Skip to main content

← insights

how to run a b2b marketing audit (template and checklist)

Mishaal Murawala·

A marketing audit run from the operating seat: revenue math first, then measurement, channels, content, team, and a 30-day fix list.

A marketing audit is a structured review of marketing spend, output, and its numbers. Most published guides start with brand and channels. I start with revenue math. A channel review built on broken measurement is a review of fiction.

I have spent 15 years as a VP of Marketing inside PE-backed B2B SaaS companies. The audits that changed anything were never the ones that produced a 40-slide deck. They found one broken definition, fixed it, and made the forecast trustworthy again. This is the order I use, the evidence I pull, and the checklist I hand over.

What a marketing audit is (and is not)

An audit answers three questions in order. Are the revenue numbers real. Is the measurement that produces them intact. Is the spend behind them efficient.

It is not a brand critique. It is not a competitor teardown. It is not a list of channels you could try. Those are strategy exercises, and they are worthless until you trust the inputs.

Most published guides treat the audit as a sweep across the marketing mix. Pipedrive and SmartBug both frame it that way: inventory the channels, score each one, build a plan. That framing is fine for a company whose CRM is clean. In PE-backed B2B the CRM is rarely clean. The sweep then draws confident conclusions from bad data.

An audit that starts with channels ends with opinions. An audit that starts with revenue math ends with a fix.

The other failure mode is scope. An audit that inspects everything finds nothing worth acting on. I cap the whole exercise at four weeks and force a ranked fix list at the end. If the audit does not name the single biggest constraint, it did not work.

Before you start: the three numbers to pull

Do not open a dashboard until you have these three. Pull them from the systems of record, not a slide.

1. Pipeline coverage against the current-quarter number. Open pipeline divided by the quota or plan for the quarter. Pull it from the CRM opportunity table, filtered to the quarter's close dates. Note the stage mix, not just the total.

2. CAC payback in months. Fully loaded sales and marketing spend divided by new logos. Then divide by monthly gross profit per new customer. If finance and marketing quote different numbers, that gap is your first finding. The full method is in CAC payback.

3. Forecast accuracy over the last four quarters. Forecast at the start of each quarter against actual closed won. Four data points. If the average miss is over 20% either way, the pipeline math is the audit. Everything downstream is decoration.

These three take a day to assemble and they set the entire agenda. A company with 3x coverage, 14-month payback, and 5% forecast error has a channel problem. A company at 1.4x coverage with 30% swings has a pipeline problem, not a marketing one.

Step 1: Revenue math

Start where the money is, because every other finding gets ranked against it.

Pipeline coverage by segment and source. Blended coverage hides the failure. Split it by ICP segment and by lead source. A healthy 3x blend often hides a 5x self-serve segment carrying a 1.2x enterprise segment.

Stage conversion rates, stage by stage. Pull the last four quarters of opportunity history. Compute the conversion rate at each stage transition. Look for the stage where the rate collapsed, not the stage with the lowest rate. A permanently low rate at first meeting is a design choice. A rate that halved last quarter is an incident.

Deal velocity and aging. Median days in stage, and the count of opportunities past two times the median. Aged pipeline inflates coverage without ever closing. In most audits I run, 15% to 25% of open pipeline is dead and nobody has closed it out.

CAC payback by channel. Not blended. If one channel runs 9 months and another 34, the blended 18 tells you nothing useful.

Red flag to look for first: a coverage ratio that never changes month over month. That is a sign the pipeline is being managed to a target rather than measured.

Step 2: Measurement integrity

This is the step every published guide skips and the step that finds the most defects.

Conversion definitions. Write down what counts as a lead, an MQL, an SQL, and an opportunity. Then check whether those written definitions match what the systems actually do. They usually do not. Definition drift is why last year's numbers cannot be compared with this year's.

Form-to-CRM plumbing. Take a real form on the site. Submit it. Follow the record all the way through to the CRM. Time how long it takes to arrive. Check which fields populated. Check whether source attribution survived. I have found forms that wrote to an automation list and never created a CRM record.

Analytics event hygiene. In GA4, key events are the conversion measure. The platform counts every occurrence of one by default. In real audits I keep finding one form submission firing more than one event. Sometimes a thank-you page view is a key event alongside the submit event. The analytics lead count then runs 1.5x to 2x the real CRM record count. The board hears the analytics number. Sales hears the CRM number. Nobody notices they describe the same thing.

Attribution model and its limits. Note which model is in use and which touches it can actually see. Then check the dark spots: direct traffic with no prior touch, offline events, and AI traffic with no referrer. State the limits explicitly in the audit. An attribution number quoted without its limits is a claim, not a measurement.

Reconcile across systems. The same conversion should appear in the ad platform, the analytics tool, and the CRM. Where the counts diverge by more than 10%, that divergence is the finding. Do not average them. Chase the gap to its cause. More on the modeling side in marketing ROI measurement.

MQL as a governing metric. If MQL volume drives budget and headcount decisions, that is itself a finding. I explain why in why we do not measure MQLs.

Step 3: Channel efficiency

Now, and only now, look at channels. With trustworthy definitions, this step is fast.

Spend by channel against pipeline sourced by channel. One table. Twelve months. Cost per opportunity and cost per closed won by channel. The CRM is the denominator, never the ad platform.

Paid search waste. Search term reports for the last 90 days. Spend against terms with zero conversions. Brand versus non-brand split, reported separately, always. Blending brand into paid search is how an account looks fine while non-brand burns.

Conversion tracking on every paid platform. Confirm each platform's conversion action points at the same event the CRM counts. Mismatched conversion actions are why platform ROAS and CRM pipeline disagree.

Outbound efficiency. Meetings booked per rep per week. Then the meeting to qualified opportunity rate. If meetings are healthy and qualification is not, the targeting is wrong.

Channel concentration. What percentage of pipeline comes from the single largest source. Above 60% is a fragility finding, regardless of how well that channel performs today.

Step 4: Content, search and AI visibility

Content is audited as an asset base, not as a publishing calendar.

Traffic and conversion by page, not in aggregate. Rank landing pages by sessions, then by conversion rate. The interesting pages are high-traffic low-conversion, and low-traffic high-conversion. The first is a fix. The second is a promotion opportunity.

Decay. Pages that lost more than 30% of organic sessions year over year. In most B2B libraries a few pages carry most of the organic traffic. Their decay goes unnoticed while aggregate traffic looks flat.

Search coverage against the buying committee. Map the keywords you rank for against the questions each role in the committee asks. Gaps are usually at the top and the bottom, not the middle.

Technical fundamentals. Indexation, canonical tags, structured data on money pages, and Core Web Vitals. Fast to check, and a broken canonical can silently remove a page from search.

AI visibility, tested directly. Ask ChatGPT, Claude, Perplexity, and Google AI Overviews the questions your buyers ask. Record whether you appear, and who does. This is a real audit line now. A growing share of buyers never reach a blue link. Check that AI crawlers can reach your key pages. Check that the content answers a question in the first 40 words.

Step 5: Team, tooling and spend

Tool inventory against actual use. Every tool, its annual cost, its owner, and its last meaningful login. Every audit I have run found a four-figure subscription nobody opened in six months.

Overlap. Two tools doing the same job is common after a merger or a leadership change. Name both and pick one.

Data flow map. One diagram: where records enter, what transforms them, where they land. Most measurement defects are visible on this diagram before you test anything.

Headcount against the motion. Compare the team shape to the go-to-market motion. A product-led motion staffed for field marketing is a mismatch no campaign fixes.

Spend split. Percentage of budget on people, tools, media, and agencies. Track it against the pipeline each bucket produces.

Most of what breaks here is organizational, not tactical. The pattern behind it is in why GTM initiatives fail.

The marketing audit checklist

Twenty-five checks, in the order I run them. Each row names the evidence to pull, so this is executable rather than aspirational.

AreaQuestionEvidence to pullRed flag
RevenueIs pipeline coverage sufficient for the quarter?CRM open pipeline / quarter planBelow 3x, or unchanged month over month
RevenueDoes coverage hold by segment?Coverage split by ICP segmentOne segment below 1.5x hidden by blended total
RevenueWhere does stage conversion break?4 quarters of stage transition ratesA rate that halved in one quarter
RevenueHow much pipeline is actually dead?Opportunities past 2x median stage ageOver 20% of open pipeline aged out
RevenueWhat is CAC payback?S&M spend, new logos, gross marginFinance and marketing quote different numbers
RevenueDoes payback hold by channel?Payback computed per channelRange wider than 3x between channels
RevenueIs the forecast trustworthy?4 quarters forecast vs actualAverage miss above 20%
MeasurementAre conversion stages defined in writing?The written definitions documentNo document, or it contradicts the CRM
MeasurementDo the systems match the definitions?CRM stage entry criteria and automation rulesStages advanced manually with no criteria
MeasurementDoes a form submission reach the CRM?Live end-to-end test submissionRecord missing, delayed, or fields empty
MeasurementDoes source attribution survive the handoff?Source field on the test recordSource blank or overwritten to Direct
MeasurementDo GA4 key events overcount leads?Key event count vs CRM record countAnalytics count 1.5x or more above CRM
MeasurementIs more than one event firing per submission?GA4 DebugView on a live submissionSubmit event plus thank-you page both counted
MeasurementDo platform, analytics and CRM agree?Same conversion counted in all threeDivergence above 10%
MeasurementAre attribution limits stated?The attribution model and its blind spotsNumbers quoted with no stated limits
MeasurementDoes MQL volume drive budget?Board deck and budget rationaleSpend justified by MQL count alone
ChannelWhat does each channel cost per opportunity?Spend by channel / CRM opportunitiesAny channel with spend and zero opportunities
ChannelIs brand separated from non-brand in paid search?Search campaign structure and reportingBlended into one performance number
ChannelHow much paid search spend converts nothing?90-day search term reportOver 20% of spend on zero-conversion terms
ChannelDo platform conversion actions match CRM events?Conversion action config per platformPlatform counting a different event
ChannelHow concentrated is pipeline by source?Pipeline share of largest sourceAbove 60% from one source
ContentWhich pages lost traffic year over year?Organic sessions by page, YoYTop pages down over 30%
ContentWhich pages get traffic but do not convert?Sessions and conversion rate by pageHigh traffic, near-zero conversion
ContentDo AI assistants cite you for buyer questions?Direct prompts to the major assistantsCompetitors cited, you are absent
TeamIs every tool actually used?Tool list, cost, owner, last loginA paid tool unopened for six months

Marketing audit template

Copy this outline and fill it in as you go. Keep the whole thing to five pages.

1. Scope and dates. Period audited. Systems inspected. What was deliberately excluded.

2. The three numbers. Pipeline coverage. CAC payback. Forecast accuracy over four quarters. Each with its source system named.

3. Findings, ranked by revenue impact. One line each. Finding, evidence, estimated impact in dollars or months of payback. Ranked, not grouped by department.

4. The constraint. One sentence naming the single biggest blocker. If you cannot write this sentence, keep auditing.

5. Measurement defects. Every broken definition, plumbing gap, and double-counted event. Name the test that proved it.

6. Fix list. Each fix with an owner, a date, and the metric that proves it landed.

7. What I did not check. The honest boundary. An audit that claims total coverage is not credible.

What to do in the 30 days after

An audit that ends in a deck is a cost. An audit that ends in shipped fixes is an investment. Here is the sequence I use.

Week 1: fix measurement, nothing else. Every downstream decision depends on the numbers being real. Deduplicate the key events. Repair the form-to-CRM path. Write the conversion definitions down and get finance to sign off on the gross margin. This is unglamorous and it is the highest-leverage week of the month.

Week 2: close out dead pipeline and re-baseline coverage. Then recompute the three numbers. The deck built on week one's fixes will differ from last month's. That difference is the first deliverable.

Week 3: cut the obvious waste and reallocate. Zero-conversion search terms. Unused tools. Duplicate subscriptions. Move that budget to the shortest-payback channel, not the best-looking dashboard.

Week 4: ship one structural fix and instrument it. One. Pick the fix that hits the named constraint. Ship it. Attach a metric that moves inside a quarter. Then write down what you will check in 30 days.

That sequence is exactly the growth diagnostic I run as a 30-day engagement. I connect the evidence, find the constraint, rank the fixes, and ship the first one. I work through the marketing leader who already owns the mandate.

Sources

BOOK A WORKING SESSION

I run this audit as a 30-day engagement.

The evidence connects first. The constraint gets named. The first useful fix ships in week one. I lead it through the marketing leader who owns the mandate.

Numbers to pull before opening a dashboard
3
Checks in the audit checklist
25
Audit to first shipped structural fix
30 days

frequently asked questions.

What is a marketing audit?

A marketing audit is a structured review of marketing spend, output, and its numbers. A B2B marketing audit runs in a fixed order. Revenue math first. Then measurement integrity. Then channel efficiency. Then content and search. Then team and tooling. It ends in a ranked fix list, not a deck.

How long does a B2B marketing audit take?

Four weeks is the right cap for a full audit including fixes. The three opening numbers take about a day to assemble. Those are pipeline coverage, CAC payback, and four quarters of forecast accuracy. Measurement testing takes a week. Channel, content, and tooling review takes another week. The rest goes to shipping the first fix.

What should a marketing audit checklist cover?

Twenty-five checks across five areas. Revenue math covers coverage, stage conversion, aging, CAC payback, and forecast accuracy. Measurement covers definitions, form-to-CRM plumbing, events, and reconciliation. Then channel efficiency. Then content and search, including AI visibility. Then team and tooling. Every check names its evidence and its red flag.

Why start a marketing audit with revenue math instead of channels?

Because a channel review built on broken measurement is a review of fiction. Coverage, CAC payback, and forecast accuracy set the agenda downstream. A company at 3x coverage with 5% forecast error has a channel problem. A company at 1.4x coverage with 30% swings has a pipeline problem. Auditing its channels first wastes the engagement.

What do marketing audits usually find first?

Measurement defects. The most common is analytics overcounting leads. More than one event fires per submission, so the analytics count runs above the CRM count. Close behind are definitions that no longer match what the systems do. Then form submissions that never reach the CRM. Then source attribution overwritten to Direct.

What should happen in the 30 days after a marketing audit?

Week one fixes measurement only. Every downstream decision depends on the numbers being real. Week two closes out dead pipeline and re-baselines coverage. Week three cuts obvious waste and moves that budget to the shortest-payback channel. Week four ships one structural fix against the named constraint. Instrument it so movement is visible within a quarter.

talk to me

tell me what needs to move.

I will come to the call ready to tell you what I would do first.