The CMO's Guide to Share of Search: From Theory to Technical Reality
In the quest for brand health metrics that are both predictive and cost-effective, "Share of Search" (SoS) has emerged as the metric of the decade. But for marketing leaders looking to operationalize this data, the path is rarely straightforward.
Share of Search is defined simply. It is the number of organic search queries for your brand, divided by the total search queries for all brands in your competitive set. It represents your share of the consumer's mental availability at the moment of active interest.
However, simply pulling data from Google Trends often leads to strategic errors. This guide covers the foundational theory, the validation from industry heavyweights, and the messy reality of cleaning data for enterprise-level reporting.
The Theory & Validation: Why It Matters
The credibility of Share of Search moves beyond mere correlation. It is rooted in the "Extra Share of Voice" (ESOV) principles championed by Les Binet and Peter Field. In their seminal work, The Long and the Short of It, they established that brands with a Share of Voice (SOV) greater than their Share of Market (SOM) tend to grow.
Les Binet's research demonstrated that Share of Search is a leading indicator of Market Share. The dynamic is predictive:
When Share of Search rises above Share of Market, market share tends to rise subsequently.
When Share of Search falls below Share of Market, market share tends to fall.
Depending on the category (e.g., automotive vs. FMCG), this lead time can range from a few months to a year. For a CMO, this turns a vanity metric into an early warning system. It allows for course correction before sales figures officially decline.
Figure 1: The famous LG example illustrates how a decline in search volume presaged a decline in global market share months in advance.
The Mark Ritson View
Marketing Professor Mark Ritson has been a vocal proponent of Share of Search, positioning it as a crucial metric for modern marketers. In a landscape where robust brand tracking can cost six figures and take months to process, SoS offers a fast, zero-cost alternative that is remarkably accurate.
"Share of Search is a proxy for mental availability. It captures the moment a consumer thinks of a category and reaches out for a brand. It is empirical, behavioral data. It captures what people are actually doing, rather than what they say they will do."
The Reality Check: Why Google Trends is a Minefield
What should be easy, accessible, and free to pull is far from it.
The process is very time-intensive, requires a lot of data cleaning, and often necessitates a team of specialists to get right—even when using automated tools. But if you want to give it a try, here's my guide on how to do it and the challenges you need to overcome.
1. The "Highest Volume Topic" Dilemma
Topics are usually better than single terms because they capture a group of variations. Your goal is generally to select the Highest Volume Topic available for the brand.
However, you must filter through multiple topic options (e.g., "Brand," "Brand (Company)," "Brand (Fashion Label)") to find the one with the highest demand that remains relevant. If you choose blindly, you might pick up irrelevant traffic.
Choosing the right topic bucket is critical to accuracy.
The Fix: You need to pull data from multiple sources—all relevant topics plus the specific "Head Term"—and merge them.
2. The Intent Trap: Assos vs. ASOS
Brand names are rarely unique. Consider Assos of Switzerland, a premium cycling brand. A simple search query for "Assos" is heavily polluted by typos for ASOS, the massive fashion retailer.
Ambiguous search terms artificially inflate data if not cleaned manually.
. Regional Confusion: The "Houdini" Problem
Houdini is a well-known Swedish sportswear brand. Within Sweden, search intent is clean. However, if you expand your report to a global view, the data becomes overwhelmed by searches for "Harry Houdini" the magician.
The Fix: We use a hybrid approach. We pull data for the specific "Topic" to capture breadth, but cross-reference it with the specific "Head Term" (e.g., "Helly Hansen") to spot anomalies.