Choose a CRM by starting with your sales process, not the feature list. Write down the real decisions a customer makes between first enquiry and signed work, make those your pipeline stages, then check which tools can hold that shape without custom hacks. After that, test three things: who owns each record, what the CRM must connect to, and whether you can export everything if you leave. Price matters last.
Start with the decisions your customer makes, not the stages your CRM suggests
Every CRM ships with default stages: New, Contacted, Qualified, Proposal, Won. They describe what your team did, not what the customer decided. That is why pipelines rot. Deals sit in "Contacted" for nine weeks and nobody can say why.
Better stages mirror a decision only the customer can make. For a service business that usually looks like: enquiry received, spoke to a human, agreed the problem is worth fixing, seen the price, made a decision. Each move forward needs evidence, not optimism.
Run this test on your current pipeline. For each stage, ask: what did the customer do to get here? If the answer is "we sent them something", it is not a stage, it is an activity. Merge it or delete it.
Keep the count between five and seven. Fewer, and you cannot see where deals die. More, and your team stops updating them, which is worse than having no CRM at all.
The two stages most service businesses are missing
The first is a qualification stage with a written definition. Not a feeling. Something like "has a budget range, a timeline and authority to sign". Without it, your close rate is meaningless because the denominator is junk.
The second is a stage between "won" and "started". The sales-to-delivery handoff is where service businesses lose money quietly, through missed context and slow starts. We covered the detail in the sales-to-delivery handoff checklist.
Decide who owns each record before you decide which tool
Ownership is the question that decides whether a CRM stays accurate. Every contact, deal, task and note needs one named person responsible for it. Not a team. A person.
Write it down before you buy. A simple version:
| Record | Owner | Who else can edit | Who must be notified |
|---|---|---|---|
| New enquiry | Whoever responds first | Sales manager | Sales manager, daily |
| Open deal | Salesperson named on it | Sales manager | Owner, weekly |
| Won deal / account | Delivery lead | Salesperson, read plus notes | Finance |
| Company record | Operations | Sales, restricted fields | Nobody |
Then check the shortlist can enforce it. Some tools let anyone edit anything, which sounds flexible and ends with three versions of the same company and a deal nobody claims. You want field-level permissions, or at least record ownership plus an audit history so you can see who changed what.
This is one of our build principles: clear permissions, and human review where money, legal commitments or messages in your name are involved. On the MRM Group build, agreement-before-payment flows and audit history existed for that reason, with human approval at sensitive financial-promotion decision points. You can read the MRM Group case study for the shape of it.
Work out what must connect, and what only needs to be visible
Integration is where CRM choices get expensive. Split your list into two columns before you look at any tool.
Must write data both ways. Your website and enquiry forms. Your calendar and booking tool. Your phone system, if calls are a real channel. Your email, so replies attach to the record automatically without anyone copying and pasting.
Only needs to be visible. Accounting, project delivery, support tickets. These often work better as a read-only view or a dashboard than as a two-way sync you spend six months debugging.
Speed to lead is the integration that pays for itself fastest. Harvard Business Review's "The Short Life of Online Sales Leads" found that firms trying to contact a web lead within an hour were nearly seven times as likely to qualify it as those who waited even an hour longer, and more than 60 times as likely as those waiting 24 hours or more. Of 2,241 US companies audited in that 2011 study, 23% never responded at all. If your form does not create a CRM record and a notification within seconds, nothing else on the feature list matters.
Watch the connector costs too. If you plan to glue everything together with Zapier, note that each successful action step counts as a task, triggers and filters do not, and at your plan limit new runs are held until you pay for more. Ten automations across a busy pipeline add up quickly. We wrote about the crossover point in Zapier vs custom automation.
Score your options: a seven-point test
Give each CRM on your shortlist a score out of two for each line. Zero is no, one is partly, two is yes.
- Can the pipeline hold your real stages without renaming things into nonsense?
- Can you set record ownership and restrict who edits sensitive fields?
- Does it connect two ways to your website forms, calendar and email?
- Can you export everything, including notes, activity history and communications, in a usable format?
- Can a new team member log an enquiry correctly with fifteen minutes of training?
- Does reporting answer your top five questions without a spreadsheet?
- Is the total annual cost, including seats, add-ons and connector tasks, something you would pay happily in year three?
Ten or above, it is workable. Below seven, keep looking. Question four is the one people skip and regret, because export quality decides whether leaving later costs a weekend or a quarter.
The signs you have outgrown the CRM you have
Not every frustration justifies a move. Migration is disruptive and the tool is rarely the whole problem. These are the signals that genuinely mean the tool is the constraint.
- You pay for features you do not use, and the ones you need are locked in a higher tier or a separate product.
- Your team keeps a spreadsheet alongside the CRM because the CRM cannot hold something they need. One spreadsheet is a warning; three is a decision.
- You cannot answer "how many enquiries came in last month and what happened to each one" without manual work.
- Reporting requires exporting to a spreadsheet every week, which we cover in the 9 numbers every service business dashboard needs.
- Permissions are all-or-nothing, so people see client financial data they should not.
- Adding a new service line or role means fighting the tool's assumptions.
TMMB Academy arrived with a fragmented funnel, rising software subscriptions and no single system. The fix was one portal with role-specific workspaces for admins, coaches, creators, sales, assistants and brands. The implementation ran for more than 60 days. The TMMB Academy case study sets out what that covered.
When not to move
If your stages are undefined, nobody updates records and there is no agreed owner, a new CRM will fail in the same way with a different logo. Fix the process in the tool you have for one quarter first. If discipline holds and the tool still blocks you, move with evidence.
Equally, if a well-configured off-the-shelf CRM covers your process, use it. Our projects start at £4,500, and a custom build is only worth it when the tool is genuinely the limit. Custom software vs SaaS sets out that line honestly.
Plan the exit before you commit
We migrated our own CRM out of GoHighLevel into a system we built. That meant more than 1,600 contacts, around 100 deals with values, over 11,000 emails, over 1,000 text messages, nearly 200 calls and several thousand activity records. We built the import to be re-runnable and keyed on original IDs, carried do-not-contact settings across, and ran both systems side by side before switching anything off. It is not yet finished.
Two lessons apply to your choice. First, communications history is the hardest thing to move and the easiest to lose, so check how a tool exports emails, texts and call records, not just contacts. Second, do-not-contact and consent flags must survive the move intact. Under UK PECR guidance, automated marketing calls need specific prior consent, and general marketing consent is not enough. Losing those flags is a compliance problem, not an inconvenience.
The step-by-step method is in CRM migration without losing data. When we do this for clients through CRM setup and migration, it is built in your own accounts, and you own the code and the data.
Do this week
Open your current pipeline and list every stage. Next to each one, write the customer decision it represents. Any stage where you cannot name a decision gets merged or deleted.
Then pick one deal from each remaining stage and check the record has an owner, a next action and a date. If most do not, you have a discipline problem, not a tool problem, and that is cheaper to fix.
If the shape of your process genuinely will not fit the tool you have, book a call and we will map how enquiries arrive and where work gets stuck before anyone talks about software.
