If You Can't Fire Them on the Numbers, You Didn't Set the Numbers

The test of whether you’ve set real accountability is simple. If performance were clearly bad, could you point at a number you both agreed to in advance and act on it?

If the answer requires a debate about context, you didn’t set accountability. You set a vibe.

Set the numbers before the work starts

Both sides should be able to state, in writing, what this engagement is accountable for and by when. Three to five numbers, no more.

Which numbers depends on your business, and any leader who arrives with a fixed list hasn’t understood it yet. But they should be commercial rather than promotional. Qualified pipeline, cost per customer, close rate, blended return, revenue by segment. Not impressions, not followers, not content published.

What to look at monthly

Blended return across the whole portfolio. Total marketing spend against revenue attributable to marketing. Crude on purpose, and the one number no platform can inflate. Why it’s the only figure that can’t be gamed.

Cost per customer, not cost per lead. Cost per lead is the easiest metric to improve and the least connected to revenue.

Stage-to-stage conversion. Visitor to lead, lead to qualified, qualified to closed. Movement here is usually the earliest real signal that something structural improved.

Whether the reporting reconciles. If marketing’s numbers exceed what finance recorded, the reporting is fiction and that’s the finding.

What takes quarters, and shouldn’t be judged monthly

Organic search, content, brand, reputation, partnerships and email list quality all compound. Judging them monthly produces exactly the wrong behaviour, which is abandoning assets before they mature in favour of paid media that reports immediately.

Track them, review them quarterly, and be explicit with each other about which bucket a given initiative sits in before it starts. That single agreement prevents most of the friction in the first year.

Metrics that punish good decisions

Some numbers actively make a good leader look bad.

Lead volume. A leader who correctly tightens targeting will reduce lead volume and increase revenue. If volume is the headline metric, they get penalized for the right call.

Cost per lead. Same problem. Better-qualified leads usually cost more.

Total spend. If reducing wasted spend shows up as “budget underutilized,” you’ve told them to keep spending.

Activity counts. Campaigns launched, posts published, emails sent. These measure motion.

The qualitative signals that matter

Some of the strongest indicators aren’t numbers.

Does leadership have a clearer picture of the business than six months ago? Do marketing and sales agree on what qualified means? Can your team explain the strategy without the CMO present? Has anyone told you something you didn’t want to hear?

That last one is the highest-signal question. A marketing leader who has never delivered an unwelcome message in six months is either extraordinarily lucky or managing your perception.

The quarterly conversation

Once a quarter, both sides answer the same question honestly: is this returning more than it costs?

That’s the real accountability mechanism for an open-ended engagement, and it’s better than an artificial end date because it’s tied to value rather than to the calendar. A long engagement isn’t a problem. An engagement nobody can justify anymore is.

If the numbers aren’t moving and neither side can explain why, the answer should be that the arrangement ends. A leader unwilling to have that conversation openly is protecting the retainer rather than the outcome.