USE CASE: Demand Gen vs Lead Gen: How to Stop Judging One by the Other’s Numbers in a Web Design Agency

Demand gen vs lead gen isn’t a semantic quibble: they’re two different jobs, and measuring both with one KPI set quietly kills the one that doesn’t fit it. This is what untangling them looks like in practice: how a web design agency separated the two disciplines, gave each its own budget and KPIs, and used shared qualification language to keep them connected.

The Context: Two Disciplines, One Line Item

A web design agency that did both kinds of marketing without quite naming them as two: some work built awareness and interest in what it does; other work turned interested people into enquiries. All of it came out of one “marketing” budget and got reported against one set of numbers. It felt efficient. It was quietly costing the agency its own future demand.

The Challenge: Two Jobs, Measured with One Ruler

Demand generation and lead generation do opposite jobs. Lead gen captures existing demand: it converts the small slice of the market that’s in-market right now into enquiries. Demand gen creates demand: it builds awareness and want among the far larger group who aren’t in-market yet but will be. The agency ran both from one budget and judged both by one KPI set: inevitably the lead-gen one, because leads, cost per lead and conversion are the numbers that show up fast and clearly. And there’s the trap: judged by lead-gen metrics, demand gen always looks like it’s failing, because it doesn’t produce cheap enquiries this month, and that was never its job. So it gets cut. And cutting it starves the very demand that lead gen exists to harvest: you make the field cheaper to harvest for two quarters, then wonder why there’s less to harvest for two years. The agency wasn’t overspending or underspending. It was measuring two different jobs with one ruler, and quietly defunding the one that didn’t fit.

Demand gen creates want, lead gen captures it: They’re two jobs, not two names for one. Put them on a single budget and a single KPI set, and it’s always the lead-gen ruler, because those numbers arrive fast; demand gen, judged by a metric it was never meant to move, looks like it’s failing. So you cut it. And you’ve just stopped sowing the field you were about to harvest.

The GenAI Workflow: Clarify Both, then Write the Language that Connects Them

The fix had two moves. First, clarify the two disciplines. The team used GenAI to sort the agency’s actual marketing activities into two buckets (which ones create demand, which ones capture it) and to flag wherever a demand-gen activity was being judged by a lead-gen metric like cost per lead. From that, GenAI proposed the right KPIs for each, so neither would be measured by the other’s ruler:

Second, the connective move: draft the shared qualification language. Separating the two disciplines risks turning them into silos, so the team used GenAI to draft an agreed, written definition of what actually counts as a “qualified lead”: one that demand gen, lead gen and sales could all sign up to, so the two stayed connected, demand feeding lead gen feeding sales, rather than optimising separate numbers in isolation. GenAI drafted the definition from how the agency really qualifies; the humans argued it to agreement. Two disciplines, two budgets, two rulers, held together by one shared definition of what they’re all working toward.

The caveat that decides whether this works: The split GenAI proposes is a starting frame, not a verdict; it knows the textbook line between demand gen and lead gen, not how this agency actually earns business, so the buckets are a draft to adapt to your real activities.

Three more considerations:

• The shared qualification language is the real point of leverage, and it’s the part GenAI can least finish for you: a definition of “qualified” only works if demand gen, lead gen and sales genuinely agree it; that’s alignment work, the same “definition debt” that breaks handoffs, and no draft substitutes for the argument that produces agreement.
• Clarity must not curdle into silos: the reason to separate the two disciplines is to fund and measure each properly, not to wall them off: demand gen exists to feed lead gen, and a version of this that turns them into two teams optimising two numbers in isolation has just rebuilt the problem in a new shape.
• Measuring demand gen honestly means resisting false precision: its whole nature is long-term and hard to attribute, so the urge to force a clean, lead-gen-style number onto it is the very mistake that got it defunded; measure it directionally, and don’t let GenAI manufacture a precision the data can’t support.

GenAI clarifies and drafts; the agreement, the connection, and the honesty about what can be measured stay human.

The Result: Two Jobs, Funded and Measured as Two Jobs

The agency stopped judging two jobs by one number. With demand gen and lead gen clarified into separate disciplines (each with its own budget and its own KPIs), demand gen was no longer failing a test it was never meant to sit, and stopped getting cut for missing a target that wasn’t its own. Lead gen kept being measured on leads and cost per lead, exactly as it should be; demand gen got measured on whether it was building the awareness and want that lead gen would later convert. And because the two were tied together by shared qualification language rather than split into silos, the clarity connected them instead of dividing them: demand feeding leads feeding sales, on one agreed definition of “qualified”. No invented figures here: the change is that the agency stopped starving the field it depended on harvesting, and started funding and measuring each discipline as the different thing it was.

Recommended KPIs to Follow

This case is itself about measuring each discipline on its own terms, so the KPIs practise what it preaches: demand gen judged on demand-gen outcomes, lead gen on lead-gen outcomes, and a shared definition holding them together. Here’s where the evidence sits and the direction this should push things. The point is the direction of travel, not a promised number.

Demand-Gen KPIs (measured on what demand gen actually moves)

Reach and engagement among future buyers, and influence on later pipeline, and not cost per lead. The whole point is to judge demand gen by the job it does (building future demand), so it stops looking like a failure and stops getting cut.

Benchmark: Direction, not a promise: at any moment only ~5% of buyers are in-market, so demand gen’s real audience is the ~95% who aren’t yet (Ehrenberg-Bass / LinkedIn B2B Institute), and the Binet-Field evidence is blunt: starve the long (brand/demand) side and activation gets cheaper for two quarters, then less efficient for two years (Binet & Field, IPA Databank). Strategic lens, not a precise rule.

Lead-Gen KPIs (leads, cost per lead, conversion)

The fast, cleanly measurable numbers, kept exactly as they are, but applied only to lead gen, where they belong. The fix isn’t to abandon CPL; it’s to stop pointing it at work it was never meant to measure.

Benchmark: No single transferable figure, CPL and conversion vary widely by channel and offer; baseline your own and track the trend. The discipline here is scope, not target: these numbers judge lead gen, never demand gen.

Shared-Qualification Adoption (and lead quality)

Whether the agreed “qualified lead” definition is actually used across demand gen, lead gen and sales, and whether lead quality improves as a result. This is the connective tissue that stops separation becoming silos.

Benchmark: Direction, not a promise: shared lead definitions are reported to lift MQL-to-SQL acceptance from a typical ~13% toward ~20-25% (LaGrowthMachine). Advocacy-sourced, treat as direction; your own acceptance rate is the real measure.

Demand gen on its own outcomes, lead gen on its own, a shared definition connecting them: the KPI set demonstrates the framework. The 95:5 figures (or 60:40 figures as found in other sources) are strategic lenses, well-evidenced, but not precise operating rules; the alignment figure is advocacy-sourced. And note the through-line to the attribution case: demand gen is exactly where clean attribution fails, so measure its contribution directionally and resist the false precision that got it defunded in the first place.

Why This Transfers

Almost every marketing function runs demand gen and lead gen together and reports them on the lead-gen ruler, because that ruler is the one that reads clearly, and so quietly defunds the demand work that makes the leads cheap in the first place. The transferable move is to name the two as two jobs, give each its own budget and its own KPIs, and bind them with one agreed definition of a qualified lead. Measure the field and the harvest by the same yardstick, and you’ll stop sowing.

Recommended Articles:

Demand Generation vs. Lead Generation: The Distinction That Changes How You Budget and Measure

Demand generation and lead generation are treated as synonyms, run on one budget, and measured with the same KPIs. They are different disciplines with different jobs, timescales, and failure modes. Here is the distinction, and what it changes about how you budget and measure.

Diana SocaciuDiana Socaciu