Answer
How do you tell if your marketing agency is working?
Count enquiries and what happened to them. Everything above that in the reporting is describing activity you cannot spend.
Count enquiries received and what became of them, from your own records. Impressions, rankings and engagement can all improve while enquiries do not. If the agency owns the only measurement, you cannot assess the agency.
The structural problem with assessing an agency is that the agency usually supplies the measurement. That is not necessarily bad faith — they have the tools and you asked them to report — but it means the metrics chosen are ones that can be moved by the work being done, and a great many of those move independently of whether the business gained anything.
Impressions can rise because a platform changed how it distributes. Rankings can improve on queries nobody with intent searches. Engagement can climb on content that attracts an audience that will never buy. Traffic can grow while enquiries stay flat, which is the single most common pattern in agency reporting and is presented as progress because the line points upward.
The only assessment that survives is downstream and it belongs to you. Count enquiries received, by channel, per month. Count how many became customers. Count what they were worth. Those three numbers are the business's own, they are unaffected by which tool measured what, and any agency doing useful work will show up in them within a reasonable period. If they do not, no amount of platform reporting compensates.
Which means the prerequisite is instrumentation rather than scrutiny. A business that cannot count its own enquiries by channel cannot evaluate anyone, and will be permanently dependent on whichever narrative is presented. Setting up that counting is a week of unglamorous work and it changes the relationship permanently, because the conversation stops being about interpretation.
The timing question deserves more honesty than it usually gets, in both directions. Paid channels should show enquiry movement quickly — within weeks — because the mechanism is direct. Organic and content work genuinely takes months, and demanding evidence at sixty days will cause the right work to be cancelled. The way to hold both is to agree at the start what should be visible by when, in enquiry terms, and then hold that agreement rather than renegotiating it whenever a report is disappointing.
There is a diagnosis that gets misattributed to agencies more than any other, and it is worth checking before concluding anything. If enquiries have risen and revenue has not, the failure may be in intake rather than in marketing — enquiries arriving and not being answered, answered too slowly, or not followed up. That is not the agency's fault and it is frequently blamed on them. Counting response times alongside enquiry volume distinguishes the two cases immediately and prevents replacing an agency that was working.
A report that shows everything improving except the number of people who contacted you is a report about a different business's problem.
Answer Production Engine, Context Theory
Related questions
Should we ask for a report that includes revenue?
Ask for a report that includes enquiries and supply the revenue yourself. Most agencies cannot see what closed and should not be given access to your accounts to find out. The productive arrangement is that they report what they can influence — reach, cost, enquiries generated — and you own the conversion side, with both looked at in the same meeting.
What if the agency says our enquiry handling is the problem?
Take it seriously and check it, because it is frequently true and it is also the obvious defence. The check is cheap: your own log of how many enquiries received a response and how quickly. If the handling is genuinely poor, fixing it will raise the return on the same spend, which is worth more than changing supplier. If it is not, you now have evidence rather than an argument.
METHOD
Every figure below carries its source and the date it was verified. Nothing on this page is asserted.
The numbers on this page.
| What | Value | Specific to |
|---|---|---|
| US SMB retainer, focused one-to-two-service engagement | $1,500–$4,000 | Category-wide |
| US SMB full-service retainer | $3,500–$8,000 | Category-wide |
| Firms that never responded to a web enquiry at all | 23% | Category-wide |
| SMB marketing spend as a share of gross revenue | 3–5% | Category-wide |
2026 agency pricing survey · per month · verified
Oldroyd, McElheran & Elkington, "The Short Life of Online Sales Leads", Harvard Business Review (March 2011) · 1.25M inbound leads across 2,241 US firms · verified
2026 SMB marketing budget survey · a $1M business ≈ $2,500–$4,200/mo · verified
What is specific to this page.
| Kind | Claim | Check it against |
|---|---|---|
| Procurement | When the agency supplies the measurement, the reported metrics are ones the work can move, and several of those — impressions, rankings on low-intent queries, engagement — move independently of whether the business gained anything. | The agency's report, checked for whether any reported metric requires a person to have contacted the business. |
| Workflow | A business that cannot count enquiries by channel from its own records cannot evaluate an agency at all, and remains dependent on whichever interpretation is presented to it. | Whether the business can produce a monthly enquiry count by channel without asking its agency. |
| Workflow | Paid channels should move enquiry counts within weeks because the mechanism is direct, while organic and content work takes months, so a single evaluation horizon will either cancel the right work or excuse the wrong work. | Enquiry counts segmented by channel over the same period, compared against each channel's committed horizon. |
| Workflow | Where enquiries have risen and revenue has not, the defect may be in intake rather than in marketing, and response-time data alongside enquiry volume distinguishes the two cases before a supplier is replaced. | The business's own count of enquiries receiving a response and the elapsed time to it, over the same months. |
Each row would be wrong on another industry's page. Where a sourced figure exists it is in the table above instead; these are the constraints that shape the work and do not happen to be numbers.
Start with the measurement.
Reading about a benchmark is not the same as knowing your own number. The audit produces yours, measured rather than estimated.
$497 · delivered in 5 business days · credited against month one