We have been called in to rescue enough abandoned CRM systems to notice a pattern. The software is rarely at fault. The system was configured around how management wished the business worked, rolled out with a two-hour training session, and then quietly ignored by the people expected to maintain it. Six months later the data is stale, the reports are wrong, and someone suggests buying a different platform.
For service businesses — agencies, engineering firms, consultancies, maintenance contractors — a few specific lessons come up repeatedly.
Model the pipeline you actually have
Default CRM stages come from product sales: lead, qualified, proposal, negotiation, closed. Service work frequently does not behave like that. An enquiry might sit dormant for eight months until a client’s budget cycle opens. A site visit may be needed before anyone can quote. Half your work may arrive as repeat business from existing clients and skip the early stages entirely.
Sit with whoever wins the work and map what really happens, including the waiting. If dormant-but-alive is a real state in your business, it needs to be a stage, otherwise those opportunities either clog the pipeline or get deleted. Forecasts built on stages that do not reflect reality are worse than no forecast, because people act on them.
Fewer fields, filled in properly
There is a strong temptation during configuration to capture everything that might one day be interesting. Industry, company size, referral source, decision-maker role, preferred contact method, budget band. Each field costs the salesperson time and none of them get filled in reliably.
Pick the small number of fields you will genuinely use in a report or a filter, make those mandatory, and delete the rest. A record with six accurate fields beats a record with thirty, of which twenty-two are blank and four are guesses. If someone later needs a new field, adding one is a ten-minute job — much easier than recovering from a form so long that nobody completes it.
Give the team something back
This is the point most implementations miss. If the CRM only extracts information from the team so that management can produce dashboards and reports, the team will treat it as an administrative tax and do the minimum. Adoption improves sharply when the system saves them work.
Generating a quotation from the opportunity record rather than editing last month’s Word file. Seeing every previous job and complaint for a client before walking into a meeting. Automatic reminders so nobody has to keep a private follow-up list. Contact history visible on a phone during a site visit. These features are what make people open the system voluntarily — the practical side of designing systems people find easy to use — and voluntary use is the only kind that produces reliable data.
Decide the rules about ownership
Service businesses often have several people touching one client — a director who owns the relationship, a project manager delivering work, an account handler chasing payment. Ambiguity about who owns a record leads to duplicate follow-ups, awkward moments with clients, and internal friction over commission.
Agree explicitly who owns an enquiry, what happens when it is passed on, and who is responsible for the record after the job completes. Then configure the system to enforce it. Written rules also make holiday cover and staff departures far less disruptive.
Clean the data before you migrate it
Importing years of accumulated contacts from spreadsheets and old mailboxes feels like progress. It usually imports duplicates, people who left their companies in 2019, and addresses that bounce. The new system inherits the old system’s reputation for being untrustworthy on day one.
Import a deliberately narrow set — active clients, live opportunities, contacts touched in the last two years — and archive the rest somewhere retrievable. A smaller trustworthy database gets used. A large questionable one does not.
None of this is about the choice of platform. It is about matching the configuration to how the business genuinely operates and making sure the people entering the data get something useful out of it. Get those two right and almost any competent CRM will work.
