Share of search (SOS) “burst” onto the marketing stage in 2020. Just to recap, SOS is the share a brand has of a markets' total brand search volume which can be sourced from Google Trends or Google Keyword Planner. If you work in digital marketing, SOS is by no means a new concept. What was new was the claim of the broad applicability of share of search as a marketing non-financial lead indicator of market share and a “free” replacement for brand tracking.
What most people don’t know, however, is that Share of search is not the full name of the measure. It is actually share of organic search of a specific brand name so it is more accurate to call it share of brand search (SoBS) or perhaps more aptly as is about to be explained, Share of BS.
Share of brand search or perhaps, share of BS
CMO’s need a near-term, non-financial lead indicator of sales. The more accurate the forecast, the greater the confidence the CMO has in taking action. Therefore, it is easy to understand why a CMO might embrace the promise of a three-month lead indicator and therefore why the marketing world initially took an interest in the SoBS claims.
If there is a forecasted contraction in sales which is three months from today, then the CMO has a range of marketing tools and tactics to address the situation. However, when a forecast is unreliable, perhaps because it naively only considered one predictor (such as SoBS), the CMO can find themselves mid-month, realising they have an unexpected contraction in sales. At that late stage, the only marketing tactic that a CMO has at their disposal to address the contraction is the potentially brand damaging, price lever. Compounding this mid-month situation is the fact that at this late stage and in the heat of battle, the CMO is unlikely to be able to tell if the contraction is for their brand or for the category.
In stark contrast, survey data along with the econometric modeling and other data sources such as media spend and economic activity, has the potential to produce a statistically validated and reliable near-term, non-financial lead indicator.
The CMO cannot tell if the contraction is for their brand or for the category.
Remember Maslow's Hammer
Is share of BS a non-financial lead indicator? Econometricians have now had sufficient time to test the veracity of the share of BS claims.
In marketing, it is common to see evidence of Maslow’s hammer. That is, the cognitive bias that results in the frequent application of the same tool. Put perfectly by Abraham Maslow (1966) ‘I suppose it is tempting, if the only tool you have is a hammer, to treat everything as if it were a nail.’ Forethought has observed the unmistakable signs of a herd instinct forming around SoBS.
Speaking only for the econometric modeling undertaken by Forethought, (although we have seen similar results from other analytics companies), if you are constructing an explanatory model to predict near-term sales and all that you have is SoBS data, then yes, there may have some (albeit) weak and noisy explanatory power. However, the moment you add other variables, this small contribution to explanatory power is overpowered and results in SoBS often being expunged from the predictive model.
The following are some of the explanatory variables we found that outperform SoBS, both in terms of significance and explanatory power:
- Measures of consumer sentiment
- Macroeconomic measures of inflation, wage growth, household savings
- Media spend
About being causal
The chief proponents of SoBS did not stop at claiming it was a reliable, non-financial lead indictor. Claims have also been made that SoBS has a causal relationship with consumer choice (fortunately some of these have since been retracted).
Consider you have decided to catch a train from London to Paris. The proponents of SoBS have discovered that you get on the train to Paris and sometime later you arrive at Paris. The brand you type into your browser is highly likely to already be in your consideration set and also likely to be the brand you choose. Searching for a particular destination doesn’t cause you to get on that train. It merely, uncovers that that destination/brand is salient for what you are searching. That does not seem like a great revelation.
SoBS proponents are only focussed on what happens once you have entered a brand name into your browser. Preceding that, brand owners want to influence further up the proverbial funnel. How do I get more people to want to travel somewhere? How do I get more people to choose to go to Paris? How do I get more people to want to use my travel package to get there? Survey data along with other forms of data form the base of an explanatory econometrics model that helps CMO’s to identify what must be done to build brand preference.
Preceding that, brand owners want to influence what brand name you type into your browser
Here, this would involve estimating the likelihood that you will be going to Paris, before buying your ticket. And more importantly, where the decision is not to go to Paris, identifying what the rational and emotional drivers are that would lead you to decide to go to Paris. This is the kind of insight captured in survey data but not in SoBS.
It is this survey data along with econometric modeling that enables brand owners to build their brand preference and share.
When you add it all up
Forethought econometric modeling has been unable to validate the claimed causal relationship between share of search and consumer choice. Indeed, some might argue that predicting sales peaks and troughs is an easier task than predicting market share movement and given that Share of BS does not even predict sales then, claiming to predict changes in market share is such a long, long bow.
Incidentally, for our econometric modeling, Forethought will continue to add SoBS as a hypothetical independent variable and allow the data to speak for itself.
We conclude that the assertion that share of BS should replace survey data for brand diagnostics and market share measurement is somewhere between inexplicable and bizarre. The undeniable sub-script coming from SoBS hype is that share of search explains market share and share of voice explains share of search. Therefore, advertise more and all else will follow.
In summary, it seems that an advertising ‘hammer’ has found its ‘nail’ in share of BS.
Ken Roberts is Executive Chairman and founder of Forethought.