Here is the thing nobody says out loud in the QBR (Quarterly Business Review): most marketing teams spent the first six months of this year optimizing metrics that stopped predicting revenue somewhere around 2023. Most scorecards are still built for how buyers behaved in 2019, tracking session volume and MQLs (Marketing Qualified Leads) as if they still predicted revenue, when the real work now happens in six places most scorecards never look: market coverage, pipeline health, velocity, spend efficiency by deal band, AI search visibility, and whether the data underneath any of it can be trusted.
Not because anyone is lazy, but because the dashboard still loads, the numbers still move, and green is still green, and a broken measurement system does not throw an error, it just quietly keeps reporting while everyone keeps nodding.
So before you build an H2 (second half) plan on top of your H1 (first half) reporting, let’s look at what the first half of 2026 actually told us. That’s the real story in two numbers: according to 6sense’s 2025 Buyer Experience Report, 94% of buying groups had already ranked a preferred vendor before talking to a seller, and bought from that favorite 77% of the time, meaning buying moved off your property. And per Gartner’s 2026 CMO Spend Survey, martech’s share of the marketing budget fell from 22.4% to 19.4% in a single year, a five-year low, meaning the systems you would use to see that shift got defunded in the same window.
Key Takeaways
- A B2B marketing scorecard needs six categories, not a pile of accumulated metrics: market coverage, funnel and pipeline health, velocity, spend efficiency by deal band, AI search visibility, and data and systems health.
- According to 6sense’s 2025 Buyer Experience Report, 94% of buying groups rank a preferred vendor before talking to a seller, and buy from that vendor 77% of the time.
- SparkToro’s analysis of Similarweb clickstream data found 68% of Google searches ended without a click in the first four months of 2026, which means organic session count is no longer a reliable demand proxy.
- Pipeline coverage targets should be calculated as 1 divided by your actual win rate, not copied from a conference-stage 3x rule. Ebsta and Pavilion’s 2025 GTM (Go-to-Market) Benchmarks, based on 655,000 opportunities, put the average B2B win rate at 19%, down from 29% the year before.
- Six commonly used metrics, including MQL volume, last-touch attribution, organic sessions, and blended Customer Acquisition Cost (CAC), should be demoted from headline metrics to diagnostics because they now point the wrong direction on their own.
- Data and systems health should gate every other category on the scorecard. According to MarTech’s 2025 State of Your Stack Survey, 65.7% of B2B marketers name data integration as their primary obstacle to measurement.
- Teams have roughly 90 days before Q4 planning to rebuild the scorecard: lock stage criteria, add self-reported attribution, rebuild CAC by deal band, and ship one scorecard per category with a named owner.
What did H1 2026 reveal about B2B buying and measurement?
Four numbers. They all point the same direction.
Buyers do 60% of the work without you. 6sense’s 2025 Buyer Experience Report put the independent-research-to-seller-engagement split at 60/40, up from 70/30 the year before. Sounds like progress. Except the same research found that 94% of buying groups had already ranked a preferred vendor before they talked to a single seller. And they bought from that preliminary favorite 77% of the time.
Read that twice. The race is mostly decided before you know you entered it.
68% of Google searches ended without a click. That is SparkToro’s analysis of Similarweb clickstream data, January through April of this year. Your session count stopped being a proxy for demand. It is now a proxy for click capture, and those are two different businesses. Ahrefs’s analysis of 300,000 keywords, updated with December 2025 data, found a 58% drop in position-one click-through where AI Overviews appear. You can hold your rankings and still lose the traffic.
67% of B2B buyers prefer a rep-free experience. Gartner surveyed 646 buyers and published in March 2026. Up from 61%. That trend line is not subtle and it is not reversing.
Half of martech leaders say their stack cannot do what they are being asked to do. Gartner’s 2025 martech leader survey of 413 leaders, published in October 2025: 50% report their organization lacks the technical and data stack readiness for AI agent deployment. Meanwhile, martech fell to 19.4% of the marketing budget, a five-year low, down from 26.6% in 2021, per Gartner’s 2026 CMO Spend Survey of 401 CMOs.
Put those together and you get the real story of H1 2026: buying moved off your property, and the systems you would use to see it got defunded.
That is the environment you are building an H2 plan inside of. So the question is not whether your numbers went up. The question is whether the numbers you are watching still describe the business.
The Six-Question Scorecard: a framework for auditing B2B marketing measurement
The Six-Question Scorecard is a framework built around six questions a marketing scorecard has to answer, gated by whether you can trust the data behind them. None of the six substitutes for another. I built it for a client engagement earlier this year, and I have used it several times since, because the problem is universal: most scorecards are a pile of metrics that accumulated over time.
1. Market coverage and demand creation
Are we present with the buyers who are not in market yet? For long-cycle categories, roughly 20% of buyers are in-market annually and about 5% in any given quarter. That is a heuristic from the Ehrenberg-Bass Institute’s research, not a measurement, and I will not pretend otherwise. But the structural point holds: if most of your market is not buying this quarter, in-quarter conversion is the wrong lens for most of your audience. Track coverage, share of search, referral share, and marketing investment rate. High Growth professional services firms spend 12% of revenue on marketing, according to Hinge Research Institute’s 2026 High Growth Study of nearly 500 firms. No Growth firms spend 5%. That gap is not a coincidence.
2. Funnel and pipeline health
Is the pipeline real, and will it cover the number? Stage conversion, win rate, coverage ratio, buying group penetration. One correction while I have your attention: your pipeline coverage target is not 3x because somebody said 3x. It is 1 divided by your win rate. At a 25% win rate you need 4x. Ebsta and Pavilion’s 2025 GTM Benchmarks, across 655,000 opportunities, put the average win rate at 19% last year, down ten points. Do that math on your own number, not the conference-stage number.
3. Velocity and conversion efficiency
How fast does work move, and where does it stall? Cycle length, stage aging, touchpoints to close. Above $100K in deal size, HockeyStack Labs measured roughly 417 touchpoints and 5,500 impressions per closed deal. If your attribution model credits one of those, the model is not a model. It is a coin flip with a logo on it.
4. Spend efficiency and return
What does an opportunity actually cost, and in which deal band? This is where most teams do the most damage, and I will come back to it. As one directional data point, Benchmarkit’s 2025 SaaS benchmarks found CAC payback stretching to 18 to 24 months for deals with annual contract values above $100K, against under 12 months for the smallest deal bands. Blend that into one number and the gap disappears from view entirely.
5. AI search and discovery visibility
Do we exist inside the answers buyers now read? Citation rate, AI share of voice by topic, AI referral traffic, zero-click exposure. This category is new and it is going to get reported badly for the next two years, because the growth rates are spectacular and the base is tiny. AI referrals are still under 1% of web traffic even though ChatGPT’s outbound referral traffic grew 206% year over year, according to Semrush’s clickstream analysis. Both facts are true. Report them together or do not report them.
The most important row in this category is the one that says not currently measurable. Prompt volume, why a source got cited, and buyer research happening inside private corporate AI tools are all structurally invisible to your analytics. Put that on the scorecard as a stated blind spot. A blind spot you have named is a risk. A blind spot you have not named is a surprise.
6. Data and systems health
Can we trust any of the numbers above? Data integration completeness, record completeness, self-reported attribution quality, AI governance. 65.7% of B2B marketers name data integration as the primary obstacle to measurement, per MarTech’s 2025 State of Your Stack Survey. Two thirds of the market is telling you where the constraint is.
Here is why the order matters. Category 6 gates the credibility of everything above it. Instrument category 5 on top of an unjoined data layer and you will produce a beautiful number that nobody can defend when a CFO (Chief Financial Officer) asks one follow-up question.
Comfort metrics: six B2B marketing KPIs that are lying to you in 2026
I call these comfort metrics. They feel like measurement. They move in ways that feel meaningful. And in 2026 every one of them points the wrong direction.
1. MQL volume as a health signal. In HockeyStack’s H1 2025 dataset, MQL volume fell 2.4% quarter over quarter while closed-won rose 4.4%. Volume and outcome have decoupled. If your team is still celebrating a lead-volume record, they are celebrating a coincidence.
2. Last-touch attribution. Against self-reported data, last touch roughly agrees on search at 53% and social at 60%. On email it agrees 21% of the time. On display, 2.5%. It does not distribute credit evenly wrong. It systematically underprices exactly the channels that build preference before contact.
3. Organic sessions as a demand proxy. With 68% of searches ending without a click, per SparkToro’s analysis of Similarweb data, category demand can rise while your session count falls. You can be winning and watching a red arrow.
4. Blended CAC and blended cycle length. Reporting a blended CAC is like reporting the average temperature of a hospital. Technically accurate, clinically useless. In Benchmarkit’s 2025 SaaS benchmarks, CAC payback ran under 12 months in the smallest deal band and 18 to 24 months in the largest. A blended number tracks your mix shift and calls it performance.
5. Imported SaaS ratios. The 3:1 LTV to CAC rule and standard payback targets came out of SaaS. Neither has services validation, and project-based revenue breaks the assumption underneath both. If you are a services firm managing to a SaaS ratio, you are managing to somebody else’s business model.
6. Marketing-sourced pipeline as the primary scorecard. Sourced share shrinks as deal size grows. Per Prooflytics’ 2026 B2B SaaS benchmarks, marketing-sourced share drops from roughly 59% to 47% once average deal size crosses $50K, and healthy teams typically run 30-50% sourced against 70-90% influenced, two different numbers that answer two different questions. Manage to sourced alone and you will systematically defund the programs that build preference before anyone fills out a form. Report sourced and influenced separately. Never blend them into one.
None of these should be deleted. They should be demoted, from headline to diagnostic, and reported next to the metric that corrects them.
How to prepare your B2B marketing scorecard for Q4 2026 planning
Before Q4 2026 planning, do four things in this order: lock stage criteria and audit your CRM, instrument attribution and segment AI referrers, rebuild CAC and cycle length by deal band, and ship one scorecard per category with a named owner. You have roughly 90 days before the H2 numbers become the story you tell for the year. Spend them like this.
Days 1 to 15. Lock written stage entry and exit criteria and get sales to sign them. Audit CRM record completeness against the fields your routing and scoring actually require. Measure buying group coverage on open deals.
Days 16 to 30. Add self-reported attribution to every form, and plan to throw out about 20% of the responses as unusable, per HockeyStack Labs’ self-reported attribution research. Segment LLM (large language model) referrers out of organic. Tag every opportunity with a deal band and service line at creation, not at close.
Days 31 to 60. Rebuild CAC, cost per lead, conversion, and cycle length by deal band and service line. Split sourced from influenced. Establish your own prompt set for AI visibility baselining.
Days 61 to 90. Ship one scorecard per category with a named owner, a refresh cadence, and the known limitation printed next to each metric.
That last part is the one people skip, and it is the one that earns you the room. Printing the limitation next to the number is not a weakness. It is the difference between a marketer defending a dashboard and an operator explaining a business.
The part that actually matters: your scorecard decides what gets funded
Your measurement system is not neutral. It decides what gets funded.
Every quarter you run a scorecard built for a buying process that no longer exists, you are not just reporting badly. You are actively reallocating budget toward the things that are easiest to count and away from the things that are actually moving preference. Gartner’s 2026 CMO Spend Survey of 401 CMOs shows exactly that happening at market scale: awareness and conversion now absorb 62.6% of media spend while retention has dropped below 15%, down 29% since 2024. The market is not choosing short-term. The market is choosing measurable, and then calling it strategy.
Optimization is not strategy. A green dashboard is not a healthy business.
Go look at your scorecard. If you cannot say what question each metric answers and what its limitation is, you do not have a measurement system. You have a habit.
Frequently Asked Questions
Who should own a B2B marketing scorecard?
Each of the six categories should have one named owner, not a shared owner across the whole scorecard. A single person accountable for all six tends to defend whichever numbers are easiest to explain, while category-level ownership, for example a demand-gen lead on market coverage and a RevOps lead on data and systems health, keeps each metric tied to the person who can actually act on it and forces the known limitation to get printed next to the number instead of glossed over.
How is a marketing scorecard different from a marketing dashboard?
A dashboard displays whatever metrics were easiest to pull into it; a scorecard is designed backward from six specific questions, and every number on it exists because it answers one of them. A dashboard grows by accretion, someone bolts on a new chart whenever a stakeholder asks for one, until nobody remembers why half the tiles are there. A scorecard has a fixed structure: six categories, one owner per category, and a stated limitation printed next to every metric.
How often should a B2B marketing scorecard be refreshed?
Each category needs its own refresh cadence rather than one blanket schedule for the whole scorecard. Fast-moving categories like velocity and spend efficiency hold up fine on a monthly refresh, while market coverage and AI search visibility move slowly enough that monthly updates mostly just add noise, and a quarterly refresh tied to the planning cycle is more honest about how fast those numbers actually change.
How do you get sales buy-in on a new marketing measurement system?
Start with the metric sales already distrusts, usually marketing-sourced pipeline, and let them help define what counts as a qualifying touch before changing anything else. Buy-in comes from co-writing stage entry and exit criteria together in the first two weeks, not from presenting a finished scorecard and asking for a sign-off. Sales will defend a system they helped define and pick apart one that just showed up in their inbox.
Should marketing scorecard results go to the CMO or the CFO?
Both, but with different framing. The CMO needs the full six-category breakdown to run the function, while the CFO needs the spend-efficiency and data-trust categories translated into cost per opportunity and payback period, the language finance already uses. Sending the CMO version to the CFO is a common way a defensible scorecard gets dismissed as marketing jargon.
What is the biggest mistake teams make when building a marketing scorecard?
Building one category at a time instead of gating everything through data and systems health first. Teams that start with market coverage or AI search visibility produce numbers that look sophisticated but sit on top of an unjoined data layer, and the first hard question from a CFO exposes the whole thing. Fix the measurement foundation before adding categories on top of it, not after.
