The lead quality problem

The same meeting happens in most companies once a quarter, and it never resolves. Marketing opens with a number that is up — leads, sign-ups, demo requests, whatever the site is built to collect — and sales opens with a complaint that the number does not mean anything, because half of what showed up this month was a student doing research, a competitor checking prices, or a company three sizes too small to ever become a customer. Marketing points at the dashboard. Sales points at the pipeline. Both are telling the truth about what they are looking at, and the meeting ends with an agreement to "tighten things up," which nobody ever defines, and the same argument returns next quarter unchanged.

Sales's complaint deserves to be taken seriously before anything else, because it is usually correct. A site that removes friction from its enquiry form — fewer fields, no forced sign-up, one click instead of three — will reliably produce more leads and a lower average lead, because the people who would have abandoned a longer form are, by selection, disproportionately the people with less at stake. That is not a flaw in the website. It is what friction removal does, structurally, and a marketing team that reports volume without reporting quality is not lying, but it is answering a question nobody actually asked.

Where the argument goes wrong is in treating quality as something a lead has, the way a lead has a name and an email address. It isn't. A lead is qualified or not relative to a rule, and the rule belongs to both teams jointly — what sales is willing to spend an hour on, what marketing is being asked to produce. The reason the argument recurs every quarter is that the rule was never written down. Everyone has a mental version of it, the two mental versions disagree, and a disagreement that exists only in two heads cannot be settled, only re-fought. Write the rule down and the fight changes shape: it becomes a disagreement about a sentence, which can be read, tested against real enquiries, and revised. That is a different kind of argument, and it is one that actually ends.

Four conditions, testable by someone who wasn't in the room

A qualification rule that works has a specific shape: a short list of conditions, each one answerable yes or no from information the enquiry itself contains, with no adjective left to a rep's mood that day. "Seems serious" is not a condition. "Company size" is, once you also write down the number that counts.

A workable example for a mid-market B2B seller might read: company size above twenty employees, a work email domain rather than a free one, a stated timeline of this quarter or next, and a role with either budget authority or direct influence over the buying decision. Four conditions, all four required, each one checkable against the form data without a conversation. The number of conditions matters less than the property that makes them useful: a new hire reading the rule cold should reach the same yes-or-no answer a five-year veteran would, given the same enquiry. If two people can read the same submission and disagree about whether it qualifies, the rule isn't a rule yet — it's a description of a feeling, and descriptions of feelings cannot be measured or improved.

The temptation is always to add a fifth condition, then a sixth, because every rejected lead that later became a good customer feels like evidence the rule is too narrow, and every accepted lead that went nowhere feels like evidence it's too wide. Both feelings are usually wrong, because a handful of exceptions is exactly what a four-condition rule applied at volume should produce — a rule with no exceptions has been tuned to the past rather than written for the future. Keep it short. A rule long enough to need a flowchart is a rule nobody downstream will apply consistently, which puts you right back at two mental versions disagreeing.

One owner, one review date

The rule needs an owner — a named person with authority over both the form and the pipeline, not a standing committee of the two department heads, because a committee is where "tighten things up" goes to become permanent ambiguity again. In most companies that person sits closer to sales than to marketing, since the cost of a bad lead lands there, but the arrangement matters less than the singularity of it. Two people who can each veto the other's changes will produce a rule that ratchets in one direction under pressure and never moves back.

Set a fixed review date — quarterly is a reasonable default — and treat it as the only legitimate moment to change the rule. The instinct to renegotiate mid-quarter after one bad week is exactly the instinct that produced the current mess, because a single bad week is usually noise at the volumes most companies see, a point made at length in the piece on measuring what a website does: below a few hundred events a month, most of what looks like a trend is the ordinary wobble of a small sample, not a signal asking to be acted on. The review date holds that impulse until there is enough data to know whether the rule is wrong or someone is reacting to three unlucky Tuesdays.

Rejection reasons are the only evidence the rule is wrong

A pass/fail rule is only as good as the log of what it rejected, and most companies don't keep one. Every lead that fails qualification should be tagged with which condition it failed — not "not a fit," which explains nothing to anyone reading the log later, but "employee count below threshold" or "no timeline stated," specifically. Six months of that log is the only honest way to find out whether a condition is doing its job.

This is where source matters, because different channels fail the rule differently, and a rule tuned on one channel's failure pattern will misjudge another's. Outbound channels report their own distinct rejection shape — the argument in cold email versus LinkedIn outreach about how each channel front-loads a different kind of interest applies here too: a cold email reply and an inbound form fill do not fail qualification for the same reasons, and a log that pools them together hides which channel needs the tighter screen. If ninety percent of a month's rejections fail the same single condition, either that condition is badly written, or a channel just changed shape and the form is catching it — both worth knowing, and only the log tells you which.

Rejection reason	Company size	No work email	No timeline	No budget authority
Share of monthly rejections	41%	22%	24%	13%

A table like that, reviewed at the quarterly meeting, turns "leads feel worse lately" into a specific claim: the company-size threshold is catching more small accounts than it used to, which is either the market shifting or the campaign targeting drifting, and both are answerable questions rather than moods.

The rule should decide what the form asks and what the site offers

Once the four conditions exist, the website's job is to collect exactly the information the rule needs and to route each visitor toward the offer that matches where they actually sit. This is the practical payoff of writing the definition down: it stops being a filter applied after the fact and starts shaping the page itself.

If timeline is a condition, the form asks about timeline — as a real field, not folded into an open "tell us about your project" box that a rushed visitor will leave blank. If company size is a condition, ask for it directly rather than inferring it from an email domain that increasingly tells you nothing, because a large share of enquiries now arrive from a personal address regardless of company size. The offer itself should split by where the rule expects someone to be: a visitor who has not stated a timeline is better served a low-commitment offer — a guide, a calculator, a newsletter — than a "book a call" button neither side is ready for. Promoting the highest-commitment offer to everyone maximizes the volume number and damages the quality number in the same motion, the exact split the piece on the only website metrics that mean anything warns against: a metric that goes up for reasons indistinguishable from success is not a result, and a lead count with no quality attached is precisely that kind of metric.

The rule can be optimized against, and that is the failure mode to watch for

The uncomfortable ending is that a written rule solves the argument and creates a new problem, because any measurable target eventually gets targeted rather than served. Once "qualifies against the four conditions" is the number marketing is judged on, a campaign that nudges visitors toward whichever answer passes the screen — a landing page that mentions the company size threshold before the form asks for it, copy that primes visitors to select a timeline they don't actually have — will raise the pass rate while doing nothing for what the rule was built to protect. The form gets smarter about the rule without the leads behind it getting any better, and everyone in the room is confused, because the metric that was supposed to end the argument now shows improvement while sales's complaint returns unchanged.

The fix isn't a stricter rule — a rule tightened in response to gaming just gets gamed at the new threshold. It's auditing the step after qualification rather than the qualification rate itself: not how many leads pass the four conditions, but what happens to the ones that do, tracked to close or to a dead pipeline, on the same cadence as the rejection log. A qualification rate can be improved by writing better landing-page copy. A close rate on qualified leads cannot be improved by writing copy at all — it can only be improved by the leads actually being what the rule said they were, which is the only version of "better leads" either department should be optimizing for in the first place.