRM was meant to solve.
What hidden CRM costs should I plan for?
The three that are your own labour rather than an invoice — training hours across everyone who touches it, ongoing admin, and the overlap period while both systems run — plus messaging costs, which spike in your busiest month because that is when the automations fire most.
How do I know if a CRM will pay for itself?
Size four things conservatively: recovered missed calls, recovered quotes, days saved on invoicing, and admin hours returned. For most businesses running 40 or more leads a month the first two alone exceed the software cost, which makes adoption, not price, the real decision.