Most guides on how to build a marketing system for CPA firms are really lists of channels. Run Google Ads, post on LinkedIn, start a blog, ask for reviews. Channels are the easy part. The reason a firm can spend for six months and see almost nothing is usually structural: enquiries arrive with nowhere defined to go, nobody owns the response, and the person answering has no way of telling a serious limited company enquiry from someone comparing four quotes on price alone.

A system is the set of connected components that takes a stranger with a problem and turns them into a client who is signed, onboarded, and in your practice software. Each component has an owner, a defined input, a defined output, and a number you can look at weekly. When one component underperforms, you can find it. That is the difference between marketing you can improve and marketing you can only hope about.

This guide sets out the five components, the order to build them in, what each one produces, and where firms typically lose money. It is written for firms already delivering good work who want the front end of the practice to behave like the rest of it: documented, measurable, and not dependent on the owner remembering to chase.

Positioning is the first component because everything downstream inherits it. Your ad copy, your service pages, your qualification questions, and your fee conversations all get easier or harder depending on how specific your answer is to the question of who you serve.

Most firms answer that question by instinct. They pick a sector they enjoy, or the one three existing clients happen to sit in. That works occasionally and fails often, because it ignores whether enough of those businesses exist within reach and whether they are actively looking for an accountant.

The two data sources that settle it

• Business registry data. Companies House filings tell you how many companies sit in a given SIC code, where they are, how recently they were incorporated, and how fast that population is growing. A niche with a shrinking company count is a niche you will fight over.
• Search demand data. Keyword volume tells you whether those businesses look for an accountant online in the language you would use. If nobody searches for a specialist in that sector, your acquisition has to come from outbound or paid social rather than search, which changes the build entirely.

What a usable position looks like

A workable position states who you serve, the specific problem you handle for them, and what proof supports the claim. Something like: we work with owner-managed construction firms in the East of England on CIS, subcontractor payroll, and year end, and we currently act for a defined number of them. That is testable. A prospect either fits or does not, which is exactly what you want, because the ones who do not fit are the ones who cost you time.

You are allowed more than one position. Run each as its own page, its own campaign, and its own set of proof. What you cannot do is run a single generic message and expect it to convert against a firm that has picked a lane.

Once the position is set, you need infrastructure that catches people already looking. Search is the highest intent source available to an accounting firm, because someone typing a service and a location into Google has usually decided to act. The task is architecture rather than volume of content.

Structure beats blogging

A search architecture for a firm has three layers. Service pages cover what you do. Niche pages cover who you do it for, one page per sector you have positioned around. Resource pages answer the questions your target sector asks before they are ready to enquire, which is how you get in front of people early and stay in their consideration set.

Each page has one job and one primary query. A single page attempting to rank for company accounts, VAT, payroll, and tax planning ranks for none of them well. Firms that build twenty tightly focused pages consistently outperform firms that publish sixty pieces of general commentary.

The conversion layer most sites skip

Ranking produces visitors. Visitors produce nothing without a conversion path. Every page needs a route to enquiry that suits where the reader is: a form for the ready, a call booking link for the confident, a downloadable resource for the researching. One of our clients rebuilt their site around conversion rather than content and saw monthly visitors go up four times over, with ten to fifteen enquiries in the first month and three new clients from that first cohort.

Local search sits inside this layer. A complete Google Business Profile, accurate details across directories, and a steady flow of reviews affect whether you appear in map results for local searches, which for most firms carry the strongest intent of anything you can chase.

Search takes time to compound. Paid acquisition and outbound give you a channel you can switch on while the search architecture matures, and both feed the same qualification and follow-up components you are building anyway.

Paid search

Google Ads works for accounting firms when the keyword list stays narrow. Bid on service plus location and service plus sector. Avoid broad terms such as accountant or tax help, where you pay for clicks from students, existing accountants, and people wanting free advice. Send every ad to a page that matches the search rather than to your homepage, and track which keyword produced each booked meeting rather than which produced clicks.

Paid social

Meta reaches people who are not searching yet. That makes it better suited to a specific offer aimed at a defined audience than to general brand presence. Sector targeting, a clear problem statement, and a low friction next step do the work. Expect a longer path from first click to signed client than you see from search.

Outbound built on registry data

Registry data lets you build a list of companies that match your position precisely: sector, size, location, incorporation date, filing history. That is a far better starting point than a bought list. One client of ours runs a registry-data outbound engine producing thirty to fifty cold leads a month, which converts into five to ten booked meetings.

Outbound carries obligations. Under UK direct marketing rules, people have an absolute right to object to marketing, you must tell them about that right no later than your first contact, and you must make objecting easy and free. When someone objects, suppress their details on a do-not-contact list rather than deleting them, because deletion means you may contact them again by accident. Build the suppression list before the first send, not after the first complaint.

This is the component firms skip, and it is the one that decides whether growth feels good or exhausting. If every enquiry goes straight into your diary, you spend your week on calls with businesses you cannot serve profitably. Firms that describe marketing as attracting nothing but price shoppers usually have no qualification layer at all.

Qualify at the form, not in the meeting

Your enquiry form should ask enough to sort. Turnover band, entity type, current bookkeeping arrangement, what triggered the search, and when they need to move. Five fields, not fifteen. That data routes the enquiry: strong fits get a booking link immediately, borderline fits get a short qualifying call, poor fits get a courteous response and useful signposting.

One firm we work with runs an automated qualification layer on the enquiry form so that the calendar only carries opportunities worth the partner's time. The volume of enquiries did not change. The composition of the diary did.

Publish your criteria

Stating who you work with on your site is itself a qualification tool. If your pages say you act for VAT registered limited companies with a bookkeeper in place, sole traders wanting a one-off return will mostly self select out before they ever fill in a form. That is a saving, not a loss.

Price positioning without a price war

Every accountant meets the prospect who has been quoted half elsewhere. The answer is not discounting. It is describing scope precisely enough that comparison becomes possible: what is included, what response times look like, what proactive contact happens through the year. When scope is vague, price is the only variable left for a prospect to compare, and someone will always be cheaper.

More revenue is lost between enquiry and signature than anywhere else in the system. A prospect submits a form on Tuesday afternoon, hears nothing until Thursday, and by then has spoken to two other firms. Nothing about the marketing was wrong. The handover was.

Response time is a system setting

Decide your target response time and build to it rather than relying on someone noticing an email. That means an automated acknowledgement within minutes, a task created for a named person, and an escalation if the task is untouched after a set period. Where the enquiry qualifies well, an instant booking link removes the back and forth over availability entirely.

Sequences for the people not ready yet

Most enquiries are not ready to sign this week. Some are between year ends, some are waiting to leave their current accountant at a sensible point, some are still deciding. Without a sequence, those contacts simply expire. A structured follow-up over several weeks, mixing email and a scheduled call attempt, recovers a meaningful share of them. Keep the content useful and the unsubscribe visible, since the same objection rules apply to your CRM as to your cold outreach.

Onboarding is part of marketing

The system does not end at the signature. Engagement letter, anti money laundering checks, authorisation with HMRC, software access, and the first data request all need to run as a defined sequence rather than a series of emails written from scratch. This is where growth either scales or creates administrative drag. A firm adding ten clients a quarter through manual onboarding will eventually stop marketing to protect its own capacity, which is the quietest way a growth programme dies.

You need few numbers, reviewed regularly, at the points where the system can break. Weekly is enough for the top of the funnel, monthly for the rest.

The numbers worth holding

• Enquiries by source. Split organic search, paid search, paid social, outbound, and referral. Without the split you cannot cut anything with confidence.
• Qualified rate. The proportion of enquiries that meet your criteria. Falling qualified rate with rising volume means your targeting has drifted.
• Enquiry to meeting. This measures your follow-up, not your marketing. A weak number here is nearly always a response time or ownership problem.
• Meeting to client. This measures the meeting itself, your scope description, and your fee positioning.
• Time to onboarded. From signature to first data received. This is the early warning for administrative overload.

Judging performance over the right window

Search and content compound. Paid and outbound respond quickly. Judging a search programme at six weeks tells you nothing useful, and pausing paid after ten days tells you less. One firm we work with moved from five to seven enquiries a month to fifteen to sixteen, and nine new clients, over roughly two and a half months of search architecture work. That is a normal shape for search: quiet, then a step change once the pages hold position.

Attribution will never be perfect

Someone finds you on Google, checks Trustpilot, sees an ad a fortnight later, then asks a friend about you. Your CRM will record one of those. Ask on the enquiry form how they first heard of you and accept that the answer is directional. Directional data reviewed consistently beats precise data reviewed never.

Build these in sequence. Each stage depends on the one before it, and skipping ahead is the usual reason a firm ends up with traffic it cannot convert.

These five failure points account for most of the wasted spend we see when we audit an existing setup.