Choosing a CRM feels like it should be simple — pick the popular one, sign up, done. Then you're six months in, half your team has quietly gone back to spreadsheets, and you're paying for forty seats of software nobody opens. The CRM market is enormous, the sales pitches all sound identical, and the wrong choice costs far more than the subscription: it costs you the customer data you never captured. These ten steps will help you narrow the field properly, so the system you choose is the one your team will actually use.
Step 1: Define the problem before you look at software
Write down, in plain sentences, what's broken today. "Leads get lost between the website and the sales team." "Nobody knows which quotes are outstanding." "We can't tell which marketing actually brings in business." If you can't articulate the problem, no software will fix it — you'll just digitise the mess. This list becomes your scorecard: any CRM that doesn't solve these specific problems is out, however shiny its demo.
Step 2: Map your actual sales process
Before evaluating anything, sketch how a lead becomes a customer in your business: the stages, who does what, where deals stall. Most CRMs are built around a pipeline model, but the stages, terminology and handoffs differ by business. A system that forces your five-stage consultative sale into a generic template will be fought by your team from day one. If you're comparing build-vs-buy, our guide to choosing a CRM as a small UK business covers the trade-offs.
Step 3: List your must-haves — and be brutal
Separate genuine requirements from nice-to-haves. Must-haves are things the business cannot function without: email integration, quote generation, specific reporting, GDPR-compliant data handling, integration with your accounts software. Nice-to-haves are everything else. Most businesses discover their real must-have list fits on half a page — which immediately eliminates most of the market and makes the decision dramatically easier.
Step 4: Count the true users, not the org chart
CRM pricing is usually per user per month, so an honest headcount matters. But "users" isn't just headcount — it's who needs full access versus who just needs to log a call or check a record occasionally. Many systems offer cheaper limited-access tiers. Over-licensing is one of the most common wastes we see: forty full seats bought, twelve people logging in. Start with the core team and expand; it's far easier than paying for shelfware.
Step 5: Check the integrations that matter
A CRM that doesn't talk to your other systems becomes another silo. List the non-negotiables: your email (Microsoft 365 or Google Workspace), your accounts package, your website forms, your phone system, your marketing tools. Check whether integrations are native, via a connector, or require custom development — there's a big cost and reliability difference between the three. If you run on Microsoft tools, read our notes on CRM and Microsoft 365 integration planning before you commit.
Step 6: Trial it with real data and real sceptics
Never buy a CRM off a sales demo — demos show the happy path with perfect data. Run a proper trial: import a slice of your real (messy) data, and give it to two groups — your keenest early adopter and your most sceptical salesperson. If the sceptic finds it easier than their current workaround, you've got a winner. If the trial only works with clean sample data, walk away. Two to four weeks of genuine use beats any number of feature checklists.
Step 7: Test reporting on day one, not month six
The reports you'll need — pipeline value by stage, conversion rates, activity levels, forecast accuracy — should be buildable in the trial without a consultant. Ask each vendor to build your three most important reports live during the evaluation. If it takes custom development to answer "how much have we quoted this month?", that's a red flag about both the product and the ongoing cost of ownership.
Step 8: Price the total cost, not the headline
The per-user monthly fee is the beginning, not the end. Add implementation and data migration, training, integrations (custom ones can dwarf the licence cost), ongoing admin, and the price rises at renewal. Build a three-year total-cost model — our CRM total cost of ownership worksheet gives you the framework. A "cheap" CRM that needs £15,000 of consultancy to implement isn't cheap.
Step 9: Check data, security and GDPR fit
Where is your customer data stored? Who can access it? How do you export everything if you leave? For UK businesses, GDPR compliance isn't optional: you need data processing agreements, sensible retention controls, and the ability to handle subject access and deletion requests. Ask vendors directly and get answers in writing — "we're GDPR compliant" on a marketing page isn't due diligence. See also our guide on how your CRM helps with GDPR.
Step 10: Plan for adoption before you sign
The best CRM is the one your team uses, and adoption is a management job, not a software feature. Before signing, decide: who owns the system internally, what the non-negotiable usage rules are (every call logged, every quote in the system), how you'll train people, and what happens to the old spreadsheets (they must die, or the CRM will). Vendors sell software; you have to sell the change. Budget time for it like you'd budget money — because it costs just as much when it's missing.
Bringing it together: the shortlist method
Run the ten steps in order and you'll typically go from "the whole market" to a long list of six or eight, to a trial shortlist of two or three, to one decision you can defend. Score each finalist against your step-1 problem list, not against each other — the winner is the system that solves your problems with the least friction and the lowest three-year cost, not the one with the longest feature list.
If you'd rather have someone run this process with you — requirements, shortlisting, trials and migration — our CRM development team does exactly that, including custom builds when off-the-shelf genuinely doesn't fit. Get a free consultation and we'll give you an honest read on whether you need a new CRM at all.
FAQs
How long does it take to choose and implement a CRM?
Selection typically takes 4–8 weeks if you run proper trials; implementation takes another 4–12 weeks depending on data migration complexity and integrations. Anyone promising "live in a week" is skipping the steps that determine whether anyone uses it. Rushing selection is the most expensive shortcut — a wrong choice costs you the implementation fee plus a second migration.
Should a small business buy off-the-shelf or build a custom CRM?
Most small businesses should buy off-the-shelf — the economics of maintaining custom software rarely work under a certain scale. Build custom only when your process is genuinely unusual and gives you competitive advantage, or when you've outgrown two or three packaged systems. A good middle ground is a packaged CRM with custom integrations and workflows around it.
What's the biggest mistake businesses make when choosing a CRM?
Letting the loudest salesperson in the company choose based on a demo, without involving the people who'll enter the data daily. The second biggest is buying for the business you hope to become in five years instead of the business you are now — you'll pay for complexity you don't need while the basics go unused.
How many users do we really need licences for?
Fewer than you think. Licence the people who live in the system daily (sales, account management, support) at full access, and use cheaper limited or read-only tiers for everyone else. Audit usage quarterly — dormant full licences are one of the easiest savings to find, and most businesses find 20–30% of seats barely get touched.
Can we switch CRMs later if we get it wrong?
Yes, but it's painful and expensive — data migration, retraining, reintegrating everything, and months of reduced productivity. That's precisely why the ten steps matter: switching is always an option, but it's an option that costs roughly as much as getting it right the first time. Choose as if switching isn't an option.
