In short: share of search is the proportion of brand searches for you against all brand searches in your category. Presented by Les Binet at IPA EffWorks 2020, it leads market share by 6-12 months, up to a year in categories like automotive (IPA, Binet, 2020). It is calculated for free with Google Trends, which makes it marketing's cheapest leading indicator. Later research by James Hankins across 30 cases in 12 categories and 7 countries found it accounts for about 83% of a brand's market share.
What is share of search?
Share of search measures how many people search for your brand relative to how many search for all brands in the category. If your category runs 100,000 brand searches a month and 12,000 are for your brand, your share of search is 12%. It is the digital equivalent of share of voice, with one important difference, because it rests on real behaviour, searching, instead of on advertising spend.
The metric was popularised by Les Binet at IPA EffWorks 2020 (IPA, Binet, 2020). Binet analysed three categories with very different purchase cycles, automotive, energy and mobile phones, and in each found that share of search correlates with market share and acts as a leading indicator of it. We use it as a first dashboard when we take on a new client, often before we have even touched the sales data.
Why does it lead market share?
The logic is simple. Interest precedes purchase, and when a brand grows in people's minds, through advertising, word of mouth or fame, searches for its name rise before those same people become customers. Share of search captures this latent demand early, while sales record it only once the purchase has already happened and the informational advantage, by then, has passed.
The time lead is documented. According to the analyses presented at the IPA and reported by WARC, when share of search rises market share tends to rise over the following 6-12 months, and when it falls share falls. Later research by James Hankins across 30 case studies in 12 categories and 7 countries found that share of search accounts for about 83% of a brand's market share.
In one of our own cases it warned us a few months in advance of a decline the client blamed on seasonality and that was in fact share erosion to a competitor more aggressive in communication. This is not statistical proof, it is an anecdote, and we give it for what it is worth. It did make us stop looking only at sales figures after the fact.
How does it compare with other brand-health metrics?
Source: based on IPA/Binet (2020) and WARC.
Share of search does not replace brand tracking, which also measures why a brand is strong, which associations it holds, how distinctive it is. It complements it as a fast, cheap signal, a dashboard you can watch weekly instead of waiting for the next paid survey.
How to calculate it with Google Trends
The basic procedure needs no paid tools. One technicality is worth remembering before you start: Google Trends does not return absolute volumes but a relative index normalised to 100, which is fine for computing shares and trends but makes it pointless to add values across different periods or categories without a common reference.
- Define the category. List the main competing brands, including yours. The category must be coherent: compare who genuinely competes for the same need.
- Extract brand search volumes. Compare brand names in Google Trends, up to five at a time; for larger categories chain comparisons through a common reference brand. Count searches for the brand name, not generic product keywords.
- Compute the share. Brand share of search equals brand searches divided by the sum of searches for all brands in the category, in the same period and area.
- Read the trend, not the absolute value. The useful signal is the direction over time: rising it anticipates share to come, falling it flags a risk to fix before it shows in sales.
Limits to know
Share of search works best where people search for brands by name, such as automotive, electronics, telecoms and many services. It is less reliable for very low-involvement products rarely searched, for brands with ambiguous names that generate irrelevant searches, and in niche B2B markets with volumes too low to be stable. It also has to be isolated from spikes caused by news or viral campaigns, which temporarily inflate searches without reflecting real share growth.
FAQ
Does share of search really predict sales?
In the categories Binet studied (automotive, energy, telecoms) and in Hankins' later research across 30 cases, it leads market share by 6-12 months and explains about 83% of it. It remains a strong correlation, not a mechanical guarantee, and should be validated on your own market.
Do I need a paid tool?
No. The basic calculation is done with Google Trends, free of charge. Paid tools, with absolute volume data, add precision but are not necessary to start monitoring the trend.
Share of search or share of voice: which do I use?
They are complementary. Share of voice measures how much you spend on advertising relative to competitors, an input; share of search measures the real interest that results, a leading output. Watching them together tells you whether the investment is turning into demand.
How often should it be measured?
It can be watched continuously, even weekly, precisely because it is cheap. The key is to reason on the medium-term trend and not react to very short-term noise.
Sources and references
- Binet, L. (2020). Share of Search, a fast, cheap, predictive metric. IPA EffWorks Global 2020.
- IPA (2020). Share of Search as a Predictive Measure. IPA.
- WARC. Share of search can predict market share. WARC.
