Knowing how to measure marketing ROI for an accounting firm is mostly a data plumbing exercise. The arithmetic is simple. The difficulty is that in most practices the link between a marketing pound spent and a client signed is broken somewhere between the enquiry form and the engagement letter, so the arithmetic gets applied to numbers that do not mean anything.

The usual symptom is a dashboard full of impressions, clicks, sessions and click through rates, none of which tells the partner whether the firm made money. AccountingWEB has made this point about paid campaigns in practices: agencies frequently optimise for clicks while the firm quietly signs nobody. The same failure applies to organic search, directories, networking groups and every other channel a practice uses.

The second difficulty is timing. A client acquired in March may bill for six years. A campaign that looks expensive in its first quarter can be the cheapest thing the firm has ever done by month eighteen. Any measurement approach that only looks at revenue booked in the same month as the spend will consistently mislead you.

What follows is the measurement framework we implement inside client firms: what to track, how to calculate the two core cost figures, how to compare channels that behave differently, what to review monthly, and what to deliberately ignore.

The standard calculation is straightforward. Take the gross fee income attributable to marketing, subtract the marketing cost, divide by the marketing cost, and express it as a percentage or a ratio.

ROI = ((fee income attributed to marketing minus marketing cost) divided by marketing cost) x 100

A firm that spent a given amount and won four times that amount in attributable fees has produced a 300 per cent return, or 4:1 gross. The formula is not the interesting part. Two definitional choices determine whether the output is useful.

Which fee income counts

Use recurring annual fee value for signed clients, not one off project revenue and not lifetime value dressed up as year one. A compliance client on a monthly retainer has a known annual value the moment they sign. If you want a lifetime figure, multiply annual fee by your actual average client tenure in years, which you can pull from your practice management system by looking at clients who left over the last three years. Most owner managed practices find tenure sits somewhere between four and seven years. Use your own number rather than a borrowed one.

Which costs count

Include everything: agency or contractor fees, ad spend, software subscriptions used for acquisition, content production, and a realistic estimate of internal time spent on marketing and enquiry handling. Firms that exclude internal time systematically overstate their returns, particularly on channels like networking and LinkedIn that consume partner hours and no cash.

Run the calculation on gross fees, then again on your delivery margin if you want the honest picture. A client worth a given fee with sixty per cent gross margin funds acquisition very differently to one at twenty five per cent.

Attribution in a professional services firm is never perfect, and chasing perfection wastes time. Aim for a system that is right most of the time and honest about the rest.

Capture source at the point of enquiry

Every enquiry route into the firm needs to record where it came from. That means a required source field on the website enquiry form, a UTM parameter carried into hidden form fields for paid and email traffic, call tracking numbers on high volume pages if you take enquiries by phone, and a manual source field in your CRM for enquiries that arrive by word of mouth or direct email.

Define enquiry stages once and stick to them

Four stages cover most practices:

• Enquiry: anyone who submitted a form, called, or emailed asking about services.
• Qualified enquiry: within your service scope, within your fee floor, and not a student, supplier or existing client.
• Proposal issued: you have quoted.
• Client signed: engagement letter returned and onboarding started.

Record dates, not just counts

Store the date of first enquiry alongside the date of signing. Without both you cannot calculate lag, and without lag you will judge slow channels on fast channel timescales. In most practices the median gap between first enquiry and signed engagement runs from a few days for urgent compliance work to several months for advisory or a mid year switch.

Keep one source of truth

The CRM holds the record. Google Analytics tells you what happened on the website. Ad platforms tell you what happened in the auction. When they disagree, and they will, the CRM wins because it is the only system that knows who signed.

Two figures do most of the work in a practice. Both need to be calculated per channel, per month.

Cost per qualified enquiry

Total channel cost divided by the number of qualified enquiries that channel produced. Use qualified rather than raw, otherwise a channel that floods you with sole traders below your fee floor will look excellent on paper while consuming your team's time.

Cost per signed client

Total channel cost divided by the number of clients signed from that channel. This is the number that connects to your P&L. Compare it against the annual fee value of the clients that channel produces.

A useful working ratio: if cost per signed client sits below one third of first year fee value, the channel is comfortably self funding within the year. Between one third and one full year of fee value, it works but only if your client tenure is solid. Above one year of fee value, it needs either a fix or a decision to treat it as a long term investment with your eyes open.

Track the conversion rates between stages

Cost per client is a product of three things: cost per click or contact, enquiry to qualified rate, and qualified to signed rate. When cost per client rises, look at all three before assuming the channel is broken. A drop in the qualified to signed rate is usually an internal issue: slow response, a proposal that arrived a week late, or a partner too busy in filing season to run the call.

Segment by service line

A channel producing self assessment enquiries and a channel producing limited company retainers should never be compared on cost per client alone. Split the figures by the service the client signed for.

The trap in channel comparison is treating a compounding asset and a metered auction as though they respond on the same timescale.

Paid search and paid social

Spend stops, enquiries stop. Measurement can be near real time, and a rolling four week window is usually fair. Cost per enquiry is visible almost immediately, so this is the channel where measurement discipline pays back fastest and where waste shows up quickest.

Organic search and content

Investment in month one produces enquiries in months three to nine. Judging SEO on a monthly cost per client figure in its first quarter will always produce a bad answer. Measure it on a rolling twelve week window, and track leading indicators in the meantime: pages ranking in the top ten for commercial terms, impressions on those terms, and enquiries per hundred visits to service pages. One of the firms we work with moved from five to seven enquiries a month to fifteen to sixteen over roughly ten weeks, and the ROI figure only became meaningful once that curve had settled.

Referral and network

Cash cost is near zero, which makes ROI look infinite until you cost partner time honestly. Price partner hours at your charge out rate and referral marketing stops being free. It usually still performs well, and it almost never scales on demand, which is the reason firms build other channels alongside it.

Outbound and registry data

Cost per meeting is the more useful measure here, because the enquiry stage does not exist in the same form. Track meetings booked, meetings held, and proposals issued, then work back to cost per client from there.

Comparing these four on a single monthly table without adjusting for lag is the most common analytical error we see in practices.

A monthly review that takes twenty minutes and gets done beats a quarterly analysis that gets postponed until it is too late to act on.

The monthly table

One row per channel, with these columns:

• Total cost, cash plus internal time
• Qualified enquiries
• Cost per qualified enquiry
• Clients signed
• Cost per signed client
• Annual fee value signed
• Qualified to signed conversion rate

Add a twelve week rolling column

Single months in a small practice are noisy. Two extra clients in a month with eight enquiries swings every ratio. The rolling twelve week figure is the one to make decisions on. Use the monthly column to spot problems early, and the rolling column to decide budgets.

What to ignore

Impressions, reach, followers, click through rate, bounce rate, time on page, and total website sessions. None of them appear in the calculation, and none should move a budget decision on their own. They have diagnostic value when a cost per client figure moves and you are looking for the cause, which is a different job from reporting.

Two questions per channel

Every month, for each channel: is cost per signed client inside the threshold you set against fee value, and is the qualified to signed rate holding. If both are yes, leave it alone. If cost per client has risen but conversion is steady, the market or the campaign has changed. If conversion has fallen, the problem is inside the firm and no amount of extra spend will fix it.

Owners usually want a number to compare against. Useful benchmarks exist, and they should be treated as orientation rather than targets.

ICAEW publishes a benchmarking guide that includes marketing spend as a percentage of sales, which is the most relevant published UK reference point for a practice. Figures circulating from US marketing firms suggest business services businesses commonly allocate around ten to twelve per cent of revenue to marketing, and that newer firms might treat 2:1 or 3:1 as acceptable early while established firms aim higher. Those figures come from agency experience rather than official statistics, so label them as estimates when you use them internally.

Build your own baseline instead

Three months of clean data from your own firm is worth more than any published average, because it reflects your fee levels, your niche, your close rate and your local market. Record your first quarter of properly tracked figures and treat them as the baseline. Everything after that is measured against your own trajectory.

The qualitative side that the numbers miss

Some marketing effects will not attribute cleanly. A prospect who read three of your articles over a year, saw you speak at a local event, then searched your firm by name will be recorded as direct traffic. Buzzacott's marketing partner has argued in Accountancy Age that returns on investment in areas such as AI depend on judgement and relationship factors that resist measurement, and the same applies to brand building in a practice. The practical answer is to measure what attributes, be honest about what does not, and add a single question to your onboarding form: how did you first hear about us. Self reported attribution is imperfect and still catches things your tracking never will.

Six steps to move from anecdote to a measurement system you can run in twenty minutes a month. Most practices can complete steps one to four in a fortnight.

These five errors account for most of the misleading ROI figures we see when we take over measurement inside a practice.