Share of Search vs Market Share: How to Measure Whether Preference Is Growing

Revenue tells you what customers bought. It does not necessarily tell you what is happening to the competitive demand underneath those purchases.

A company can continue producing healthy revenue while gradually losing relative customer demand to competitors. The reverse can also happen. Interest in the brand can strengthen before the improvement becomes fully visible in sales, particularly in categories where purchase cycles are long or where changes in preference take time to flow through the commercial system.

This is why Share of Search is one of the first things I want to look at when assessing the quality of a company's growth.

Share of Search measures a brand's share of branded search demand relative to a defined competitive set. Rather than asking whether searches for your company increased in isolation, it asks how much of the relevant branded search activity belongs to you compared with the companies customers could choose instead.

That competitive context is what makes it considerably more useful than absolute branded search alone.

In our work at Growth Dynamics, we have found Share of Search to be a particularly strong signal of underlying brand demand. We give it the most weight in the first layer of our growth-quality measurement because it is relative. Most other indicators can rise simply because the category is growing. Share of Search forces us to understand whether the brand is gaining or losing ground within that growth.

One first-party example is a good illustration of why I take it seriously. In one of our datasets, marketing spend was essentially unchanged while Share of Search increased by 15% and revenue increased by 60%.

I am not claiming that the 15% increase in Share of Search caused the 60% increase in revenue. There were other things happening in the business, as there always are. What made the result interesting was that a directly observable measure of relative customer demand strengthened materially alongside a major commercial improvement while additional marketing spend was not the explanation.

That changes the measurement question. Instead of immediately asking which channel should receive credit for the revenue increase, I want to understand what changed in the demand system underneath it.

This is where Share of Search becomes commercially valuable.

Why relative branded demand matters

Branded search volume already contains useful information. If substantially more people are searching for a company, that is real observable customer behaviour and I want to know about it.

The problem is that absolute growth has no competitive context.

Your branded demand can increase while your relative position deteriorates if competitors are growing faster. The entire category might be receiving more interest. A major external event could be increasing search activity for everyone. Looking only at your own search trend can therefore make the business appear stronger while competitors are gaining ground more quickly.

Share of Search introduces the denominator.

The company is no longer asking simply whether people are searching for it. It is asking what proportion of the relevant branded search demand it owns.

That is why competitor selection matters so much.

The denominator should contain companies that represent realistic alternatives for the customers you are trying to win. Adding a huge category brand that operates in a materially different market can distort the signal. Omitting a fast-growing competitor can conceal exactly the change you are trying to detect.

Our own rule is to keep comparable rivals in the denominator and use category giants as a separate reference where they provide useful context.

Kantar reaches a similar conclusion in its work on Share of Search. It describes the calculation as simple in principle but highlights the importance of competitor selection, keyword coverage and search-data quality in producing something commercially meaningful. Its own client validation work has found Share of Search correlated with both salience and sales.

The simplicity of the metric should therefore not be confused with simplicity in its construction.

What the published evidence actually says

The external evidence around Share of Search is strong enough that I am comfortable using it seriously, but it is important to describe the findings correctly.

The IPA Share of Search Think Tank analysed 30 cases across 12 categories and seven countries. Across those cases, it found that Share of Search appeared to represent 83% of a brand's Share of Market on average. It also found a relationship between changes in Share of Search and subsequent changes in market share, while noting that the exact ratio varies between categories and countries. Importantly, the IPA explicitly describes these findings as correlations rather than causal relationships.

The IPA's current guidance continues to describe organic Share of Search as an early indicator of future movements in market share, in some cases up to a year ahead.

That is consistent with what makes the metric valuable to me, but I would not interpret it as a formula that converts a percentage increase in Share of Search into a predetermined increase in market share.

Too much sits between search demand and the eventual sale.

Price matters. Distribution matters. Product availability matters. Competitor behaviour matters. The buying experience matters. A customer can strongly prefer a company and still be unable or unwilling to purchase it.

Share of Search is therefore best understood as an observable signal of the demand side of the market rather than as a replacement for market share itself.

The fact that we have also seen it work well in our own operating data gives me more confidence in it than I would have from the published literature alone. That is practitioner evidence, not a reason to overstate what the research proves.

Share of Search and market share answer different questions

Market share describes a realised commercial outcome. Share of Search describes relative customer behaviour occurring before all of that demand has necessarily become revenue.

The relationship between the two is where much of the value sits.

When both are strengthening, the story is relatively coherent. The company appears to be gaining relative customer demand and converting enough of that demand to improve its commercial position.

When they move differently, the analysis becomes more interesting.

Share of Search can strengthen while market share remains flat because demand has not yet matured into purchases. That becomes particularly plausible in a category with a long consideration journey. The customer may begin considering, researching and searching for the company months before they eventually transact.

There can also be a more immediate commercial constraint. Interest in the company might be increasing while product availability, pricing, distribution or the buying experience prevents the business from converting that preference into sales.

Kantar uses Share of Search in a similar diagnostic way. Its work recommends combining Share of Search with other brand and commercial indicators to distinguish situations where media appears to be generating interest but something later in the journey is preventing that interest from becoming market share.

The opposite combination deserves attention too. Revenue or market share can remain healthy while Share of Search weakens. Current performance may be supported by previous brand strength, repeat customers, distribution improvements, pricing or category growth while relative customer demand has already started deteriorating.

That is one of the reasons I see Share of Search as an early commercial signal rather than simply a marketing KPI.

By the time the deterioration reaches the P&L, the underlying movement may have been happening for some time.

The first-party case is useful because the signals moved together

Our own +15% Share of Search and +60% revenue example is worth returning to because it illustrates how I want the measure to be used.

Marketing spend did not increase. Relative branded demand did. Revenue increased very substantially at the same time.

The wrong conclusion would be to say that Share of Search caused the revenue increase.

The useful conclusion is that there was clearly something happening in the underlying customer-demand system that an analysis focused purely on changes in media spend would have missed.

The revenue increase could not simply be explained by saying that the company spent more money.

That leads to better questions.

Had the brand become more likely to enter consideration? Were customers increasingly searching for it relative to competitors? Was the company monetising existing preference more effectively? Had distribution or product availability changed? Did creative improve? Were competitors weakening?

Share of Search does not answer every one of those questions. It tells us where to investigate.

That is much more valuable than forcing the entire growth story into an attribution model and allocating the additional revenue between whichever channels appeared before the sale.

It is also why the strongest interpretation comes from reading Share of Search alongside other evidence rather than in isolation.

Geography makes Share of Search much more actionable

A national Share of Search figure is useful for understanding the overall direction of a brand, but national averages can hide substantial differences between markets.

Brand strength is not geographically uniform. Competitors vary. Distribution differs. Marketing investment can be uneven. Customer mix changes, and the same company can occupy a very different position in one region from another.

Looking at Share of Search geographically allows us to identify where relative branded demand is strengthening and where it is losing ground.

That becomes particularly valuable when geographic commercial data is also available.

For a business with substantial offline sales, for example, we can compare regional branded search behaviour with store or regional revenue. Our own Growth Quality framework explicitly adds this offline bridge for clients where online revenue alone would provide an incomplete commercial read.

This does not mean treating every regional search as a precursor to a store transaction. The purpose is to understand whether the observable footprint of customer demand and the commercial outcome are telling a coherent story across markets.

When they are not, the gap becomes diagnostic.

If relative search demand is strengthening in a region while revenue is disappointing, the problem may sit in monetisation rather than demand creation. If revenue remains healthy while relative branded demand has been deteriorating consistently, current performance may be concealing a weaker future competitive position.

This geographic view also creates useful hypotheses for later causal work. It can show which markets behave differently before the business decides where to run an incrementality test or how representative particular geographies might be.

Share of Search is evidence of demand movement, not proof of causality

One of the most important boundaries around Share of Search is that it remains observed evidence.

If a company increases brand advertising and Share of Search subsequently rises, I find that interesting. I do not automatically conclude that the advertising caused the entire movement.

Advertising itself can generate branded search, which is one reason Share of Search can provide a useful observable footprint of marketing activity. The IPA's work has found a relationship between advertising Share of Voice and subsequent Share of Search.

But product launches, news coverage, word of mouth, distribution changes, competitors and other factors can move branded demand as well.

This creates a useful distinction between observing the footprint and estimating the causal effect.

If an upper-funnel campaign is supposed to increase future customer demand, I want to know whether anything observable changes before accepting a modelled claim that the investment will produce a large long-term effect. Share of Search may provide one such observation. Buyer evidence, branded organic demand, consideration and meaningful customer behaviour can provide others.

None proves the campaign's incremental ROI.

Together they can materially improve the evidence base around the investment.

This is particularly important for long-term effects because those effects can become extremely difficult to identify statistically. A small commercial effect spread over many months can disappear into normal revenue variation. An econometric model can represent a long-term effect mathematically, but as the signal weakens, assumptions about lag, priors, controls and baseline behaviour begin carrying more of the answer.

I would rather have observable evidence of demand movement available to challenge that model than ask the model to infer the entire story on its own.

Preference and monetisation need to be separated

This is why the first Growth Dynamics measurement layer asks two related but distinct questions:

Is preference growing, and is the business monetising it?

Share of Search carries the most weight on the demand side because it is relative to competition. Branded organic search provides another observable measure of people actively looking for the company. The commercial side then examines whether that customer demand is turning into profitable revenue.

The combination is more informative than either part independently.

Strong Share of Search growth with weak revenue can indicate that preference is moving but the business is failing to monetise it. That should lead to questions around pricing, availability, distribution, website or store experience and purchase timing before anyone concludes that the marketing responsible for generating demand should be cut.

Strong revenue with weakening relative demand creates a different concern. The business may be monetising its current customer pool very efficiently without replenishing it at the same rate.

Our first-party case was particularly interesting because both sides moved strongly while marketing spend remained stable. Relative demand strengthened and the commercial outcome improved substantially.

That is exactly the sort of relationship I want this layer of measurement to reveal.

Paid brand search should not be counted as demand creation

Paid brand search needs separate treatment because it purchases access to demand that is already visible.

When more people search for a company, the paid-search account receives more opportunities to serve advertisements. Paid clicks and attributed revenue can increase because underlying brand demand increased rather than because paid search created the demand.

This makes paid brand search a poor indicator of whether preference is strengthening.

We use it instead as a cannibalisation check against branded organic search. Growth Dynamics' Step 1 methodology explicitly separates paid brand spend from the demand lines for that reason.

If paid coverage increases while branded organic clicks decline, there may be substitution occurring. If paid brand investment is reduced and organic search absorbs a substantial portion of the lost traffic without a comparable reduction in total revenue, that provides evidence about the incremental economics of the paid investment.

Whether that warrants a designed experiment depends on how much money is involved.

But for the simpler question of whether underlying branded demand is strengthening, I want to observe the search demand itself rather than the amount of that demand the company chose to buy through paid advertising.

Keeping those two questions separate prevents demand capture from being mistaken for demand creation.

The disagreements are often more useful than the agreements

A good measurement system should not require every indicator to tell the same story.

If Share of Search, branded organic demand and revenue all increase together, interpretation is relatively straightforward.

The more valuable analysis often begins when they diverge.

If Share of Search improves but revenue does not, I want to understand what is preventing stronger customer demand from being monetised.

If revenue improves but Share of Search deteriorates, I want to know whether the company is becoming more efficient at harvesting a demand pool whose relative competitive position is weakening.

If advertising investment rises substantially but there is no movement in Share of Search or any other meaningful demand indicator, I want to understand what the advertising was intended to accomplish and whether there is evidence of that effect anywhere else.

If buyer evidence says customers increasingly prefer the company while branded demand falls, I want to understand whether the customer sample represents the wider market.

The objective is not to select whichever measure supports the story marketing would prefer to tell.

It is to use the differences between the signals to locate where the uncertainty actually sits.

That is one reason Kantar recommends using Share of Search alongside other indicators rather than as a standalone replacement for brand tracking or market share.

What Share of Search cannot answer

I am positive about Share of Search because we have found it commercially useful and because the external evidence gives us good reason to take it seriously.

That does not require asking it to answer questions outside its boundary.

Share of Search does not establish which marketing channel caused a change in customer demand. It does not provide an incremental ROI. It does not measure every dimension of brand strength, and the strength and timing of the relationship between Share of Search and market share vary by category.

The underlying search data also matters. Some categories generate much richer branded-search behaviour than others, and search terms can contain noise unrelated to purchase demand. Kantar specifically warns that categories with insufficient search volume may not produce the expected relationships and that data quality needs to be assessed before relying heavily on the measure.

Those are boundaries rather than reasons to dismiss the metric.

Within those boundaries, Share of Search answers an extremely valuable question: is observable customer demand for the brand strengthening or weakening relative to its competitors?

I want that answer before moving immediately to a more assumption-heavy technique.

If the decision requires causal evidence about a particular intervention, design an experiment. If the business needs a whole-mix decomposition and response curves, use MMM.

Share of Search does not need to perform either of those jobs to be useful.

Use Share of Search as a commercial signal, not another dashboard metric

The best use of Share of Search is not adding another line to a brand dashboard.

It is connecting the movement in relative customer demand with a business decision.

If Share of Search is strengthening while revenue is not, investigate the monetisation problem before cutting demand-generating activity.

If revenue is strong but relative customer demand is weakening, understand whether current financial performance is being supported by a competitive position that is becoming less sustainable.

If both strengthen, as we saw in our own same-spend, +15% Share of Search and +60% revenue case, the business has a much stronger reason to investigate what changed in the customer-demand system and whether those conditions can be reinforced.

The published research supports using Share of Search as a proxy for the consumer-demand side of market share and as an early indicator of market-share movement. Our own experience makes me more confident in its practical value than that description alone.

I still want the claim to remain honest.

Share of Search is not the brand. It is not causality, and it is not a complete marketing measurement system. Our broader strategy explicitly treats it as a better proxy for brand demand than a model-implied long-term coefficient while still recognising that it remains a signal rather than the asset itself.

That is enough.

A strong observable signal of competitive customer demand is extremely valuable when the alternative is waiting for revenue to tell you what happened months later or asking a model to infer something the business could have observed directly.

Growth Dynamics uses Share of Search as the primary relative-demand signal in the first layer of our marketing measurement framework. We connect it with branded organic search and commercial performance to understand whether customer demand is strengthening and whether the business is successfully monetising that demand.

The purpose is not to prove everything about brand.

It is to see an important part of growth while there is still time to act on it.