Answer in brief
Most marketing reports contain numbers nobody acts on. Four figures change a decision — how many enquiries, what each cost, how many became customers, and what a customer is worth — and everything else is context for explaining them.
Most reports contain numbers nobody acts on
A marketing report can be twenty pages long and still not answer whether to spend more next month. That happens when it reports activity — posts published, impressions served, positions moved — rather than the small number of figures that change a decision.
There are four such figures for most small businesses: how many enquiries arrived, what each one cost, what share of them became customers, and what a customer is worth. Together they answer spend more, spend differently, or stop.
Everything else has a job, but the job is explanation. Traffic explains a change in enquiries. Impressions explain a change in traffic. Position explains a change in impressions. None of them is the result.
This piece takes the four in turn, then the three ways a number can mislead you, and then what an honest reporting rhythm looks like when someone else is producing it.
Number one: enquiries, defined in one sentence
Before counting anything, write down what an enquiry is, in one sentence, and make it something a person could verify. A submitted form, a phone call longer than thirty seconds, a booking, a reply to a message.
The definition matters more than the tooling. Two people counting different things will disagree for months without noticing, and the disagreement usually surfaces in the meeting where a decision has to be made.
Count it in one place. An enquiry recorded in three systems is an enquiry counted three times or none, and reconciling them at the end of a quarter is a job nobody plans for.
Then hold the definition still. Changing what counts halfway through a period is the fastest way to produce a chart that shows something that did not happen.
Number two: what each enquiry cost
Divide the spend in a period by the enquiries in the same period, per channel. That is the cost per enquiry, and it is the number that makes two channels comparable when nothing else does.
Decide whether you are including only media or media plus work, and then keep the choice consistent. Both are defensible; changing which one you mean between reports is not, because the comparison becomes meaningless.
The figure means little in isolation and a great deal against your own previous period. There is no correct cost per enquiry for an industry, and any supplier quoting one as a benchmark is describing somebody else's business.
Watch what happens to it as budget rises. A cost per enquiry that climbs steadily as you spend more is telling you the audience is being exhausted, which is a scaling limit rather than a failure.
Number three: the share that become customers
Enquiries are not customers, and a channel that produces many cheap enquiries that never convert is more expensive than one producing fewer expensive ones. Only the conversion share tells them apart.
This is also where marketing and sales stop being separable. If the share falls while the enquiry count rises, the change may be in what the marketing promised rather than in how the sales conversation went.
Track it per channel where you can. The same offer often converts at different rates depending on whether the person was already looking or was interrupted, and averaging the two hides a decision.
Keep the definition of a customer as strict as the definition of an enquiry. A signed order and a verbal agreement are different events, and mixing them makes a good month look better than it was.
Number four: what a customer is worth
Without this figure, the other three cannot be judged. A cost per enquiry of forty is excellent for one business and ruinous for another, and only the value of a customer settles which.
Use what you can actually observe rather than a projection. An average first order is a real number; a lifetime value estimated over five years for a business that is two years old is a hypothesis wearing a number's clothes.
If repeat purchase matters in your business, measure it over a period you have actually lived through, and say which period you used. A figure without its window is not comparable to anything, including its own past self.
This is the number that turns a marketing conversation into a business conversation, and it is the one most often absent from an agency report, because it lives in your accounts rather than in a platform.
The numbers that are context, not results
Traffic, impressions, followers, positions, engagement: each of these can rise while enquiries stay flat, and each of them is genuinely useful for explaining why a result moved.
The mistake is promoting them to results. A report that leads with traffic growth is answering a question nobody asked, and it tends to appear in months when the four real numbers did not move.
Position in search is a particularly slippery one. It varies by person, place and device, so a single reported position is an average of things that were never the same measurement.
Use context numbers to answer why. If enquiries fell and traffic held, look at the page and the offer. If traffic fell and the conversion share held, look at the channel. That is what context is for.
Why a metric without a baseline cannot be judged
The most common reporting failure is not a wrong number, it is a number with nothing to compare it against. Forty enquiries is neither good nor bad until you know what a normal month looks like.
Record a normal period before you change anything, even roughly. A month of counts written in a spreadsheet is enough, and it is the difference between a later conversation about evidence and a later conversation about impressions.
Where seasonality exists, note it. A quiet August compared against a busy May produces a conclusion about the calendar dressed up as a conclusion about marketing.
And write down what changed and when. A dated list of changes is often the fastest available explanation when a number moves and nobody can remember why.
The three ways attribution misleads
First, last click. The channel a person used immediately before enquiring gets the credit, which systematically flatters channels that capture demand and starves the ones that create it.
Second, invisibility. Analytics counts what happens on the site. Calls made from a phone screen, enquiries through a message, and visits where scripts were blocked produce gaps that no configuration closes entirely.
Third, over-precision. A report giving a channel a share to one decimal place implies a measurement accuracy that the underlying data does not have, and decisions made on that precision are decisions made on noise.
The cheap correction for all three is a question at the point of enquiry: ask how the person found you, record the answer, and keep it next to the analytics rather than instead of it.
Making the site's own numbers trustworthy
Check that the events you rely on actually fire, and check them again after any change to the site. Form submissions and click-to-call are usually implemented as page-specific code, and page-specific code is what a redesign rewrites.
Test the form yourself on a phone and confirm the submission arrives everywhere it should — the inbox, the CRM, the notification channel. A form that accepts and delivers nowhere produces a silent zero that looks like a marketing problem.
Check the form is usable by keyboard as well. The W3C's WCAG 2.2 quick reference is the practical list, and a field that cannot be reached without a mouse is a lost enquiry that never appears in any report.
Do this before drawing conclusions from a drop. A measurable share of sudden declines are measurement failures rather than demand failures, and the two are indistinguishable in a chart.
What an honest report contains
The four numbers, against the previous period. What changed during the period, dated. One interpretation per number. And a decision, or an explicit statement that no decision is needed this month.
A monthly figure with no interpretation is a receipt. A monthly figure with a decision attached is the thing you are actually paying someone to produce.
The Monthly Growth Retainer is $330 per month on a monthly cycle with 30 days notice to stop, covering monthly planning, a reporting rhythm and channel optimisation. Channels and volume are agreed each cycle; ad spend and third-party tool subscriptions are excluded. The article applies that principle to a defined result: A metric with no baseline cannot be judged, so record a normal period before changing anything.
Ask what a report contains before the first one arrives. It is a reasonable question, the answer is short, and it prevents a year of documents that describe activity.
Measuring a diagnosis rather than a campaign
Not every purchase produces a number. A diagnosis produces a decision, and the way to judge it is whether the decision was made, not whether a chart moved that week.
Search Snapshot is priced at $70 over 1-2 working days and covers a visibility check on your top queries, three competitors side by side and one page of findings with no deck, carrying one clarification round. Implementation, content and ad spend are excluded. For this case, the decision criterion is specific: Most marketing reports contain numbers nobody acts on. Four figures change a decision — how many enquiries, what each cost,…
Judge it by what it settles. If the argument about which channel to fund is shorter afterwards, it did its job, and that outcome will not appear in any analytics platform.
The same applies to positioning work. A rewrite changes the response of every channel at once, which makes it valuable and makes it hard to attribute — another reason to record a baseline before changing the words.
A reporting routine that takes twenty minutes
Once a month, write down the four numbers and the previous month's four next to them. Add a dated list of what changed. Write one sentence per number saying what you think it means.
Then write the decision: spend more here, stop this, change that page, or nothing this month. "Nothing this month" is a legitimate and underused outcome.
Keep the file. A year of these is the most useful marketing asset a small business can own, and it costs twenty minutes a month to build.
It also changes the conversation with any supplier. A client who arrives with four numbers and a history gets a different quality of answer than one who arrives asking whether things are going well.
Practical checklist
- Write down the single action you count as an enquiry, in one sentence.
- Record a normal month before changing anything, so you have a baseline.
- Divide spend by enquiries to get a cost per enquiry, per channel.
- Track what share of enquiries become customers, not just how many arrive.
- Ask every enquiry how they found you and keep the answers next to the analytics.
- Read the numbers on a fixed rhythm rather than daily.
Questions and answers
What are the most useful marketing metrics for a small business?
Four: how many enquiries arrived, what each one cost, what share became customers, and what a customer is worth. Together they answer whether to spend more, spend differently or stop. Every other figure in a report exists to explain one of those four.
Is traffic a useful metric?
It is context rather than a result. Traffic explains a change in enquiries — more people arrived, or the same people converted better — but on its own it cannot tell you whether to spend more. A rise in traffic with flat enquiries is usually a message or page problem rather than good news.
What does reporting cost as part of a marketing arrangement?
The Monthly Growth Retainer is $330 per month on a monthly cycle with 30 days notice to stop, and includes monthly planning, a reporting rhythm and channel optimisation, with channels and volume agreed each cycle. Ad spend and third-party tool subscriptions are excluded.
How do I calculate cost per enquiry?
Divide what you spent in a period by the number of enquiries in the same period, per channel. Include the media spend and, if you want the honest figure, the work you paid for as well. The number matters less in isolation than as a comparison against your own previous period.
Why do analytics and my own count disagree?
Because they measure different things. Analytics counts what happened on the site; your inbox counts what reached you. Blocked scripts, calls made from a phone, and enquiries that arrive through a message all produce gaps. Asking every enquiry how they found you is the cheapest way to close it.