Every quarter, a marketing team is asked the same question: how much pipeline did you create? So they pour budget into the people who are buying right now, the form-fillers, the demo-requesters, the ones already shopping. It works, briefly, and then it stalls, because the pool of people ready to buy this quarter is tiny and everyone is fishing it. The growth was never going to come from harvesting demand. It comes from creating it, months before anyone fills in a form.

The quick version

  • Demand generation creates and builds buying interest across your whole market, most of whom aren't ready to buy yet. Pipeline creation (often "lead gen") captures the interest that has matured into intent and turns it into qualified opportunities sales can work.
  • At any moment, roughly 95% of business buyers are not in-market, only about 5% are actively buying in a given quarter. Chasing only that 5% caps your growth (Ehrenberg-Bass Institute's "95:5 rule", John Dawes, 2021).
  • The two motions need different budgets and different metrics. Binet & Field's analysis of nearly 1,000 case studies found long-term brand-building and short-term activation both pay off, but on different timescales.
  • The move: fund demand creation as a long-term asset, instrument pipeline creation as a near-term conversion engine, and stop scoring the first by the metrics of the second.

The idea in depth

The words get used interchangeably in budget meetings, and that is exactly where the money goes wrong. "Demand generation" and "pipeline creation" describe two different jobs, separated by where the buyer is in their decision, and by how long you have to wait for a return.

Demand generation is the work of making buyers want, and remember, your category and your brand: the content, advertising, events, point of view and reputation that build interest across your whole addressable market, including the large majority who won't buy for months or years. Pipeline creation is the work of capturing the interest that has matured into intent: the gated content, inbound enquiries, outbound prospecting and qualification that produce named opportunities a salesperson can pursue this quarter. Marketers sometimes split this as demand creation versus demand capture; the practitioner guides at ZoomInfo and Cognism make the same distinction. One fills the top of the funnel; the other pulls people through it.

flowchart LR
    M(["Whole market
most not buying yet"]) --> DG(["Demand generation
create & build interest"])
    DG --> I(["Interest matures
into buying intent"])
    I --> PC(["Pipeline creation
capture & qualify"])
    PC --> O(["Qualified opportunities
sales can work"])
Two jobs, one funnel, demand generation creates interest across the market; pipeline creation captures the share of it that has turned into intent. Leaders Loop

Why most of your market isn't buying, and why that matters

The single most useful number in this whole subject comes from the Ehrenberg-Bass Institute. In a 2021 paper for LinkedIn's B2B Institute, Professor John Dawes set out what is now called the "95:5 rule": across many categories, only about 5% of business buyers are in-market in a given quarter, because the average B2B purchase cycle runs to several years. Put bluntly, at any moment roughly 95% of your potential buyers are not buying anything. That doesn't make them worthless, it makes them future pipeline.

This reframes the whole argument. If you only ever speak to the 5% who are shopping today, you are competing with every rival in the most expensive, most crowded sliver of the market, and you have done nothing to influence the 95% who will be shopping next year. The companion idea, from Byron Sharp and Jenni Romaniuk's work at the same institute, is mental availability: the job of demand generation is to get your brand linked, in memory, to the buying situations (the "category entry points") that will eventually trigger a purchase, so that when a dormant buyer wakes up, you are already on the shortlist (Marketing Week on Ehrenberg-Bass).

Treat demand generation as building an asset, not running a campaign. Reach the 95% who can't buy yet, with a memorable point of view tied to the moments they'll buy in, and accept that you can't measure the return this quarter. You're not generating leads here. You're making sure you're remembered when the lead generates itself.

The budget split: long-term brand vs short-term activation

The most-cited evidence on how to divide the money comes from Les Binet and Peter Field, whose The Long and the Short of It analysed the IPA's databank of effectiveness case studies, close to a thousand campaigns across hundreds of brands. Their headline finding: the best long-run return came from splitting budget roughly 60% to long-term brand-building and 40% to short-term sales activation. In their later B2B-specific work with the LinkedIn B2B Institute, the optimal split moves closer to 50/50, with somewhat more weight on activation, because B2B cycles are long and buying groups are large. The central lesson holds either way: you need both, and they work on different clocks.

The mechanism is the part leaders miss. As Binet puts it, the "long" makes the "short" work better over time, brand-building lowers the cost and lifts the conversion of your lead generation, but it doesn't run in reverse. No amount of activation spend builds a brand. Pour everything into capture and each new quarter starts from a cold, price-competing standstill.

Demand generation is future pipeline; lead generation is current pipeline. You can't capture demand that was never created.

Run two budgets with two scorecards, not one. Ring-fence a brand/demand-creation budget judged on long-term measures, share of search, brand recall, inbound volume over quarters, and judge the pipeline-creation budget on near-term conversion: opportunities created, cost per opportunity, win rate. The failure mode is a single "pipeline this quarter" target that quietly starves the long-term engine to feed the short-term number.

Pipeline creation needs a shared definition, the demand waterfall

Where demand generation is fuzzy by nature, pipeline creation lives or dies on definitions. The reference model is the Demand Waterfall, created by SiriusDecisions (now part of Forrester) and the de-facto industry standard for the stages a lead passes through: inquiry, marketing-qualified lead (MQL), sales-accepted lead (SAL), sales-qualified lead (SQL), and on to a won deal. Forrester's later Demand Unit Waterfall updated it to track buying groups rather than individuals, a sensible fix, since B2B purchases are rarely made by one person.

The point of a waterfall isn't bureaucracy; it's a shared language so marketing and sales agree on what a "qualified" lead actually is, and so you can see where leads leak between stages. Without it, marketing celebrates MQLs that sales quietly bins, and nobody can tell whether the problem is too few leads or too poor a hand-off.

Define each stage with sales, in writing, and instrument the drop-off between them. Agree what makes an MQL worth a salesperson's time, then watch the conversion rate at every step. A waterfall that leaks at MQL→SAL is a quality problem, you're passing junk. One that leaks at SQL→won is a sales or fit problem. Neither diagnosis is possible until the stages are named.

The honest limitation: this gets gamed, and the long bet is hard to defend

Two cautions keep this from becoming dogma. First, the metrics invite gaming. Because pipeline targets are immediate and visible, teams optimise for the number that gets them through the QBR, stuffing the funnel with low-intent MQLs, or counting "demand" the brand never created (the buyer who was already shopping and would have found you anyway). A rising MQL count can mask a falling win rate. Always pair a volume metric with a quality one.

Second, the long-term case is genuinely hard to hold. The 95:5 rule and the 60/40 split are robust, repeatedly-found patterns, but they are averages across many categories, not a guarantee for your specific product, segment or sales cycle. A startup that needs revenue this quarter to survive cannot allocate like an established brand playing a five-year game; the right split depends on your runway, your category's purchase frequency and how known you already are. Treat these as a strong prior to argue from, not a law to obey, and measure your own funnel rather than importing someone else's benchmark.

A worked example

The figures below are illustrative, chosen to show the mechanics rather than to report a real company.

Picture a B2B software company selling to finance teams, with a sales cycle that averages about a year. The board wants more pipeline, so last year the team put nearly all of its budget into capture: paid search on high-intent keywords, gated "request a demo" pages, an outbound SDR team hammering a bought list. It worked for two quarters, cost per opportunity was low because they were skimming the buyers already in-market, and then it flattened. They had drained the 5% pool, and CPLs climbed as they bid against rivals for the same shrinking set of ready buyers.

This year they rebalance. Roughly half the budget still funds capture, but the qualification is tightened: marketing and sales rewrite the MQL definition together, and the SDRs stop chasing anyone who downloaded a guide and instead pursue accounts showing real buying signals. The other half funds creation: a genuinely useful point of view on a finance-team problem, distributed where finance leaders already spend attention, tied to the moments those teams actually start a buying search (a new system, an audit, a growth spurt).

flowchart TD
    A(["~95% of market
not buying yet"]) --> B(["Demand gen: useful POV
tied to buying moments"])
    B --> C(["Mental availability
remembered on the shortlist"])
    C --> D(["A buyer enters the 5%
and thinks of you first"])
    D --> E(["Pipeline creation
capture & qualify vs agreed MQL"])
    E --> F(["Lower cost per opportunity
higher win rate over time"])
The long bet feeding the short one, demand created among the 95% lowers the cost of capturing the 5% later. Illustrative. Leaders Loop

For two quarters the new mix looks worse on the spreadsheet: half the capture budget, so fewer raw leads, and the creation spend shows no direct pipeline at all. Then the curve bends. Branded and direct enquiries rise; sales report that more first calls start with "we've heard of you" instead of a cold explanation; win rate ticks up because the brand did some of the convincing before the salesperson arrived. The same pipeline target is now hit at a lower cost per opportunity, not because capture got cleverer, but because demand got created upstream. The lesson is the order of operations: you cannot harvest a field you never planted, and the planting never shows up in this quarter's harvest.

Frequently asked questions

What's the real difference between demand generation and lead generation?

Demand generation creates interest across your whole market, most of whom aren't ready to buy; lead (or pipeline) generation captures the interest that has turned into intent and hands sales a named opportunity. A quick test: if the activity reaches people who can't buy yet and builds future preference, it's demand generation; if it converts existing intent into a contactable, qualified lead now, it's lead generation. You need both, demand gen makes lead gen cheaper, but not the other way around.

If only 5% of buyers are in-market, isn't spending on the other 95% a waste?

It's the opposite, it's the only way to grow beyond the 5% everyone else is also chasing. The Ehrenberg-Bass "95:5 rule" (John Dawes, 2021) says the in-market pool is small at any moment, so competing only there caps your ceiling and bids your costs up. Reaching the 95% builds the memory and preference that means you're already shortlisted when they enter the market. You just can't measure that return this quarter, which is why it gets cut first and shouldn't be.

Who should own pipeline creation, marketing or sales?

Both, against one agreed definition. Pipeline creation falls apart when marketing is measured on MQL volume and sales on closed revenue, with no shared standard for what "qualified" means in between, marketing then passes leads sales won't touch. Use a demand-waterfall model so the hand-off (MQL → sales-accepted → sales-qualified) is defined jointly and the leak between stages is visible. The owner of the number can sit in either function; the definition must be shared.

How do I justify demand-generation spend when the board wants pipeline now?

Run two budgets with two scorecards. Judge pipeline creation on near-term conversion (opportunities created, cost per opportunity, win rate) and demand generation on leading indicators that move over quarters (share of search, brand/aided recall, direct and branded inbound volume, the share of deals that start warm). Binet & Field's roughly 60/40 brand-to-activation split, moving closer to 50/50 in B2B, is the defensible starting point for the argument, adjusted for your runway and how known you already are.

Won't more MQLs always mean more revenue?

No, and assuming so is the classic trap. MQL count is a volume metric and it's easy to inflate with low-intent leads that never convert; a rising MQL number can hide a falling win rate. Always pair a volume metric with a quality one, track conversion from MQL to sales-accepted to won, not just the top-of-funnel total. If MQLs are up but win rate and cost per opportunity are getting worse, you're manufacturing activity, not pipeline.

Related in the Toolkit

Where to go next