Share of Search:
An interesting metric if you use it right

Share of Search if often presented as the saviour of marketing metrics. Especially by small brands now getting access to free market research. And it is truly a fantastic measurement that most brands should use. But are there no downsides, ifs, buts or caveats?

Yes many! Is it easy and straight forward to use? Not always. Is it for everyone? Maybe… Here are some real-world cases.

In practice, we have now lived with Share of Search (“SoS”) as a method of measuring the relative interest in brands for just over a year. We at Kapero have been able to use it practically in a number of client projects. In many ways, it is a great measure since it indicates genuine interest in a brand by using behavioural data from reality and not by being based on hypothetical questions to a consumer panel. In addition, it is free and happens in real time as you primarily download data from the open tool Google Trends. I even allows you to look back in time and to later re-define what you want to study. The measure has also proven to have a strong correlation to future market share.

But the very fact that the data is not gathered for a specific need (e.g. measuring a brand’s strength), but an interpretation of search behaviour by comparing search volumes, raises a number of challenges that for many brands means that SoS is not the savior that many may think. Especially not for smaller brands. This is paradoxical since many small brands seldom have the budget for traditional marketing research, and with SoS are hoping to finally get metrics for free. And above all, SoS is not a tool for strengthening a brand, as has been claimed by agencies providing SoS services.

The challenges are based on how Share of Search works:
1. It measures specific search terms in relation to a number of other specific search terms. For example, “Apple” against “Samsung” etc., or “Apple iPhone 12” against “Samsung Galaxy S21 Ultra 5G”. It is therefore easy to miss a large part of the search volume since you cannot measure all possible searches that may include the brand name as some brands are seldom searched on brand name alone.

You need to decide which search terms to measure and which to oppose. What is your category and your competitors? Do you just want to measure “SAS” against “Ryan Air”? Or do you want to weigh in railway operators as well? Do you even want to measure against alternatives like “camping” or “stay-cation”?

SoS is therefore relatively simple and reliable when comparing very well-known and established, unique brands with large search volumes (especially product brands) in a well-defined category where competing brands are obvious and the product type cannot be easily replaced, and the category generates research before purchase (e.g. cars, fitness watches, lawn mowers, sneaker brands, electricity companies, gym memberships).

SoS is therefore relatively simple and reliable when comparing very well-known and established, unique brands with large search volumes (especially product brands) in a well-defined category where competing brands are obvious and the product type cannot be easily replaced, and the category generates research before purchase (e.g. cars, fitness watches, lawn mowers, sneaker brands, electricity companies, gym memberships).

But Share of Search is much more challenging, for example:
> For lesser-known brands with few searches – especially if competitors have many searches

> When the category or competitors are difficult to define (often broad or new categories)

> When your brand or the competing brands operates in many categories (e.g., Samsung, Best Buy)

> When the brand is often misspelled (Häagen dazs, McDonald’s/Mc Donalds), has generic or category-linked name (Target, Best Buy, Norwegian) or nickname (Coke, Coca Cola, Cola).

> If you need to study a specific segment or target group.

> When interest in the brand is not only linked to sales (e.g. searches about the company itself or when the brand has been involved in scandal, research or debate)

> When you usually combine the brand with other words such as any product model or retailer brand name
you are a retailer brand who sells many product brands, and you are not a product brand yourself

> When the product type primarily is purchased spontaneously; due to store location (e.g., fast food, construction products, convenience products)

> When the product can easily be replaced by an alternative method (e.g. the fast food burger is replaced by a kebab, a banana or the leftovers at home in the fridge)

> When the brands requires no or very little pre-purchase online research (chewing gum, hair shampoo, petrol)

> Because many brands operate in small geographical markets with few searches per search term to give reliable data.

Brands often find themselves in a number of the above situations. This means that SoS is seldom the saviour that many brands with limited budgets for measurements hope for.

A few real-world-examples
Let’s look at a few cases we have worked with: a major bank and a major insurance/savings company – both with a substantial market share, well established brands and millions of customers.

First you must understand that most Google-searches for the brands are generated by customers who regularly need to access their online accounts (by typing in brand name into Google or in browser URL-field), not by non-customers doing research to become customers. This means that just small change in behavior of current customers impacts SoS a lot, for instance when:

> Customers increasingly start using an online service, e.g. online stock trading, logging in more often makes the searches increase for the reason above but does not reflect brand strength.

> An increase in number of customers, driving more traffic for the reasons above, due to a great tactical campaign (a hard-to-resist interest offering). Not related to brand strength but e.g. an aggressive pricing tactic.

> Searches increase due to an international scandal. Related to brand in a way but by negative interest that has generated an increase in SoS.

The challenge here is to separate the search contributions above from changes in general brand interest that may indicate a change in future market share. Please note that these issues are not related to the brand size or search volumes being small.

The conclusion here is that in many real-world situations the reality is quite complex, forcing you to be very cautious when using and interpreting SoS.

Share of Search as a part of the insight work

Mark Ritson has pointed to SoS as SoV and ESoV have shortcomings when it comes to estimating changes in market share. The challenge is that many of these weaknesses also are applicable to SoS – such as the difficulty of defining the category and determining competitors. Should Apple compare their SoS to Samsung’s even though Samsung also sells a heck of a lot of TVs, fridges and washing machines?

An equally big challenge is that most brands, especially in smaller markets, do not operate in the international arena of established well-known product brands (as in the example above) that marketing gurus often do. This results in issues both with small search volumes and questions regarding what local and international brands to compare us with.

To some brands SoS is quite straight forward, but to many brands it is quite challenging as I have highlighted above. But of course, SoS can be a good complement to other marketing measurements. “A complement to what, we do not have the budget for doing comprehensive traditional brand tracking”, you may be wondering? To begin with, a lot of data that can indicate where you are going with your brand can be taken from owned channels such as traffic data (online or physical), the amount of “brand traffic” to your site or by following the development of click-through rate (CTR) on your digital advertising (which has proven to be strongly linked to brand strength).

Your brand and product marketing teams might be really excited about this metric we’re running into called “Share of Search”.

It’s simple to understand: Who do people search for the most in our category?

And it’s easy to calculate, too – get the monthly search volume for your brand, and see what percentage that is of all brands in your category. For example:

If you have 10,000 searches a month, that’s 10,000 searches for you, and… Your 5 competitors all have 5,000 each, that’s 25,000 searches for them, then… Share of search is about 30% – your searches (10,000) over total searches (35,000)!

(A quick and very rough way to get a similar analysis is Google Trends, so start there if you don’t feel like doing a lot of math right now.)