A custom CRM is paid for once at build time, then every year in maintenance. A subscription is paid per seat, every month, for as long as your team uses it. Which one costs less depends mostly on how many people log in, and the calculation fits on the back of an envelope.
Your team keeps a spreadsheet next to the CRM
You pay for a CRM and someone still updates a spreadsheet alongside it. That spreadsheet has two possible causes, and they call for very different spending.
First cause: the step does not exist in the software. Your deals go through three internal approvals, the CRM only knows a "proposal sent" status, so the rest gets tracked elsewhere. Commissioning a build makes sense here.
Second cause: the step exists and nobody fills it in. The fields stay empty because completing them takes four clicks too many, or because nobody explained what they feed into. A new tool will be empty the same way.
Open the spreadsheet and read the column headers. If they map to fields that already exist in your CRM, the software is not your problem. If they describe a stage of your business the software has never heard of, you have your answer.
Running the five-year numbers
Subscriptions are billed per seat, per month. Take the listed price, multiply by your seat count, then by sixty months, and add the setup fee where the vendor charges one.
Published list prices at the time of writing: Pipedrive charges US$39 per seat per month for Growth on annual billing, US$59 for Premium. HubSpot charges US$90 per seat per month for Sales Hub Professional on annual billing, plus a mandatory one-time US$1,500 onboarding fee, and from US$150 per seat for Enterprise with a US$3,500 onboarding fee.
| Seats | Subscription at US$39 per seat, over 5 years | Subscription at US$90 per seat + US$1,500 onboarding, over 5 years |
|---|---|---|
| 5 | US$11,700 | US$28,500 |
| 8 | US$18,720 | US$44,700 |
| 12 | US$28,080 | US$66,300 |
| 20 | US$46,800 | US$109,500 |
A custom tool works the other way round: paid at build, then maintained. My rates are published on the internal tools page: a targeted tool starts at €7,140 and a business application at €16,660, at a day rate of €595. Maintenance depends on how much you change. Budget six days a year and you add €3,570 annually, or €17,850 over five years, which puts a €16,660 application at roughly €34,500 across the period, hosting excluded.
US vendor prices are listed in dollars and exclude VAT. Convert them at the day's rate and add VAT before comparing them with a euro quote, otherwise you are comparing two columns in two different currencies.
Three situations where building is the wrong call
The arithmetic does not settle everything. Three cases give a negative answer even when your calculation points to building.
Your process is still moving. A custom tool encodes one way of working. If you have reworked your sales cycle twice this year, every change will come back as billable days. Wait until you have held the same process for a year.
Data entry is the real problem. If your records are patchy because nobody fills them in, a new system will be patchy too. Look first at who enters what, and at what point in the day. It is the same precondition as marketing automation: without machine-readable data, there is nothing for the system to act on.
Your team is small. Below five or six seats, none of the subscriptions above reaches the cost of a build over five years. The crossover never arrives, however frustrating the current tool is.
A fourth case is worth naming: the things vendors ship with the subscription that are expensive to rebuild alone. Bulk email deliverability, mobile apps, security certifications, access to data enrichment databases. If those matter to you, keep them on the subscription side and have the rest built around them.
What the quote leaves out
A development quote covers the build. Four lines rarely appear on it and get paid anyway.
Hosting and backups, billed monthly like a subscription, for much less.
Dependency upgrades. The tool sits on components that ship security patches. Skipping them for three years turns a routine update into a project.
Code ownership. Ask for it in writing, along with the repository and the documentation, before you sign.
The exit. If your provider stops, who picks it up? An outside developer should be able to reinstall the project from the documentation alone. Ask that at the first meeting.
That is what drove the architecture behind Coddy, my own urban escape game product: the booking, payment and instruction-sending flow across nine countries did not fit any subscription on the market, so it was built, with maintenance budgeted from day one.
Where to start if the numbers point to building
- Count the seats actually in use, not the licences you pay for.
- Multiply the per-seat price by sixty and add any mandatory onboarding fee. That is your five-year subscription cost.
- Write down the three things your team does outside the CRM. Those define the scope worth building.
- Ask for a quote split into two lines: build and annual maintenance. Maintenance left off the quote still gets billed, by the hour.
- Compare, and keep 20% headroom on the build line.
If the gap between the two totals stays under 20%, stay on the subscription and have the missing pieces automated around it. Automation built around a subscription goes live faster and can be removed without breaking the rest.
Key takeaways
- Do the seat maths before any other conversation: per-seat price × sixty months, plus mandatory setup fees.
- Mandatory onboarding fees of US$1,500 to US$3,500 change the answer at small headcounts, because they spread across very few seats.
- If your process changed twice this year, do not commission a build now. Every change will come back as billable days.
- Insist on a quote split into build and annual maintenance. Without the second line you are not comparing like with like against a subscription.
- If your records are patchy because nobody enters the data, a new tool will be patchy too. Fix data entry before buying anything.
