Not because they lack data. Not because they lack tools.
Because they are running campaigns the way a chef cooks without a recipe — adding ingredients by instinct, tasting at the end, and hoping the result lands somewhere near what the guest ordered.
The gap between what marketing *does* and what marketing *predicts* is one of the most quietly expensive problems in growth. It is the reason CMOs get replaced after two quarters, the reason sales and marketing fight in every QBR, and the reason “we need more budget” is met with raised eyebrows instead of a signed PO.
A marketing forecast does not just tell you what is coming. It exposes what you actually believe about your own funnel — and forces you to defend it with numbers.
How Drift turned forecasting into a competitive weapon
Around 2018, Drift was a mid-sized conversational marketing platform competing in a crowded category against much larger players. They were growing, but chaotically. Marketing was generating leads. Sales was closing deals. Nobody could reliably connect the two in advance.
Dave Gerhardt, then VP of Marketing, started treating the forecast not as a finance artifact but as a marketing instrument. The team mapped every content channel, every demand source, every conversion step — and assigned a rate to each. Blog traffic converting to demo requests. Demo requests converting to qualified pipeline. Pipeline converting to closed revenue. They built what amounted to a living funnel model, updated weekly, reviewed in every marketing stand-up.
The result was not magic.
Drift did not suddenly 10x overnight.
What changed was something more durable: the team stopped arguing about whether a campaign was working and started arguing about which assumption in the model was wrong. That is a completely different conversation. One is emotional. The other is fixable.
By the time Drift was acquired by Salesloft in 2021 for a reported 850 million dollars, their marketing operation had become known for one thing beyond the product itself — predictability. They knew what their inputs would produce before they produced them. That is the real compounding advantage a forecast gives you.
What a marketing forecast actually is — and why most teams build it wrong
A marketing forecast is an estimate of future marketing outcomes — leads, pipeline, revenue — built on historical conversion data and explicit assumptions about activity. That sentence sounds simple. It is not practiced simply.
Most teams confuse a forecast with a goal. A goal is what you want. A forecast is what the math says will happen if you execute your planned activity at your historical conversion rates. These two numbers should sit next to each other in every planning document. The gap between them is your problem statement.
The mechanism underneath forecasting
The mechanical logic is straightforward. You start with planned inputs — ad spend, content volume, email sends, event registrations, whatever your demand channels are. You apply known conversion rates at each stage of the funnel. You arrive at an expected output: MQLs, SQLs, pipeline value, closed revenue. The model is only as good as the conversion assumptions you feed it, which is why the quality of your historical data is not a reporting problem — it is a forecasting problem.
As Lenny Rachitsky, product and growth advisor to dozens of high-growth companies, has put it: “The best growth teams treat their funnel like a system. They know their inputs, they know their rates, and when output changes, they go looking for which rate moved.” That discipline — looking for which rate moved — is the entire job of a growth marketer in a forecast-driven team.
Where most marketers go wrong
The first mistake is building the forecast backwards from the revenue target rather than forwards from planned activity. When you start with “we need 2 million in pipeline” and reverse-engineer the inputs required to hit it, you are writing a wish, not a forecast. You are choosing a number and then finding justification for it. This is the origin of every sandbagged target and every embarrassing miss.
The second mistake is treating conversion rates as fixed. They are not. Rates change with channel saturation, with message fatigue, with competitive intensity, with product-market fit shifts. A forecast model that uses last year’s rates without questioning them is a model that will mislead you in exactly the moment you most need clarity.
The third mistake — arguably the most damaging — is building the forecast once at the start of a quarter and leaving…
That’s it!
Comments