Marketing is working if, and only if, it's producing enquiries from the right kind of buyer at a cost that makes sense against what a customer is worth. Five metrics tell you that: qualified enquiries per month, cost per opportunity, pipeline sourced by marketing, conversion rate from visitor to enquiry, and revenue by source. Everything else (impressions, followers, likes, raw traffic) is context at best and decoration at worst. If your monthly marketing report doesn't connect to your pipeline, you don't have a measurement problem; you have a report designed to look good.
Not leads. Not form-fills. Enquiries from people who match your ideal customer and have a problem you solve. This is the single most honest indicator of marketing health, and it should be counted by a human applying a written definition, not by whatever the analytics platform calls a "conversion". Track it monthly, by source. A business getting eight qualified enquiries a month knows more about its marketing than one drowning in dashboards.
Total marketing spend divided by the number of qualified opportunities created. This is the number that makes budget conversations rational: if an opportunity costs $800 to create, you close one in four, and a customer is worth $60,000, the argument for spending more makes itself. It also exposes channel truth. A channel producing cheap leads that never qualify is expensive; a channel producing few but excellent opportunities is cheap. Most businesses that calculate this properly for the first time immediately reallocate budget.
The dollar value of open opportunities that started with a marketing touch. This is the bridge metric between activity and revenue. It moves months before revenue does, which makes it the best early-warning system you have. If pipeline sourced by marketing is growing quarter on quarter, revenue will follow; if it's flat while activity metrics glow, the activity is theatre. This requires basic CRM discipline: every opportunity gets a source. That discipline is worth more than any analytics tool you can buy.
What percentage of commercial-page visitors take a meaningful step? Typical B2B sites convert 1–3%. This metric matters because it's a multiplier on everything else. Doubling it doubles the return on every dollar spent driving traffic, permanently. It's also the metric most likely to reveal that a "traffic problem" is actually a message or proof problem. Measure it on pages with buying intent, not blended across the blog, or the number means nothing.
Ultimately: which customers came from where, and what were they worth? Software attribution catches the trackable journey. A single question ("how did you hear about us?", asked on every enquiry form and in every first call) catches the untrackable one: referrals, LinkedIn lurking, podcasts, AI assistant recommendations. The two views disagree constantly, and the disagreement is informative. Businesses that collect self-reported attribution routinely discover their most valuable channel was one their analytics couldn't see.
Impressions and reach: measures of possibility, not outcome. An impression is someone scrolling past. Follower counts: an audience is only worth what it does. Raw traffic: meaningless without intent. Ten thousand visits from a viral tangent are worth less than a hundred from buyers. Likes and engagement rate: weak proxies, easily gamed, largely generated by peers rather than prospects. Email open rates: unreliable since privacy changes; clicks and replies are the real signal. These metrics seduce because they're large, they always go up if you spend, and they make bad months look survivable. They're fine as diagnostics inside a channel. They are not evidence that marketing is working.
One page. Qualified enquiries (by source, versus last month and last year). Cost per opportunity. Marketing-sourced pipeline. Conversion rate on commercial pages. New revenue by source. Then three sentences a dashboard can't write: what worked, what didn't, and what changes next month. If your current agency or team sends fifteen pages of charts and none of these five numbers, ask for this page instead. Anyone doing real work will be glad you asked; the reaction to the request is itself a diagnostic.
We don't have a CRM. Can we still measure this? A spreadsheet with columns for enquiry, date, source, outcome and value covers all five metrics. The discipline matters more than the software.
How long before these metrics should move? Conversion fixes: weeks. Paid channels: within a quarter. Content and SEO: two to four quarters. Judge each channel on its own clock.
What's a good cost per qualified opportunity? It's entirely relative to customer value. Sensible numbers for a $5,000 customer and a $500,000 customer differ a hundredfold. The test is your unit economics, not a benchmark.
Which single metric matters most for an SMB? Qualified enquiries per month. It's simple enough to count by hand and honest enough to build every decision on.