Most startup marketing audits are graded homework: a spreadsheet of channels, a red-yellow-green for each, a list of tactics to add. That format produces the same advice for every company, which is how you know it’s wrong. A real audit is a diagnosis — and diagnosis is a sequence, because each answer changes what you check next.
The sequence below is the one we run inside Kindling’s free audit, written out so you can run it yourself with a CRM export and an honest hour. We call it the Seven-Check Audit.
Why founders misdiagnose their own marketing
Founders usually arrive at an audit already holding a conclusion: we need more leads, or we should run ads. The actual recurring damage looks different: surviving on word of mouth with no repeatable engine, no tracking of spend or return, pipeline that can’t be traced to any marketing activity, and a go-to-market built by skipping phases — straight from first customers to “scale acquisition” with nothing in between.
The measurement gap alone is bigger than most people admit. In CaliberMind’s 2025 State of Marketing Attribution report, four in ten B2B marketers said they don’t track the pipeline marketing creates at all, and only about half could measure the opportunities it opens. 6sense’s 2025 survey of 634 B2B marketers found just 13% report closed-won revenue to their boards. If teams with full-time marketers can’t draw the line from activity to revenue, a founder doing marketing on the side almost certainly can’t — which is exactly why the audit starts there.
The Seven-Check Audit
1. Marketing contribution split
What percent of leads, and of closed customers, did marketing drive? If marketing drives more than half your sales, suspect a sales problem. If it drives less than 10% of leads, you likely have a marketing problem. Two different diseases; founders routinely treat the wrong one.
2. Motion mix vs deal size
Inbound vs outbound share of leads. Inbound leads have done their research and arrive readier to buy — but they skew smaller. The real question: does your lead-source mix match the deal size your revenue goal requires? The bigger deals almost always live on the outbound side, and outbound is slow to build.
3. The funnel middle
Opportunity-to-won rate and cycle length. Average B2B new-business win rates ran 19% in 2025 across 655,000 opportunities analyzed by Ebsta and Pavilion; typical healthy ranges sit at 20–30%, with top performers above 30%. If you’re well under that, “not enough leads” isn’t your problem — leads aren’t closing.
4. Presence vitals
Site load speed, organic traffic share, and LinkedIn cadence — the company account and the founder measured separately, because they do different jobs. These come from a scan, not from opinion.
5. Brand vs market
Walk your competitors’ homepages, then apply the 10-second test to your own: can a stranger say what you are, who it’s for, and why it’s different? Nielsen Norman Group’s research on page abandonment found the first 10 seconds decide whether a visitor stays at all.
6. The ICP test
The number that keeps everyone honest: what percent of your current pipeline is inside your ICP? Heavy inbound reliance usually means the ICP isn’t actually being reached — verify, don’t assume.
7. Structural checks
ACV-motion coherence (an ACV under ~$5K can’t feed outbound math; over ~$25K rarely grows self-serve), channel concentration (one source above ~70% of leads is fragility, not focus), proof assets (is there a case study with a number in it anywhere?), and churn — the leak marketing can’t fix.
The word-of-mouth trap
The most common finding at seed stage deserves its own section. Word of mouth feels like product-market fit — people love the product enough to refer their friends, and the friends close fast. But referred customers are friendlies. The test of a repeatable engine is strangers, and word of mouth is not a channel you can turn up when the quarter needs it. An honest audit gives word of mouth its due — it’s real evidence people love the product — and then asks the uncomfortable question: what happens when the founder’s network runs out?
Placement, not verdicts
A good audit never ends with a grade. It ends with a placement — twice.
First, on product-market fit. First Round Capital’s framework describes four levels — nascent, developing, strong, extreme — and the useful question is never “do we have PMF” but which level are we at, because the right marketing at the wrong level is wasted. Inbound emerging and referrals compounding is a “strong” signal; a handful of happy friendlies is nascent, and that’s fine — it just prescribes different work.
Second, on go-to-market maturity. Most seed-stage companies with word-of-mouth traction have effectively skipped phases: they went to market without a market map, a customer-evidence ICP, or a measurement spine, and jumped straight to acquisition. The audit names the skipped phases explicitly, because the 90-day plan writes itself from the gaps. The gate we use for the measurement phase is one question: can you forecast next month’s qualified pipeline with reasonable confidence? If not, that phase isn’t done — and spending on acquisition before it’s done is how ad budgets disappear without a trace.
What to do with the results
An audit that ends in a document failed. It should end in three things: the one or two problems that explain most of the symptoms, a placement that says which phase of work comes next, and the inputs for a plan — the pipeline math worked backward from your revenue goal, and two or three plays that fit your motion and your hours. That plan is its own discipline, and we’ve written the 90-day version of it here.
Run the checklist yourself this week — it costs an afternoon. Or let Kindling run it for you in about ten minutes: the free audit reads your site the way a stranger would, walks your competitors’ homepages, runs this exact sequence, and names the biggest leaks plainly. Either way, someone should run it before the next dollar gets spent.