Lead quality

Every lead generation account eventually has the same conversation. Cost per lead has risen, or volume has fallen, and the agency explains the auction while the sales team says the leads are rubbish anyway.

Both sides are describing something real and neither can see the other half. The agency has platform data that stops at the form submission. The sales team has an impression formed from the leads they happened to call. Nothing connects them.

The agency loses that argument by default, and it is worth being clear why: volume is a number and quality is an anecdote, so the side with the number is assumed to be defending itself. A rising cost per lead with no qualification data attached looks like an excuse regardless of being true.

With qualification data it becomes a different conversation entirely. Cost per qualified lead falling while cost per lead rises is a good quarter, and it is a sentence that cannot be argued with. That single metric is usually worth more to an agency relationship than anything else in the reporting.

The data lives in the client CRM, and the timing of the request matters more than the mechanism. Asked at kickoff, alongside analytics and ad account access, it is routine. Asked after a quality complaint, it reads as an agency building a defence.

By that point the narrative has usually formed, and the numbers arrive into an argument rather than into a report. The same figures that would have been neutral in month one are contested in month six.

The request does not have to be technical to begin with. A monthly export of leads with a status column is enough to compute a qualification rate, and it can be a spreadsheet emailed by whoever runs the CRM. Integration can come later once the number is proving useful.

What matters more than the pipe is agreeing the definition. What counts as qualified, who decides, and at what point. Two people using the word differently will produce a metric that moves for reasons nobody can trace, which is worse than not having it.

Qualification data is useful in a report and considerably more useful pushed back into the platforms, where automated bidding can use it.

Google Ads offline conversion import and Meta Conversions API both exist for this. The pattern is the same: the platform gives each lead an identifier at the click, the client CRM stores it against the record, and when the lead qualifies the identifier is sent back with the outcome attached.

The effect is that bidding stops optimising for form submissions and starts optimising for the ones that turn into business. That is a genuine change in what the account buys, and on lead generation accounts it is usually the single largest improvement available.

It needs the CRM to cooperate, which is the real constraint. Storing an identifier against a lead record is a small development task and it is somebody else backlog, so it belongs in the kickoff conversation rather than being discovered as a dependency later.

The feedback loop is only useful if it closes faster than the optimisation window. A lead qualified three weeks after the click cannot inform this week bidding, and in long sales cycles it may not inform this quarter.

Where the cycle is short, weeks rather than months, importing the qualified outcome works well and is worth the setup. Where it is long, the honest answer is that the qualified conversion is a reporting metric rather than a bidding signal.

The workaround is to find an earlier proxy that correlates with qualification and arrives sooner. A completed discovery call, a demo attended, a proposal requested. Each is imperfect and each arrives within days rather than weeks, which is what makes it usable for bidding.

Choosing the proxy is a judgement worth making with the client rather than for them, because it decides what the account optimises toward for the next quarter. And it is worth revisiting: a proxy that correlated with qualification in January may not in July if the mix of enquiries has changed.