Demand Generation Audit Worksheet

Why this frame matters in 2026

The create/capture/convert frame matters more now because the buyer moved out of your funnel. G2’s March 2026 survey of 1,076 B2B software buyers found 51% now start research with an AI chatbot rather than a search engine, 71% rely on chatbots at some point, and a third bought from a vendor a chatbot surfaced that they’d never heard of (Demand Gen Report, May 2026). Gartner’s 2026 survey of nearly 650 B2B buyers found 67% prefer a rep-free experience and 45% used AI during a recent purchase (Demand Gen Report, March 2026). Meanwhile the parts of the journey you can still see are getting more expensive: Gartner’s 2025 CMO Spend Survey pegged marketing budgets flat at 7.7% of company revenue with paid media consuming 30.6% of that (Demand Gen Report, June 2025), and ChatGPT-referred visits to B2B sites grew 303% year over year to 2.6 million a month by June 2026 (Demand Gen Report, August 2026). Buying groups average nine members and buying cycles have compressed to seven months (INFUSE/G2, via Demand Gen Report, July 2026). Score the three stages separately, calibrate against the benchmarks below, and leave with owners.

Scoring rubric (use for all three scorecards)

Evidence rule: no evidence, no score above 2. Attach a dashboard link, URL, or call note for every score. An assertion without evidence is a hope, and hopes don’t get funded.

1. Scope

Fill this before scoring anything. An audit without a scope produces scores that apply to no one.

Context: with buying groups now averaging nine members and cycles at seven months (INFUSE/G2, 2026), a scope built around a single contact and a single quarter will miss the machine you’re trying to fix. Score committees and multi-quarter windows.

2. Create scorecard (1–5)

Demand creation is the work that builds intent before it can be captured. Score each item, add one line of evidence, and mark the owner if the item is broken.

Scoring context:

• AI-answer presence is now a creation output, not a nice-to-have. 51% of buyers start research in AI chatbots, and 85% think more highly of a vendor a chatbot recommends (G2, 2026). Run your top 20 buying questions through two or three answer engines and note where you’re cited, where competitors are, and where nobody is.
• Distribution beats production. If assets ship without an owner for the path to the audience, the create stage is 2 regardless of how good the content is. Freeze new production and fix the paths first.
• Gartner 2025 CMO Spend context: budgets are flat (7.7% of revenue) while paid media eats 30.6% of budget. That’s the budget reality creation must justify — measure it in influenced pipeline, not output volume.

3. Capture scorecard (1–5)

Demand capture converts existing intent into qualified pipeline. Score each item with evidence.

Benchmarks to score against (labeled with sources):

• Pipeline coverage: 3–4x qualified pipeline to target is the common baseline, but it is a starting point, not a standard — required coverage is 1 ÷ win rate (25% win rate needs 4x; 20% needs 5x). Enterprise motions with longer cycles may need 4x–7x (Clari, June 2026).
• MQL-to-SQL conversion: 20–30% is a solid benchmark; SQL-to-opportunity: 40–60% (Guideflow, June 2026). If sales rejects more than half of what you send, the capture stage is producing volume, not qualification.
• Form conversion isn’t the whole funnel. Buyers research in chatbots and build shortlists before they fill forms (G2, 2026; TechTarget 2025 survey cited by Miniloop, June 2026). Score capture on qualified pipeline per program, not on lead count.
• Signal-aligned capture outperforms activation alone: programs aligned with buyer intent signals deliver 21% more leads per account and are 1.6x more likely to reach the decision stage (INFUSE/G2, 2026). If your capture programs ignore intent signals, cap the score at 3.
• CPL honesty check: CPL matters less than influenced pipeline. A $200 CPL program that produces real pipeline beats a $30 CPL program that produces nothing downstream (Guideflow, June 2026). Flag any program reporting CPL without pipeline influence.

4. Convert scorecard (1–5)

Conversion turns captured demand into sales-accepted pipeline and closed-won. Score with a sales counterpart in the room — this stage cannot be scored by marketing alone.

Benchmarks to score against (labeled with sources):

• Speed-to-lead is a conversion lever with hard data. In the HBR lead response study (1.25 million leads, 29 B2C and 13 B2B companies), firms that tried to contact a lead within an hour were nearly seven times as likely to qualify it as those that waited longer. The average response time across the audited companies was 42 hours, and 23% never responded at all (HBR, 2011). If your average response is measured in days, this item scores 2 — and fixing it is usually cheaper than buying more leads.
• Alignment pays: tightly aligned sales and marketing organizations show 38% higher win rates and 36% higher customer retention (MarketReach research, cited by Miniloop, June 2026). Misalignment alone adds roughly 25% to sales cycle length (Guideflow, June 2026).
• Score committees, not contacts. With buying groups averaging nine members (INFUSE/G2, 2026), a conversion system that routes one contact and calls it done is structurally incomplete. Check how many stakeholders per account are engaged before scoring above 3.

5. Decision table

For every program in scope, one row. “Fix” is capped at what you can staff — an unfunded fix is a pause with nicer wording. Mark the owner of each decision.

Rules:

• A program with no evidence of pipeline influence after two quarters goes in Kill, not Pause.
• A program with rising CAC and flat branded search is a creation problem, not a capture problem — decide accordingly.
• Force-rank pauses and kills. If everything is “strategic,” nothing is, and the table has no information.

6. 90-day plan (max 5 workstreams)

Each workstream gets an owner, a weekly metric, a definition of done, and a 90-day target. The plan is a betting slip: state what you’re betting on, what you stopped, and what evidence would make you abandon the bet.

Constraints:

• Max five workstreams. More than five means you haven’t prioritized; cut until the list fits on one page.
• One owner each. An ownerless workstream is a wish.
• Weekly metric, not quarterly. If you can’t name what you’d look at every Friday, the workstream is too vague to run.
• Kill at least one thing. Every audit cycle ends with at least one pause or kill. It’s how the portfolio stays honest.

7. How to run

• Format: one 90-minute scoring session with a sales counterpart present, followed by a 45-minute planning session that turns scores into the decision table and the 90-day plan. Attach evidence links as you score; never score from memory.
• Who attends: the demand or growth owner (facilitator), one RevOps counterpart (data), one sales rep (conversion and acceptance reality). A founder or CMO attends the planning session only — scoring without evidence pressure produces inflated numbers.
• Cadence: weekly 45-minute pipeline stand-up (stuck deals, SLA breaches, intent spikes, fastest-moving accounts); monthly review of MQL volume, QLVR, CAC, and win rate by source — anything that moved more than 10% gets a causal explanation; quarterly portfolio decisions on the table above.
• Failure mode to avoid: the audit becomes a 60-slide report with no owners and no kill list. The deliverable is the decision table and the five-row plan. If a finding doesn’t produce a fund, fix, pause, or kill decision, it doesn’t belong in the document.

For the full field process behind this worksheet, run the 90-Day Demand Generation Audit, and use Pipeline Metrics That Matter to settle which metrics earn a place in the weekly stand-up.