Knowing how to choose and implement a CRM for accountants matters more than knowing which CRM is best, because most of the value sits in the configuration and the habits around it rather than in the software itself. We have implemented client acquisition and onboarding systems across dozens of accounting firms, and the pattern is consistent: two firms buying the same platform get completely different results depending on how the first six weeks are handled.

There is a lot of published material on this topic and most of it says similar things. Centralise your data. Automate your workflows. Get a single view of the client. All true, and all close to useless as a purchasing decision, because none of it tells you whether your firm needs a CRM at all, which category to buy in, or what to do on the Monday after you sign.

This guide is structured around the decisions you actually have to make. What a CRM is for in a practice context. Whether your firm has enough enquiry volume to justify one. How to evaluate against your own process rather than a feature grid. The implementation order that produces adoption. The data protection obligations that come with holding prospect records. And the failure modes we see repeatedly when firms buy first and think later.

CRM stands for customer relationship management. In a practice context, the useful definition is narrower: a CRM is the system of record for people who are not yet clients, plus the commercial relationship with people who already are.

That covers a specific set of jobs.

• Capturing every enquiry from every channel into one queue, so nothing sits unanswered in a personal inbox.
• Holding the history of an enquiry: where it came from, what was discussed, what was quoted, why it did or did not convert.
• Running the follow-up sequence automatically, because most enquiries that go cold do so through silence rather than rejection.
• Producing and tracking proposals and engagement letters through to signature.
• Recording attribution, so you can see which acquisition channel produced which fee income.
• Flagging existing clients who have gone quiet or who fit a service they do not currently buy.

What it does not do

A CRM does not schedule jobs, track time, manage workflow capacity, or run your compliance calendar. If you are looking for something to tell you which VAT returns are due next week and who is behind on them, you are looking for practice management software, and buying a CRM will not help.

The distinction matters commercially. Practice management is graded on delivery: who is doing what for which client, by when, at what recovery rate. A CRM is graded on acquisition and retention: how many enquiries arrived, how many converted, how long it took, and what the pipeline looks like next month. Firms that conflate the two end up with an expensive system that answers neither question well.

Vendors will tell you every firm needs a CRM. That is not our experience. The honest test is volume and complexity, not firm size.

Signals you are ready

• You receive more than roughly ten inbound enquiries a month and cannot say from memory how many converted last quarter.
• More than one person handles enquiries, so context lives in separate inboxes.
• You are spending money on acquisition, whether that is Google Ads, SEO or outbound, and cannot attribute new clients back to a source.
• Prospects are going cold because follow-up depends on someone remembering.
• Proposals are produced ad hoc in Word and nobody knows which ones are outstanding.

Signals you are not

If your entire client base arrives through referral, you take three or four enquiries a month, and you personally handle all of them, a CRM will add administrative overhead without solving anything. A shared mailbox with labels and a single tracking sheet will hold the line. Buy the CRM when the volume arrives, or when you build the acquisition system that creates the volume. Buying it first is the most common sequencing error we see.

The other legitimate reason to buy early is that you are about to turn on demand generation. If you know that in eight weeks you will have thirty enquiries a month arriving from paid search and organic, standing the CRM up in advance is sensible. What you should not do is buy a CRM as a substitute for having a way to generate enquiries. Software does not create demand.

Practically, firms are choosing between three shapes of product, and the right answer depends on where the pressure is.

Practice management with client records built in

Suitable for firms whose main problem is serving the clients they already have, with low inbound volume. The client record sits where the work sits, so there is no duplicate data entry. The trade-off is that pipeline reporting, marketing sequences and attribution are usually weak or absent. Uku, whose buyer's guide argues most firms under around fifty staff belong in this category, is itself an example of the type. Treat that fifty-staff figure as their estimate rather than a rule.

A standalone CRM with marketing automation

Suitable when acquisition is the constraint: you are running paid campaigns, you need attribution, you need automated nurture over weeks or months, and someone in the firm owns conversion as a job. This category does pipeline, forecasting and campaign reporting properly. It does not know what a VAT return is, and it will not manage your job list.

A general-purpose CRM configured for practice use

Cheapest to licence, most expensive to configure. Everything specific to accounting, such as agent authorisation status, service scope, or year-end dates, has to be built as custom fields and maintained by hand. Workable if you have someone internally who will own the build. A liability if you do not.

The integration question

Whichever category you choose, the connection between the CRM and your accounting or practice software decides how long the data stays accurate. If contact and engagement details have to be keyed twice, they will diverge within a few months and staff will start trusting neither system. Test the integration during evaluation with real records, not the vendor's demo data.

Feature comparison grids reward the vendor with the longest feature list, which is rarely the right choice. A better method is to write down your current enquiry-to-client sequence first, then score each system against it.

Map the sequence

On one page, list every step from the moment an enquiry arrives to the moment the first piece of work starts. A typical practice sequence runs: enquiry received, acknowledged, qualified, call booked, call held, proposal sent, proposal signed, identity checks and anti-money-laundering completed, professional clearance requested, agent authorisation submitted, onboarding pack issued, first job created. Twelve steps, and most firms have never written them down.

Score against it

For each step, ask three questions of the system in front of you. Can it hold the record for this step. Can it trigger the next action without a human remembering. Can it report on how long this step takes on average. A system that scores well on the steps where you currently lose time is the right system, regardless of what else it does.

Demand a real trial

Run at least ten genuine enquiries through the trial before signing. Not test records. Real ones, handled by the people who will handle them daily. You are looking for friction: how many clicks to log a call, whether the mobile experience is usable, whether email sync works reliably with your mail provider.

Ask about exit

Ask how you export every record, including notes, attachments and email history, and in what format. Vendors that answer this crisply are usually confident about retention. Ones that deflect are telling you something.

A practice CRM holds personal data about people who never became clients, which sits outside your engagement letters and outside most firms' existing data protection thinking.

The ICO's guidance on data protection by design and by default is the relevant reference. Articles 25(1) and 25(2) of the UK GDPR require appropriate technical and organisational measures to be built into a system at design stage and maintained through its life, and require that personal data use is limited to what each purpose actually needs.

What that means at configuration

• Decide your lawful basis for prospect data before importing anything, and record the decision.
• Set a retention period for unconverted enquiries and configure automatic deletion. Indefinite retention of prospect records is difficult to justify.
• Restrict field visibility by role. Not every user needs to see every note attached to every contact.
• Keep marketing consent as a discrete, auditable field rather than an assumption inferred from an enquiry.
• Check where the vendor hosts data and confirm what their processor terms say about sub-processors.

The ICO has said it takes the measures a controller put in place into account when considering regulatory action. Configuring this properly at setup takes a couple of hours. Retrofitting it across four thousand contact records does not.

Treat the CRM as in scope for your existing records of processing activity, and update your privacy notice so the enquiry form on your website reflects what actually happens to the data.

Every failed CRM implementation we have reviewed failed for the same reason. Using the system was optional and manual, so within a quarter the record was incomplete, and once a record is incomplete nobody trusts it, and once nobody trusts it everyone reverts to their own notes.

Remove the manual step

The single highest-return configuration decision is making data capture automatic. Website enquiry forms should write directly into the CRM. Calls should be logged from the phone system. Email should sync both ways without anyone copying anything. If entering a lead requires opening a second application and typing, it will not happen consistently during January.

Give one person ownership

Someone has to own the pipeline, review it weekly, and chase records that are stale. In a firm of five that is the principal. In a firm of fifteen it might be a practice manager. Distributed ownership means no ownership.

Report on it in a meeting people already attend

Pull three numbers into your existing weekly or monthly meeting: enquiries received, proposals outstanding, conversion rate. Once the numbers are discussed in front of the team, the record gets maintained, because incomplete data becomes visible.

Start narrow

Turn on new enquiry capture and follow-up sequences first. Leave client retention scoring, advisory opportunity flags and campaign automation until the basic pipeline has been running cleanly for a month. Firms that switch everything on at once usually switch most of it off again.

A workable order for a firm of 2 to 20 staff, assuming the selection decision is already made. Expect six to eight weeks from signature to a system the team relies on.

Five patterns account for most of the CRM projects we are asked to rescue.