Your Marketing Is Responsible for Less Revenue Than You Think
Table of Contents
The Uncomfortable Baseline Question
When a national specialty retailer with over 300 stores commissioned a marketing mix model across 2.5 years of data, they expected the model to tell them which channels were working. TV versus radio. Digital versus out-of-home. Performance versus brand.
They got that. But the first number out of the model was the one no one had asked for, and the one that changed how they thought about the whole exercise.
93% of their revenue would have happened without any marketing at all.
Not “most of it.” Not “a lot of it.” Ninety-three percent.
What Baseline Actually Means
In marketing mix modelling, the baseline represents everything that drives revenue that isn’t a current marketing action. It includes brand equity built over years, organic search and word-of-mouth, repeat customer behaviour, and general category demand driven by factors outside your control like seasonality, economic conditions, and competitor activity.
For this retailer, the baseline was exceptionally strong. They operate in a category with genuine organic demand, a loyal customer base, and stores in markets where they have strong brand recognition. That baseline is a competitive asset.
But it also means that the 7.5% of revenue directly attributable to marketing initiatives is what your media budget is actually buying. And when you reframe it that way, the ROI calculation changes.
The $11 Million Reframe
Across 2.5 years, the model attributed approximately $11 million in revenue directly to marketing activity. Against an annual spend of around $3 million, that works out to roughly $3.90 in revenue for every $1 spent across all channels.
That’s a solid return. But the more useful question isn’t “what’s our blended ROI?” It’s: which specific dollars within that $3M are generating the return, and which are subsidising channels that aren’t pulling their weight?
Because the blended $3.90 hides significant variation. Some channels return $6+ per dollar at current spend levels. Others return $1 for every $1 spent, which, after margins and agency fees, is running at a loss.
When Exciting Channels Disappoint
One of the clearest examples from the model was broadcast video on demand (BVOD) - streaming ad placements delivered across connected TV platforms. It’s a channel that attracts significant attention in media planning conversations because it combines the visual impact of TV with digital targeting.
The model showed it returning approximately $1 in revenue per $1 spent.
That sounds neutral. But revenue ROI isn’t the same as profit ROI. When you subtract margins and media fees from a 1:1 revenue return, you’re spending money to lose money. The channel isn’t contributing, it’s consuming.
Contrast this with radio, a channel with a significantly longer history and less novelty appeal. The model showed radio returning $2+ per dollar spent, with strong confidence in the result.
The gap between perception and performance is exactly what measurement is supposed to surface. BVOD gets attention in media meetings. Radio gets taken for granted. The data disagreed with both assumptions.
The Problem With Correlated Spend
One of the more challenging findings was around TV. The model flagged high confidence in most channels, but noted that the TV result carried more uncertainty than usual.
The reason: TV spend was highly correlated with peak promotional periods. The retailer’s biggest annual promotion ran every September and October, the same windows when TV activity was heaviest. When two variables move together consistently, it’s difficult for a model to separate which one is responsible for the revenue spike.
The model’s best estimate was that TV was a strong performer, but the correlation introduced enough noise that the team decided to re-examine the data and test TV in an off-season period to get a cleaner read.
That decision, to wait for cleaner data rather than act on a result with known limitations, is exactly the right response to confidence intervals. The model told them what it knew and what it didn’t. Acting on the “didn’t” category without additional validation is how media budgets go wrong.
What the Baseline Number Actually Tells You
Going back to the 93% figure: a high baseline isn’t a problem. It’s an asset. It means the business has genuine momentum, real brand equity, and a customer base that comes back.
What it changes is how you approach marketing investment.
If 93% of revenue is baseline, your marketing budget isn’t responsible for keeping the business alive. It’s responsible for incremental growth on top of a healthy base. That’s a fundamentally different brief. It means you can afford to be more selective about where you invest the 7.5%, because the stakes of getting it slightly wrong are lower than they would be for a business where marketing accounts for 40% of revenue.
It also means that protecting your baseline, through brand investment, customer experience, and category presence, may be more valuable than optimising your performance channels. A retailer that loses its organic demand foundation is in a far worse position than one whose BVOD ROI isn’t quite where it should be.
The Optimisation That Follows
With the channel-level data in hand, the next step for this retailer was saturation curves: modelling how ROI changes as spend in each channel increases. The goal was to find where existing budget could be reallocated to improve returns without increasing total spend.
The targets were clear: channels running at or below 1:1 revenue ROI were candidates for reduction. Channels with high returns at low spend levels were candidates for investment, with the expectation that marginal returns would decline as spend increased.
The rough market condition at the time was tight. No budget increases were on the table. Which meant the opportunity wasn’t to spend more - it was to get more from the same $3M by moving money from channels that had stopped earning it to those that still had room to grow.
That’s the value of measuring your baseline. Not to diminish what marketing does, but to understand precisely what it does, at what cost, and where the next dollar should go.
Key Takeaways
- A strong revenue baseline is an asset, not an excuse to cut marketing spend
- Marketing-driven revenue for this national retailer was 7.5% of total, worth $11M over 2.5 years at $3.90 per $1 spent
- Channel ROI varied significantly: radio outperformed BVOD by 2x or more
- Channels with revenue ROI at or below 1:1 are likely running at a loss once margins are applied
- Highly correlated spend makes channel attribution difficult - consider out-of-season testing to isolate individual channel performance
- Budget optimisation in a flat-spend environment means moving dollars from saturated or underperforming channels to those with headroom