Paid Media for B2B Demand Generation, Beyond Click to Lead

Most B2B paid media is graded on the wrong moment. The click and the form fill feel like the result, so that is what teams optimise. But the click is the start of a 22-person buying process, not the finish line. The average B2B buying decision now involves 13 internal stakeholders plus 9 external influencers (Forrester, 2026). A single captured lead is a fragment of that committee. This piece reframes b2b demand generation paid media as two distinct jobs, creating demand and capturing it, and shows why measuring both on cost per lead is a category error.

Key takeaways

• Only about 5% of your market is in-market in a given quarter (Ehrenberg-Bass, 2024). Click-to-lead paid media harvests the 5% and ignores the 95% who decide future shortlists.
• An 11-person buying group throws off 150 to 200 touchpoints per vendor, yet multi-touch models credit only 3 to 4 (6sense, 2025). You are steering on roughly 2 to 3% of the data.
• 95% of deals go to a vendor already on the buyer's Day One shortlist (6sense, 2025). Paid media's real job is mental availability before the trigger.
• Demand generation and lead generation are not synonyms. One creates future buyers; one captures present ones. Budget and metrics should differ.
• Owned data, self-reported attribution and account-level signals beat last-click for measuring what paid media actually moved.

What is the difference between demand generation and lead generation?

Demand generation creates future buyers; lead generation captures present ones. Only about 5% of B2B buyers are in-market in a given quarter (Ehrenberg-Bass, 2024). A gated PDF harvests that in-market 5%. It does almost nothing for the 95% who will build next year's shortlists.

Practitioners like Chris Walker at Refine Labs reframe this as brand to demand to expand (Refine Labs, 2024). The argument is blunt. Webinar registrations, content downloads and MQL volume are capture mechanisms. They target people who already want the category. They do not create want where none existed.

This matters for budget. Capture spend competes for a small, fixed pool of active buyers, so it inflates cost as you scale. Creation spend grows the pool. If you only run capture, you cap your own ceiling and call it efficiency.

Demand generation builds future pipeline by reaching out-of-market buyers; lead generation captures the roughly 5% who are in-market now (Ehrenberg-Bass, 2024). Most teams label both as demand gen and measure both on cost per lead, which rewards harvesting and starves creation. Splitting the two is the first fix.

Why MQL volume is a misleading scoreboard

MQL volume tells you how many people filled a form, not how many were going to buy anyway. Only 22% of organisations measure closed-won deals from target accounts, and just 13% report closed-won revenue to the board (6sense, 2025). Meanwhile 39% still report MQAs upward. The scoreboard measures activity, not revenue.

Why does click-to-lead attribution miss most of the buying journey?

Because your model only sees the touches that happen on properties you can track. A typical 11-person buying group generates 150 to 200 digital touchpoints per vendor, yet multi-touch attribution credits only 3 to 4 of them (6sense, 2025). That is roughly 2 to 3% of real engagement.

The missing 97% is mostly dark social. Slack threads, peer DMs, podcast mentions, community posts, a forwarded newsletter. None of it carries a UTM. So buyers commonly type "Google search" into your form when the real trigger was a colleague's recommendation (6sense, 2025). Last-click then credits Google for a deal it merely witnessed.

This is expensive to get wrong. The blind spot grows as deal size grows, and survey buyers report median deals worth six figures annually (6sense, 2025). Optimising paid media to a model that sees 2% of the journey is steering a car using one dashboard light.

A typical 11-person buying group produces 150 to 200 digital touchpoints per vendor, but multi-touch attribution credits only 3 to 4, roughly 2 to 3% of real engagement (6sense, 2025). The other 97% is dark social with no UTM. Optimising paid media on last-click means tuning a campaign on a fraction of the evidence.

Where European teams feel this most

Europe sharpens the problem. Privacy regulation weakens cookie and click tracking further, and 40% of the 6sense 2025 sample sits in the UK, Ireland and Continental Europe (6sense, 2025). GDPR-era measurement makes click-level attribution even thinner here, which strengthens the case for asking buyers directly.

How should you split a B2B paid media budget between brand and demand?

Run two budgets with two jobs, not one budget chasing one metric. Binet and Field's effectiveness research for the LinkedIn B2B Institute sets out five principles, including balancing brand building and activation across the funnel (IPA, 2021). The widely cited working split lands near 50/50 over time; confirm the exact figure in the full report.

The 95:5 rule explains why the brand half is not a vanity line. With only ~5% in-market now, demand creation has to reach the 95% before their trigger event (Ehrenberg-Bass, 2024). You are buying mental availability so you sit on the shortlist when they enter the market.

That is not soft. 95% of deals are won by a vendor already on the buyer's initial shortlist, and 94% of buyers rank that shortlist before contacting any seller (6sense, 2025). Brand spend is how you buy your way onto Day One.

Two jobs, two metrics

The two budgets need separate scorecards. Judging brand reach on cost per lead guarantees you cut it, because creation rarely produces a same-week form fill. The table below maps the split.

Dimension	Demand creation (brand)	Demand capture (lead/activation)
Target	The ~95% out-of-market	The ~5% in-market now
Goal	Mental availability, shortlist entry	Efficient conversion of active demand
Channels	Reach campaigns, thought leadership, podcasts	Search, retargeting, high-intent LinkedIn
Wrong metric	Cost per lead	Reach or impressions alone
Better metric	Share of voice, branded search lift, self-reported influence	Pipeline and closed-won from target accounts

Binet and Field's LinkedIn B2B Institute research recommends balancing brand building and activation, with a long-run split near 50/50 (IPA, 2021). The 95:5 rule explains the brand half: with only ~5% of buyers in-market now, creation must reach the 95% before their trigger so the brand sits on the Day One shortlist that wins 95% of deals (6sense, 2025).

Why optimise for the buying committee instead of the lead?

Because the lead is one voice in a crowded room. The average B2B decision now involves 13 internal stakeholders and 9 external influencers, and procurement is a decision-maker in 53% of buying cycles (Forrester, 2026). A captured lead is a single seat at a 22-person table.

Consensus is also what produces good deals, not bad ones. Buying groups that reach agreement are 2.5x more likely to report a high-quality, low-regret purchase (Gartner, 2024). Paid media that only reaches one persona leaves the rest of the committee uninformed and the deal fragile.

So design campaigns for buying-group coverage. Reach multiple personas across multiple departments inside the same account, then measure account-level engagement rather than counting individual MQLs. The unit of demand is the account, not the email address.

The average B2B buying decision now involves 13 internal stakeholders plus 9 external influencers, with procurement decisive in 53% of cycles (Forrester, 2026). Buying groups that reach consensus are 2.5x more likely to report a high-quality deal (Gartner, 2024). Paid media should cover the committee and report account-level engagement, not individual MQLs.

What does the rep-free paradox mean for your ads?

It means optimising for the form fill can actively hurt deal quality. 67% of B2B buyers now prefer a rep-free experience, up from 61% in mid-2025, and 45% used AI during a recent purchase (Gartner via Demand Gen Report, 2026). Buyers want self-service. The trend is accelerating.

Here is the paradox. Buyers who go fully self-serve buy worse. They are 1.8x more likely to close a high-quality, low-regret deal when they combine supplier digital tools with a sales rep, and 59% regret at least one software purchase from the prior 18 months (Gartner, 2024). Independence feels good and ends badly.

The takeaway for paid media is precise. Do not optimise ads purely for the cheapest form fill. Optimise to make the rep-assisted, committee-wide journey easier. Feed buyers the self-service content they want, then connect them to a human at the moment that lifts deal quality. The click starts that process; it does not end it.

67% of B2B buyers prefer a rep-free experience, up from 61% a year earlier (Gartner, 2026), yet buyers are 1.8x more likely to close a high-quality, low-regret deal when they pair supplier digital tools with a rep, and 59% regret a recent software purchase (Gartner, 2024). Optimise paid media to enable the rep-assisted journey, not just the cheapest form fill.

How do you measure demand generation with owned data?

You ask buyers directly and you watch account-level signals, because the click data only shows 2 to 3% of the journey (6sense, 2025). Self-reported attribution closes the gap that last-click cannot. A "how did you hear about us" field surfaces the dark social that no UTM ever captured.

This is where owned infrastructure earns its keep. When you run outbound on owned domains, dedicated IPs and your own first-party data, the engagement signals belong to you, not to a platform you rent. Empra has sent 1.6 million-plus emails at a 7.4% reply rate across 40-plus B2B teams, and the data from that motion stays measurable inside the account, not locked behind an ad platform's modelling. Pair that owned outbound layer with paid demand creation and you can connect specific accounts to specific touches.

Then change the report. Replace MQL counts with self-reported influence, branded search lift, and pipeline plus closed-won from target accounts, the metrics only 22% of teams currently track (6sense, 2025). And treat AI as a human-in-the-loop role inside that measurement layer, not an unattended tool. A person reviews what gets sent and what gets reported.

If you want the channel-level version of this, our note on paid media for B2B and the breakdown of LinkedIn ads for B2B lead generation go deeper on execution. The GTM glossary defines the supporting terms.

Click data exposes only 2 to 3% of the real buying journey (6sense, 2025), so demand generation must be measured with owned, self-reported and account-level data. Asking buyers what influenced them recovers the dark social no UTM captured. Empra runs this on owned domains and first-party data, with 1.6M+ emails sent at a 7.4% reply rate.

A practical measurement stack

1. Self-reported attribution: an open "what prompted you to look" field on every form.
2. Account-level engagement: track touches per account, not per lead, across the committee.
3. Branded search lift: a proxy for demand created upstream of any click.
4. Pipeline and closed-won from target accounts: the board-level metric most teams skip.
5. Owned first-party signals: outbound replies and engagement you keep, not rent.

Frequently asked questions

Is demand generation the same as paid media?

No. Paid media is a channel; demand generation is a job that channel can do. Paid media can create demand by building mental availability with the ~95% out-of-market, or capture it from the ~5% in-market now (Ehrenberg-Bass, 2024). Most teams only do capture.

Should I stop using cost per lead entirely?

No, but stop using it as your only metric. Cost per lead is fine for the capture half of the budget. It is the wrong scoreboard for demand creation, which rarely produces a same-week form fill. Only 22% of teams measure closed-won from target accounts (6sense, 2025), the metric that actually matters.

How much budget should go to brand versus activation in B2B?

Binet and Field's LinkedIn B2B Institute research points to balancing brand and activation, with a long-run split commonly cited near 50/50 (IPA, 2021). Confirm the exact figure in the full report. Newer brands often start activation-heavy, then shift toward brand as the in-market pool gets too expensive to scale.

Why does dark social matter for paid media attribution?

Because most influence happens where no UTM exists. An 11-person buying group creates 150 to 200 touchpoints, but multi-touch models see 3 to 4 (6sense, 2025). Slack shares and peer recommendations get logged as "Google search" on forms, so last-click misreads the trigger.

Can AI run demand generation measurement on its own?

It should not run unattended. AI is useful for clustering self-reported attribution and surfacing account-level patterns, but treat it as a human-in-the-loop role. A person reviews the inputs and the outputs. For the full argument, see Empra's note on the human gate; book a call to map it to your stack.

Frequently asked questions

Is demand generation the same as paid media?

No. Paid media is a channel; demand generation is a job that channel can do. Paid media can create demand among the roughly 95% out-of-market, or capture it from the roughly 5% in-market now (Ehrenberg-Bass, 2024). Most teams only do capture.

Should I stop using cost per lead entirely?

No, but stop using it as your only metric. Cost per lead suits the capture half of the budget. It is the wrong scoreboard for demand creation. Only 22% of teams measure closed-won from target accounts (6sense, 2025), the metric that actually matters most.

How much budget should go to brand versus activation in B2B?

Binet and Field's LinkedIn B2B Institute research points to balancing brand and activation, with a long-run split commonly cited near 50/50 (IPA, 2021). Confirm the exact figure in the full report. Newer brands often start activation-heavy, then shift toward brand over time.

Why does dark social matter for paid media attribution?

Because most influence happens where no UTM exists. An 11-person buying group creates 150 to 200 touchpoints, but multi-touch models see only 3 to 4 (6sense, 2025). Slack shares and peer recommendations get logged as Google search on forms, so last-click misreads the trigger.

Can AI run demand generation measurement on its own?

It should not run unattended. AI helps cluster self-reported attribution and surface account-level patterns, but treat it as a human-in-the-loop role. A person reviews the inputs and the outputs before anything ships or gets reported to the board.