Boardroom question: “How do we know what’s working and what’s not?”
That was the stress test in my latest Causal CMO chat with Mark Stouse, CEO at Proof Causal Advisory.
The problem isn’t the question. It’s the math most teams are using to answer it.
If you’re struggling with GTM effectiveness, it’s not because you have a weak dashboard.
You’re struggling because your model is reading from an outdated map.
It’s the reason why leadership hears one story while reality delivers another.
Correlation worked for decades because the world was stable enough. Extrapolate the last four quarters, get a reasonable next quarter. Econometric models, actuarial tables, sales forecasts, marketing mix models. All leaned on the same bet: past was prologue.
Past is no longer prologue. (Was it ever?)
Insurance is the cleanest tell. Several carriers have pulled out of entire states. Not because they’re bad at pricing risk, but because their models can no longer price the risk with confidence. They’d rather walk away from the revenue than write policies they can’t defend.
Swiss Re’s latest catastrophe data shows why that pressure is building. Wildfires, storms, and floods accounted for a record 92% of global insured natural-catastrophe losses in 2025. The underlying conditions shifted. The models are still calibrated to the old ones.
The same pattern is showing up in GTM. In a recent MarTech analysis, Mark reported B2B GTM effectiveness fell from 78% in 2018 to 47% in 2025 across 478 companies. That is not a rounding error. That is a model that no longer fits its market.
“What has been promised is not what’s actually happening. And that’s the dead giveaway.”
Most plans describe the actions on the left and the outcomes on the right. They leave out the middle.
The middle is everything you don’t control: tariffs, rates, inflation, war, competitor moves, buyer budget pressure, category fatigue, shifts in buying committees. The middle accounts for 70 to 80% of what actually drives the outcome.
If you don’t measure it, you can’t explain why the plan half-worked. And you definitely can’t tell the CFO.
Most of that data is available. Financial institutions publish it. Governments publish it. Competitor movement is tractable. You just have to actually include it in the model. Google’s Meridian documentation makes the same point in plainer terms: causal inference estimates effect under real conditions, not correlation in historical data.
Mark put the gap between correlation-guided and causation-guided decisions at 90 to 100 degrees off on a compass rose.
Not off by a few points. Pointed in the wrong direction.
Drift diving: You drop into a current and go neutral buoyancy. The current carries you and it feels great. Then you try to turn around.
What was a tailwind is now a headwind. You’re burning oxygen fast and going nowhere.
GTM spend into a headwind works the same way. The same activity costs more to produce the same result. If you’re not documenting the headwind, it just looks like the team underperformed.
This is why cutting GTM spend during turbulence is usually the wrong reflex. Not the tired “your competitors went quiet, be loud” story. The real reason: staying even in a headwind already costs more. Cutting into that accelerates the decline. If leadership can’t quantify the headwind, they’ll blame execution for a market condition.
Looking inside the four walls of the company instead of outside is an all too common bad habit. Pipeline, velocity, rep activity, campaign throughput. All internal. All controllable. All missing the 70 to 80%.
Teams stay inside because that’s where control lives. It’s comfortable. It’s defensible. You can put it in a deck.
But none of it is reality.
As of 2025, more than half of B2B GTM spend is now ineffective, and it’s not because teams suddenly took stupid pills. They just stopped looking outside. The externalities got louder while the dashboards stayed the same.
The deeper resistance is different. If a causal model shows the old playbook didn’t work, what does that do to my credibility?
When the environment has shifted this much, retrospective blame is a waste of time. Nobody called this environment cleanly with a correlation model. The question is not whether the old playbook was right. The question is whether the current one is.
“Causal AI is not something to be afraid of. Causal AI is reality. Its whole goal is to show you a model, a digital twin of reality, so that you can navigate it more successfully.”
A GPS doesn’t grade your past driving. It reroutes when conditions change. That’s the point.
Use GenAI to generate high-fidelity synthetic data for a strategy you haven’t run yet, then pressure-test it through a causal model.
Use case: Your big agency walks in with a hot proposal to change the game for your business. Deeply insightful. Expensive. Before you put a dime behind it, upload the proposal to a causal model and ask: What is the likelihood this actually works? What would have to be true in the market? Three-year play or twenty-year play?
This kind of tooling leans heavily on one real strength of pattern matching: it’s more reliable at telling you something is a bad idea than a great one.
Worth having before you sign the SOW.
Reality is not a matter of opinion. It’s gravity.
“Reality is what you run into when you’ve made a mistake.”
You can get the signal early by modeling externalities, or you can get it late from a missed quarter. One costs less than the other.
These are the two questions worth writing down:
Two questions. No technology required. They will force the conversation outside the four walls.
You can do the same gut-check on your own buying behavior. Same muscle. Different mirror.
Write down your top three GTM assumptions for the next two quarters. List the external conditions that must hold true for each. Flag the ones that aren’t holding now.
Check the date range on your forecasting data. If it reaches back more than three years, the model is averaging a world that’s gone.
Before the next board update, add a slide on headwinds and tailwinds with a number attached. If you can’t quantify it yet, say so and commit to a date.
Missed the session? Watch it here.
If you like this content, here are some more ways I can help:
Cheers!
This article is AC-A and published on LinkedIn. Join the conversation!
