You can't tell whether a sales function works by looking at it. Busy reps, full calendars and a fat pipeline are all consistent with a business that closes nothing. Three numbers cut through it. They take an afternoon to calculate and they'll tell you more than a quarter of dashboards.
1. Does marketing own anything past the click?
Ask your marketing team, or your agency, one question: how many of the leads you generated last quarter became closed-won revenue?
If the answer is a number, marketing is accountable and you can work with it. If the answer is traffic, form fills, MQLs or "that's a sales question", you have two separate functions pretending to be one pipeline. Leads get handed over, sales complains about quality, marketing complains about follow-up, and nobody owns the outcome.
The fix is not a better attribution tool. It's a decision that marketing is measured on what converts, made by the owner and enforced in the reporting. In our engagements that's usually done alongside our partner Full Signal, but the principle holds whoever runs demand: one pipeline, one number, both halves accountable to it.
Test: marketing can state its closed-won contribution for the last quarter without asking sales.
2. Is sales force effectiveness above 15%?
Take closed-won deals for a period and divide by sales-qualified leads for the same period. That's your sales force effectiveness — the share of real opportunities the team actually converts.
Below about 15% in mid-market B2B and something in the deal mechanics is broken: qualification is letting in prospects who were never going to buy, demos aren't advancing deals, or proposals are going out into silence. The team isn't lazy; it's leaking.
The number is only useful if "qualified" means something. If any lead a rep is willing to work counts as an SQL, the denominator is inflated and the ratio is flattering. Define qualified — for us it's access to the decision-maker, a tested timeline, agreed value and a booked next step — and the number becomes honest, which usually means it gets worse before it gets better.
Test: closed-won over SQL for the last two quarters, with a written definition of SQL that the team uses.
3. Can leadership audit the CRM without asking a rep?
Open the CRM right now, pick five deals at random, and check: is there a logged email trail, a meeting record, a named decision-maker, a next step with a date, a close date that's been justified rather than defaulted, and a deal source that's true?
If you can answer from the system, the forecast is worth something. If you have to ask the rep, the CRM is a notebook, not a source of truth, and the forecast is a collection of opinions.
This is almost never a rep-discipline problem, however much it looks like one. It's a system problem: stages without exit criteria, fields that aren't required, close dates that default to a month out. Fix the gate rules so a deal can't advance without the evidence, and the data cleans itself up because it has to.
Test: five random deals, every field answerable from the CRM alone.
What to do with the results
Three passes and you have a sales engine; the work is optimisation. Two passes and there's one leak worth finding. One or none, and the pipeline you're planning the business on is a guess — which is the situation most founder-led businesses are in when they call us.
We run these three tests in the first week of every GTM Diagnostic. See how we work together →
