Why businesses that depend on several separate groups keep reporting a single awareness figure that describes none of them.

Key takeaways

1. Plenty of businesses depend on two or three groups who have nothing to do with each other. A marketplace needs suppliers and buyers. A building products manufacturer needs the architect who specifies, the contractor who installs and the owner who signs. Each is a separate market with its own reasons to choose.
2. Most trackers sample only the group the business thinks of as the market, which is almost always the one that pays. The other sides are absent from the data entirely, so their weakness cannot appear in the reporting however bad it gets.
3. Where all sides are sampled, aggregation finishes the job. A blended figure lets a strong result in the large audience offset a weak one in the small, and the small one is usually the side that constrains growth.
4. The finding is the gap between sides, not the level of either. A business can be well known to everyone who buys from it and unknown to the people it needs in order to sell more, and report that as healthy awareness.

A trades marketplace grows for three years on the strength of its consumer brand. Homeowners know it, use it and rate it. Then growth flattens in two states at once, and it is not a demand problem: jobs are being posted and going unfilled because there are not enough tradespeople on the platform to take them.

The board asks what the brand tracker showed. The tracker showed awareness rising every wave for three years. It was fielded among homeowners, because homeowners are who the business thinks of as its market, and homeowners are indeed where the revenue comes from. Nobody had ever measured what a plumber in Geelong thought of the platform, or whether they had heard of it at all.

Nothing in that tracker was wrong. It answered the question it was designed to answer, waved after wave, accurately. It simply could not see the side of the business that had become the constraint, and by the time the constraint showed up in the numbers it was showing up as flat revenue, which is the most expensive place to find it.

Where one score hides two markets

This is a general condition rather than an edge case, and it is worth recognising the shape rather than the sector, because the same structure recurs in businesses that look nothing like each other. It is the commercial version of Awareness Asymmetry, which we defined in the context of purpose-led organisations serving several separate audiences.

Marketplaces and platforms. Supply and demand are two distinct markets with two distinct brand jobs. Being the obvious choice for buyers does nothing to recruit sellers, and the side that is scarcer is the one that sets the ceiling.

B2B where the specifier is not the purchaser. A product gets written into a specification by one party, bought by another, and lived with by a third. Each of them chooses on different grounds. A tracker fielded among "decision makers" quietly picks one of the three and calls it the market.

Franchise and dealer networks. Customer brand and operator brand are separate assets. A business can be popular with the public and close to invisible to the operators it needs in order to open in a new region.

Labour-constrained businesses. Where growth is capped by who you can hire rather than who you can sell to, employer awareness is a commercial metric sitting in an HR budget, measured on a different cycle by different people, if it is measured at all.

Channel businesses. The shopper and the category buyer at the retailer are separate audiences with separate criteria, and a brand that shoppers reach for can still lose the shelf it needs in order to be reached for.

What these share is that the sides are not interchangeable, and that the aggregate is arithmetically valid and commercially meaningless.

Why does a tracker miss something this large?

Three mechanisms, and they compound.

Sampling decides the answer before the questionnaire does. Research gets commissioned by the part of the business with the budget and the question, which is usually marketing, and marketing's audience is usually the buyer. The other sides are not deliberately excluded. They are simply never named as audiences, so they never appear in a sampling frame. Everything downstream is then a precise description of one side of a business that has more than one.

This is sharpened when the missing side is hard to reach in the first place. Tradespeople, specifiers, franchise prospects and skilled candidates are all expensive, low-incidence samples. There is a standing temptation to field the easy side well rather than the hard side adequately, and it is a false economy when the hard side is the one that binds.

Aggregation hides what sampling let through. Where two audiences are in the same study, results are reported as a total because a total is what the summary slide wants. A strong result in the larger audience mathematically offsets a weak one in the smaller. The blended figure then moves reassuringly for years while one of its components falls.

Weighting buries the decisive side. If the study is weighted to population, the side with fewer people contributes proportionally less, which is exactly backwards. A marketplace might have hundreds of thousands of buyers and a few thousand suppliers. Weight that to population and the supply side effectively disappears from the number, at the precise moment the supply side is what the business needs to know about.

None of this requires anyone to do anything wrong. It is what careful, methodologically sound research produces when the question of who the audiences are was settled by default rather than deliberately.

The side that matters most is usually the smallest

This is the part that makes the problem worth solving rather than merely noting.

In most multi-sided businesses the audiences are wildly different sizes, and the commercially decisive one is frequently the small one. There are always fewer suppliers than buyers, fewer specifiers than end users, fewer prospective franchisees than customers. Scarcity is what gives that side its leverage.

The consequences are asymmetric too. Losing ground among buyers usually costs something proportionate: fewer sales, a bit less pricing power, a gradual line on a chart. Losing ground among the scarce side can close a region, stall a product launch or cap growth outright, because the business cannot operate at all without enough of them.

So the measurement effort and the commercial stake point in opposite directions. The side that is cheap to sample and easy to reach is the side where the downside is gentle. The side that is expensive to sample is where the downside is a cliff. Left to inertia, research budget flows toward the first one, and the asymmetry widens quietly for as long as nobody asks.

It is the same failure of averages that lets a brand lose relevance inside one group long before any aggregate metric moves. The difference here is that the group is not a segment of one market. It is a separate market the business forgot it was in.

How to measure it instead

Four changes, none of which requires a larger research budget so much as a differently allocated one.

Define the sides by what they decide, not by who they are. Demographics and job titles will mislead you. The useful question is which distinct decisions have to go your way for the business to grow, and who makes each of them. If two groups make genuinely different decisions on genuinely different criteria, they are two audiences, whatever the org chart says.

Field each side as its own study, sized to be read alone. Not a booster, not a quota within a general population sample. Each side needs a base that supports reading it on its own, because the whole point is to compare sides, and you cannot compare a robust number with a fragile one. This usually means accepting a smaller sample on the large, easy side in order to afford an adequate one on the small, hard side.

Never report a blended figure. There is no decision that a combined awareness score across two unrelated markets improves. Report the sides separately, always, and let the difference between them be the headline rather than something a reader has to reconstruct.

Track the gap, not just the levels. The ratio between sides is the diagnostic. Both sides rising is growth. One side rising while the other holds is a widening exposure that will eventually cap the first. Watching that ratio over time is what turns this from an interesting observation into an early warning, and it is only visible if the sides were measured separately from the start.

Most of this is a design decision made before fielding, which is why the problem is cheap to prevent and expensive to retrofit. A tracker that has run for three years on one audience cannot be made to tell you about the other, and starting the second one now means waiting three years for a comparable trend.

Frequently asked questions

How do we know whether we are actually a multi-sided business?

Ask what would stop growth first if it went wrong. If the honest answer involves a group you are not currently measuring, suppliers, specifiers, operators, candidates, retail buyers, you are multi-sided and your tracker is single-sided. The test is not whether you sell to more than one type of customer. It is whether the business needs more than one group to say yes before anything happens.

Can we just add the missing audience to our existing tracker?

Sometimes, but rarely well. The questionnaire is built around one audience's decision, so the attributes, the competitive set and the occasion framing will all fit one side and not the other. It is usually cleaner to run a parallel instrument that shares the core metrics, so the two are comparable, while letting everything else differ.

What if the second audience is too small or too hard to sample affordably?

Then it is measured less often rather than not at all. An annual read on the scarce side alongside a quarterly read on the main one is far better than a precise quarterly figure that describes half the business. Frequency is the right thing to trade away here. Coverage is not.

What this changes

The practical consequence is not a new metric. It is a different unit of analysis.

Most brand reporting implicitly treats the business as having one market, then measures that market carefully and reports levels rather than the relationships between them. For a business that genuinely has one market, that is fine and the rigour is well spent. For a business that has two or three, it produces a number that is accurate, defensible, and describes a market that does not exist: an average of groups that share no criteria, no competitive set and no reason to care about each other.

The businesses that get caught by this are rarely careless. They are usually the ones measuring diligently, on a long-running instrument, with a stable methodology and a trend they trust. That is exactly what makes it dangerous. A figure that has been rising consistently for three years is not interrogated, and the side of the business it cannot see is not missing from the report in any visible way. There is no blank space where it should be. There is just a number that looks like the whole picture.

The fix is neither expensive nor clever. It is deciding, deliberately and in advance, how many markets you are actually in, and then refusing to average them.

If your growth depends on more than one group saying yes, and your brand tracking describes only one of them, the gap between those groups is the number worth having. Brand Health designs brand tracking and brand audit programmes for businesses with several distinct audiences, measuring each side separately so the asymmetry between them is visible before it caps growth.

Schedule a free 30-minute consultation to discuss which of your audiences you are not currently measuring.

Tom Morris is the Managing Director of Brand Health, an Australian brand research consultancy. He works with senior marketing leaders to design measurement programs that connect brand performance to commercial outcomes.