Demand generation metrics only change behaviour when the replacement number is defined, instrumented and reported before the old one is retired. Replace a lead volume target with four measures: qualified pipeline created, pipeline coverage for next quarter, win rate by originating source, and cost per opportunity. Run both numbers for one quarter, then drop the old one with evidence.
The argument for quality over volume is won on paper and lost in practice, because nobody can present a quality argument to a board in the slide format the volume number used to occupy. We are not going to re-run that argument here. Our piece on lead quality versus quantity in B2B covers it in full. This one is about the mechanism, which is the harder half: how you build the replacement scorecard, how you run both systems through a transition quarter, and how you defend the change to the people who approved the budget on the old number. DevCommX rebuilds demand measurement for B2B revenue teams, so what follows is the sequence we run.
The short answer: you cannot drop the lead number until something replaces it
Every failed attempt looks the same. A marketing leader argues that the lead count is a bad proxy, wins the argument in the room, and then discovers there is no number to put in the weekly deck. Within two months it is back, because a forecast meeting cannot run on an absence. The metric survives not because anyone defends it but because it is the only fully instrumented number marketing owns end to end.
So sequence the work. Build the replacement, instrument it, prove it reconciles with finance, run it beside the old number for one quarter, then retire the old one. Forrester's analysis of the end of MQLs notes that the typical inquiry to closed won conversion rate of a lead centric process is under 1 percent, which is the strongest available argument that the old number was never predictive. It is not, on its own, something anyone can act on next Tuesday.
The deeper reason to change is that the buying unit stopped resembling a lead years ago. Forrester's B2B Revenue Waterfall is built on opportunities and buying groups rather than individual leads, because the object worth counting is the group making the decision, not the one person who filled in the form. A measurement system built on individuals cannot see the thing it is trying to predict, which is why better scoring never fixes it.
Why lead volume survives every argument against it
Four reasons, and only one is inertia. Knowing which is operating in your company tells you what the replacement must provide.
It is the only number with a clean line from spend to output. Budget goes in, leads come out, the ratio is computable weekly. Every replacement metric lags, and a lagging metric is harder to manage a team on. It is compensable. Somebody's variable pay is tied to it, so changing the metric routes the decision through finance and people operations, which is slow enough that it survives another year by default.
It is comparable. The board has seen the same chart for eight quarters and can read the trend without explanation. A new metric resets that history to zero, which reads as a loss of visibility even when it is a gain in accuracy. It is defensible when things go badly. If pipeline misses and the marketing leader hit the lead target, there is a story. That is the real reason it persists, and the replacement must give the leader an equally defensible position or it will not be adopted. Forrester documents three objections of its own, measurement anxiety, CRM tooling risk and implementation complexity, and only the first is the same objection as ours.
None of these is an argument that the lead count is accurate. They are arguments about operability, which is useful: it makes the replacement an engineering problem rather than a persuasion problem. The signal based framing underneath it is in why MQLs are dead and what signal based alignment replaces them with.
The four demand generation metrics that actually predict pipeline
Pick pipeline metrics that a rep's behaviour cannot manufacture and a finance team can reconcile. These four do both, and between them cover output, forward risk, quality and efficiency.
1. Qualified pipeline created. The value of opportunities created in the period where the primary source is a marketing motion, counted at creation and held to the creation quarter even when the deal slips. This is the closest thing to a like for like replacement for the old target, and it is the number a board will anchor on.
2. Pipeline coverage for next quarter. Open pipeline for the coming quarter divided by that quarter's target, split by source. Coverage is the earliest honest warning in the set. A lead number can stay green all quarter while coverage quietly falls below the ratio your historical win rate requires, and nobody notices until the quarter has already been decided.
3. Win rate by originating source. Closed won divided by closed total, segmented by originating motion, measured only on cohorts old enough to have closed. This is what exposes a channel that produces volume and no revenue, and it is the metric that turns the quality argument into arithmetic. Calibrating the fit model that feeds it is covered in win rate calibrated ICP scoring.
4. Cost per opportunity, not cost per lead. Fully loaded marketing spend divided by qualified opportunities created. Moving the denominator to the object sales actually works is what makes a finance team engage with marketing efficiency. Two of these four lag by construction, which is why the scorecard also carries leading indicators. Account engagement depth, buying group coverage and intent signal density all move earlier than pipeline, and our guide to intent data and buying signals in B2B outbound covers where they come from. Carry them as context, not as the target.
Building the demand generation metrics scorecard
A demand gen scorecard has five columns and no more: metric, definition, owner, cadence, and the threshold that triggers a conversation. If a metric cannot be given a threshold, it is context rather than a scorecard line, and it belongs in an appendix where it will not dilute the four numbers that matter.
Note the last row. Keep the old number visible and explicitly untargeted through the transition. That single line removes most of the political risk, because nobody can accuse you of hiding a decline behind a definition change.
Write the definition column in full sentences and store it where both teams can read it. Ambiguity there turns a scorecard back into an argument six weeks later. The same discipline applies to the qualification boundary underneath it, covered in MQL versus SQL lead qualification, and the ownership model for maintaining each definition sits with revenue operations.
The transition quarter: running both numbers at once
One quarter. Both numbers. Neither target changed mid flight. The purpose is not fairness to the old metric. It is to build the reconciliation evidence you will need in the board meeting at the end of it.
Weeks one to four. Publish both, with the new definitions written out in full, then reconcile the new pipeline number against the CRM report finance already trusts. If marketing's figure and finance's figure differ by more than a rounding error, stop and fix that first. A metric that does not reconcile will be dismissed the first time it delivers bad news.
Weeks five to eight. Run the correlation. Take the last four to eight quarters, plot the old lead number against pipeline created in the following quarter, then do the same for each new metric. You are looking for a defensible result, not a dramatic one, and whatever your data says is what you present. If the lead number does correlate in your business, say so and adjust the proposal.
Weeks nine to twelve. Change the compensation plan for the following quarter and give the affected people the arithmetic in writing before they hear it in a meeting. Replacing MQL targets fails at this step more often than at any other, because the reported metric changes and the incentive does not, so behaviour follows the money and nothing moves. The distinction between a signal worth working and evidence worth opening an opportunity on is drawn in signal based selling versus intent data, and you will need it when the comp conversation starts.
How to present the change to a board that funded the old number
Boards do not object to better metrics. They object to losing a time series and to being asked to trust a number they cannot audit. Address both directly, in that order.
Open with continuity rather than with the critique. Show the old number for the transition quarter beside the new one and demonstrate that they tell a consistent story. Then show the correlation work, then propose the change. In our experience, leading with the case against the old metric makes the room feel sold to. Separately, Harvard Business Review's guidance on making the business case for a marketing budget holds that the case that lands is framed in the finance team's terms rather than marketing's.
Give them a metric they can audit. Qualified pipeline created and cost per opportunity both trace to CRM records and ledger entries a board member can ask the CFO to verify. Auditability is worth more in the room than any argument about predictive power.
Then name what you are giving up. You are trading a fast, weekly, precise number for a slower, lagging, more accurate one, and you will not have a clean read for the first six to eight weeks of a quarter, which is exactly why coverage is on the scorecard. Saying this before somebody else notices it is what buys you the change. It also helps to state the buyer side reason: Gartner predicted in 2020 that by 2025, 80 percent of B2B sales interactions between suppliers and buyers would occur in digital channels, so a measurement system anchored on a single form fill was always going to describe less and less of how buying actually happens.
What breaks in the first 60 days, and what to do about it
Reporting volume collapses. The new numbers are monthly or quarterly by construction, so the weekly deck looks empty. Fix: put the leading indicators on the weekly cadence and keep the lagging ones monthly, which is what the cadence column is for.
Attribution arguments restart. The moment pipeline carries marketing's name, every source assignment becomes contestable. Fix: freeze the primary source field at opportunity creation, make it read only afterwards, and audit the logic monthly instead of debating individual deals.
Campaign selection drifts to safe accounts. If the number is opportunities, the fastest route to opportunities runs through accounts already in cycle, and pipeline quality quietly narrows to the installed base. Fix: carry a new logo split on the scorecard so the number cannot be hit from existing relationships alone.
Sales disputes the opportunity definition. Now that marketing's number depends on opportunity creation, the creation bar matters to both teams for the first time. This is the argument worth having, and better in month two than in the quarter you miss. Settle it in writing, in the same document as the scorecard definitions, and give one person authority to arbitrate. The operating model for that sits in our guide to revenue operations consulting for mid market teams.
Build Your Demand Gen Scorecard With DevCommX
One check decides whether any of this survives its first bad quarter: does qualified pipeline created, as marketing reports it, match the CRM figure your finance team already trusts? If the two disagree by more than a rounding error, the scorecard will be dismissed the first time it carries bad news, and no amount of correlation work will rescue it afterwards. So that reconciliation is where our engagements start, not where they finish. From there we write the four definitions out in full sentences, instrument them, set the thresholds that trigger a conversation, and build the evidence you will need in the board meeting at the end of the transition quarter. Our revenue operations practice carries the build; the weekly cadence stays with your team. Our own benchmark on the demand side is 40+ qualified demos in ~6 weeks, from our AI SDR work on a fully scoped programme with a defined ICP. Send us the deck you present today and we will tell you which line fails reconciliation.
References
- Forrester, The Revenue Process Alignment Series Part 1: The End Of MQLs, source for the finding that inquiry to closed won conversion in a lead centric process runs under 1 percent
- Forrester, Transform Your Demand Process: The B2B Revenue Waterfall Guide, source for the opportunity and buying group centric model that replaces the lead centric waterfall
- Forrester, Three Objections To Moving From MQLs To Opportunities, source for its three objections, measurement anxiety, CRM tooling risk and implementation complexity
- Harvard Business Review 2021, Making the Business Case for Your Marketing Budget, source for framing the case in the finance team's own terms
- Gartner press release, 15 September 2020, Gartner Says 80% of B2B Sales Interactions Between Suppliers and Buyers Will Occur in Digital Channels by 2025, source for the prediction that by 2025, 80 percent of B2B sales interactions between suppliers and buyers would occur in digital channels
FAQ
What demand generation metrics should marketing be measured on?
Measure demand generation on qualified pipeline created, pipeline coverage for the next quarter, win rate by originating source, and cost per opportunity. Together they cover output, forward risk, quality and efficiency, and all four reconcile to records a finance team can audit. Keep leading indicators such as buying group coverage on the scorecard, but do not set the target on them.
Should we still track MQLs?
Yes, as a diagnostic rather than a target. An MQL count still spots a broken form or a sudden channel change within a week, which the lagging metrics cannot do. What changes is that nobody is compensated on it and no forecast conversation is built on it. Report it, leave it explicitly untargeted, and review it monthly alongside the rest.
How do I change what marketing is measured on?
Build the replacement before retiring the old number. Define the new metrics in writing, instrument them, reconcile them against the pipeline report finance already trusts, then run both systems side by side for one quarter. Present the correlation work to leadership, change the compensation plan for the following quarter, and only then stop targeting the old number.
How long should the transition quarter be?
One quarter is usually right. Any shorter and you have no reconciliation evidence and no correlation to present. Any longer and the organisation treats the new numbers as a side report rather than the real ones. Publish both from week one, close reconciliation gaps by week four, and change the compensation plan before the next quarter opens.
What goes on a demand gen scorecard?
Five columns and no more: the metric, its definition written as a full sentence, a named owner, a reporting cadence, and the threshold that triggers a conversation. Anything that cannot be given a threshold is context rather than a scorecard line. Include the old lead number for the transition quarter, clearly marked as reported but not targeted.
How do you present a metric change to a board?
Lead with continuity rather than with the critique. Show the old and the new numbers side by side for the transition quarter, show the correlation work behind the new ones, then propose the change. Offer metrics the board can audit through CRM and ledger records, and state plainly what you are giving up, which is weekly precision in exchange for accuracy.
.webp)
