Sales teams frequently report that leads got worse in the same quarter marketing reported that leads got more numerous. Both are usually accurate, and the relationship between them is mechanical.
The best sources are used first
A marketing program starts with the channels closest to buying intent: branded search, direct referrals, high-intent keywords. These are cheap, small, and produce the highest conversion rates.
Those sources are finite. Once they are fully worked, additional volume has to come from audiences that are further from a purchase decision at the moment they arrive.
Average quality falls as a result of growth rather than as a result of error. The first thousand leads and the third thousand were never drawn from the same population.
Broader content attracts broader readers
Content that ranks widely usually answers a general question, which is why it ranks. The people it reaches include students, competitors, job seekers and practitioners with no purchasing authority.
Narrower content converts better and reaches fewer people, so a team measured on volume drifts toward the general end. The incentive produces the outcome without anyone choosing it.
The same drift affects paid campaigns. Broad match terms and lookalike audiences deliver scale precisely because they relax the targeting that made the original segment valuable.
Gates change who fills the form
Lowering friction on a form raises submissions. Removing fields, dropping the phone number requirement or ungating a download all increase count and reduce the effort a visitor is willing to spend.
Effort is itself a qualification signal. A person who completes a longer form has demonstrated something about intent that a one-field submission does not demonstrate.
Teams that add friction back usually see volume drop and conversion rise, with total closed business roughly unchanged. The number moved; the underlying demand did not.
Sales behavior adjusts to the dilution
When a list is mostly poor, representatives triage rather than work it. Follow-up gets shallower, response times lengthen, and genuinely good leads sitting in the same queue are treated as if they were not.
That reaction destroys real pipeline. The loss is invisible in reporting because nobody records the qualified buyer who received one voicemail and moved on.
The behavior is rational at the individual level and harmful at the company level, which is why it persists until routing changes rather than until people are asked to try harder.
Scoring works only if it is calibrated
Lead scoring is intended to sort a diluted list, but scores built on early data describe the buyers of an earlier stage. As the mix shifts, the model's assumptions expire.
Useful calibration compares scores against closed outcomes, not against engagement. Opens and clicks measure interest in content, which is a different thing from readiness to buy.
Where scoring is honest, marketing and sales can argue about the same number instead of about impressions. That is usually the practical resolution to the quality dispute.