Insights

What to fix before you choose a CRM⁠

29 September 2026By Dan Chapman

Many CRM projects begin with a demo. Someone sees a pipeline view they like, a shortlist forms around three vendors, and the real questions turn up later: who owns the data, what counts as a qualified enquiry, where leads should come from. By then the platform has been chosen, and it now shapes the answers.

The platform is the last decision, not the first. A lot of what decides whether a CRM works is settled before you open a trial account. This guide covers what to settle, how to turn it into requirements, and how to test vendors against your own process rather than theirs. If you want the wider picture of why projects go wrong, read why CRM implementations fail first.

Seven things to fix before you shortlist

None of these needs software. Each has a plain test for whether it is done.

# Fix Done when If you skip it
1 The real process One page shows the stages an enquiry passes through from first contact to paid invoice, as they happen today, not as they should You configure the vendor’s default pipeline and people work around it
2 Owners Every stage has one named person, and the system as a whole has one owner Nobody is responsible for whether it is used well after go-live
3 The decisions it must support You can list the three to five questions it must answer, such as what is in the pipeline, where enquiries come from and what is overdue You buy features nobody uses and still cannot answer the questions
4 Fields and definitions The minimum fields are agreed, and “enquiry”, “qualified” and “won” each mean one thing Reports disagree because people mean different things by the same word
5 Data worth moving You know what you hold, where the master copy of each record type lives, and what you will leave behind You migrate duplicates and out-of-date records into a clean new system
6 Marketing email basis You can record, per contact, what you rely on to send marketing email and when it was given The CRM quietly turns every enquiry into a mailing list
7 Connections You have listed what it must connect to (forms, email, calendar, invoicing, analytics) and who keeps each connection working Data still gets retyped between systems, and the copies drift apart

A few notes on the ones that catch people out.

Process (1) and owners (2) are the ones software cannot do for you. If you have not written down the stages, the lead management process sets out seven you can start from, with suggested standards for each. The useful test is to trace your last five enquiries from first contact to outcome. Wherever the trail goes cold is the first thing to design for.

Decisions (3) keep the field list short. If a field never changes a decision, it probably does not deserve to exist. That is also why I leave lead scoring out at small volumes; an owner and a next step do more. The KPIs for small volumes guide covers which counts are worth reviewing monthly.

Data (5) is more than housekeeping. UK GDPR says personal data must be “adequate, relevant and limited to what is necessary” for your purposes, and the ICO’s guidance on data minimisation says to periodically review what you hold and delete anything you no longer need. A migration is the natural moment. Decide what to move, what to archive and what to delete, and write down why.

Marketing email (6) is where a CRM can create risk that did not exist before. Someone who enquires has not necessarily agreed to marketing email. Two separate rules apply. Under PECR, the ICO’s business-to-business marketing guidance says the rule on electronic mail does not apply to corporate subscribers, while sole traders and some partnerships are treated as individual subscribers. Separately, UK GDPR still applies wherever an address identifies a person, and every marketing email needs a clear way to opt out. The ICO page notes it is under review because of the Data (Use and Access) Act, so check current guidance or ask an adviser; this is general information, not legal advice. What a CRM can do is hold the basis, the date and the source for every contact, and make an opt-out stick.

Turn it into a requirements list

Once the seven are settled, the requirements mostly write themselves. Sort them into three groups:

  • Must: without it the CRM cannot run your process.
  • Should: it would clearly help, but you could work around it.
  • Could: nice to have. Nothing here should affect the decision.

Here is an illustrative example for a professional services firm of about a dozen people. The figures and choices are invented to show the shape, not to recommend anything.

Group Requirement Tied to
Must Website form enquiries create a record with the source captured automatically Fix 1, 3
Must Stages match our six, and each record has an owner and a next-step date Fix 1, 2
Must Marketing email basis, date and source stored per contact; opt-outs respected Fix 6
Must Full export of all data in a standard format, and the ability to delete one person’s record Fix 5, 6
Must Data processing terms we can sign, and clarity on where the data is held Fix 6
Should Calendar sync for discovery calls Fix 7
Should Monthly counts by source, stage and loss reason without a manual rebuild Fix 3
Could Sales forecasting none yet
Not needed Lead scoring at our volumes Fix 3

The last row matters. Naming what you do not need protects you from paying for it.

Test vendors against your process, not their demo

Every vendor can make a demo look good, because they have chosen the data. Give each one the same script instead, built from your own recent enquiries:

  1. Log a real enquiry from a web form, with its source.
  2. Move it through your stages, with an owner and a next-step date at each.
  3. Show everything that is overdue for a reply.
  4. Produce the monthly counts from your definitions.
  5. Record the marketing email basis for two contacts, one an individual and one at a company.
  6. Export everything, then delete one person’s record.

Score each step as yes, partly or no, and note what the vendor had to build or buy to do it. Then ask the questions a demo will not answer:

  • What does this cost at our number of users in year two, including add-ons, integrations and any implementation help?
  • Who configures it, and who looks after it afterwards?
  • What are the contract length and renewal terms?
  • What does leaving look like? In what format can I take my data?

A platform that passes your script, costs what you expected in year two and lets you leave cleanly is usually a better choice than the one with the longest feature list.

When not to buy yet

If you cannot complete fixes 1 to 4, a new system will only make the gaps easier to see, and you will pay for the privilege. The better first step is the process work, and it can start on paper. If your volumes are small and one person handles every enquiry, a spreadsheet with a named owner and a logged next step may serve you for a while. The lead management field map is a free starting point either way.

If you are partway through a shortlist

I am not tied to one platform and I do not earn commission on licences, so the choice comes from your process and your budget. At a B2B marketing agency group I implemented the group CRM as part of a wider operating model; the case study has the detail, and the CRM implementation page sets out the options: a review of an existing CRM, an implementation project, or a rescue.

If you would like a second pair of eyes on the requirements or the vendor script, a CRM review or platform selection is part of the CRM implementation service. If the wider question is whether the constraint is the CRM at all, the Strategic Marketing and Operations Review is the place to start. Either way, I’d be glad to compare notes. I work with your team, not around it: the requirements come from the people who will use the system, and the project ends with a named owner on your side. If you would rather check the wider set-up first, the Strategic Review is the usual first step.

FAQs

How do I choose a CRM for a small business?

Settle the process, the owners and the few decisions the CRM must support first. Then write a short must, should and could list, give every vendor the same script built from your real recent enquiries, and score what they show you. Choose on how well it fits your process, what it costs in year two, and how easily you can get your data out.

What should I fix before implementing a CRM?

Seven things: the real process from enquiry to invoice, an owner for each stage and for the system, the decisions the CRM must support, the minimum fields and their definitions, which data is worth moving, the basis for marketing email, and what it must connect to. None needs software to start.

What are the requirements for a CRM?

Write them as must, should and could. The musts are the few things without which the CRM cannot run your process, such as capturing the source of every enquiry, stages that match how you sell, an owner and a next-step date on every record, and a full data export. Anything you cannot tie to a decision belongs in could.

Should I clean my data before moving it to a new CRM?

Yes, and decide what not to move at all. Under UK GDPR, personal data should be adequate, relevant and limited to what is necessary, and the ICO advises periodically reviewing what you hold and deleting what you no longer need. A migration is a natural moment to do it.

Do I need a CRM at all?

Not always. With a handful of enquiries a month, a clear owner and a logged next step for each, a well-kept spreadsheet can do the job for a while. A CRM starts to earn its place when more than one person handles enquiries, when you cannot trace where an enquiry came from, or when follow-ups depend on someone remembering.

Let’s talk

A 30-minute discovery call is free and there’s no pitch. Bring the situation as it is. If I’m not the right fit, I’ll say so, and suggest who might be.

Book a discovery call (opens Cal.com) or send an enquiry