Für mehr Sinn und Verstand in der Redaktionellen Gesellschaft
Der Newsletter der Looping Group
How the ‘digital truth serum’ can unlock business growth
In wenigen Sätzen
What exactly is the effect of marketing? How can the strength of a brand be measured digitally? Can changes in market share be predicted months in advance? A new method of marketing measurement promises to deliver such insights: "Share of Search.'' It analyzes which brands are searched for online and how often. Renowned researchers have come to the conclusion that market share rises and falls in step with search share - albeit with a certain delay. Is someone who is searched for automatically a winner? In the editorial society, this could permanently change communication strategies. Christian Polman, Managing Director UK of LOOPING GROUP, explains the fascinating new way of thinking of leading data analysts.
With more than half of all global marketing dollars now invested in digital, marketers are increasingly looking for new and effective ways to assess the impact of this spend on both the health of their brands and on business performance. The shift to digital has been accelerated by the pandemic, both in terms of consumer behaviour and allocation of ad budgets. The first three months of the initial lockdown in Q2 2020 fast-forwarded the uptake of eCommerce by four years, right across the generations. Advertising investment has naturally followed eyeballs.
For many years it’s been difficult to establish meaningful, let alone predictive, relationships between brand health metrics and consumer behaviour, in particular in digital. For one, brand trackers are often incredibly slow-moving and reactive, taking months to register the impact of any marketing activity. They’re also notoriously expensive, a particular problem when budgets are under close scrutiny and continuous threat of being cut. What’s more, they’re often – though not always – difficult to make actionable. Picking the right brand health metrics can be a dark art in itself, and giving the data any kind of predictive power requires complex maths.
Yet despite increasing fragmentation in the media and marketing ecosystem, a new way of measuring marketing impact has been quickly gathering momentum since emerging earlier this summer. The latest approach uses ‘share of search’ to enable marketers to not only understand the relationship between brand health and market share, but importantly to also predict market shares months in advance. This represents the first true challenger to the widely used and well understood ‘share of voice’ approach, which has – until recently – been the go-to metric today for predicting changes to market share.
The predictive power of share of search
Share of search has a lot going for it. As the runaway market leader for search in almost all markets other than China, Google Trends is the single source of data for calculating share of search. The database goes back to 2004, is updated constantly to 36 hours before a query is submitted, and generates comparisons by market, including both at the country and the city level. This allows near-real time comparisons of share of search on a weekly basis, enabling analysts to tease apart the impact of different marketing activities. Marketing professor Mark Ritson is an early convert to share of search, calling it “a decent proxy for salience, mental availability, and consideration”. 1 Beyond the power of the metric, it’s worth noting it’s both inexpensive and straightforward to run.
One of the pioneers of this approach is Les Binet, head of effectiveness at adam&eveDDB. His team analysed share of search on Google for three different categories with different buying patterns – energy suppliers, mobile phone handsets, and the automotive sector. In each category, as share of search increases, so market share increases in lock-step, typically with a lag of 3-6 months. For energy, Binet found the lag was 0-3 months, for mobile handsets up to six months, while for automotive purchases it was 12.
As a result of this lag factor, share of search is a powerful tool for predicting business performance – in particular market share changes – well in advance of it actually changing. Critically, this gives businesses a longer runway to change course by reversing, or perhaps accelerating, certain strategies and tactics as early as possible. Used effectively, this can provide a real advantage over slower competitors who aren’t able to react and adapt early enough. The differentiator will soon shift from those who are simply using this new tool, to those who are using it effectively to steer and alter key decision making across the business.
Strategist James Hankins published a separate, extensive analysis of share of search on WARC in August, titled “The most important metric you’ve never heard of”.He also found a reliable and strong predictive relationship between share of search and share of market and in categories other than those assessed by Binet, including retail (supermarkets, department stores, large retailers), broadband, and premium make-up.2
Modelers using share of search don’t make any causal connection in the correlation between the metric and market share – nor do they need to. This is because share of search is effectively a measure of the total impact of all marketing activity combined – what Hankins calls “an output metric of all of business’ endeavours”. What is also clear is that share of search is a lead indicator of market share – share of search increases before and predicts the increase of market share. A typical laydown is shown in Figure 1.
Figure 1. Share of search: a promising predictive metric
The marketing community’s adoption of share of search as an unmediated measure of marketing impact echoes the findings of the data scientist and economist Seth Stephens-Davidowitz. In his book Everybody Lies: What the Internet Can Tell Us About Who We Really Are, he observed that people lie to researchers, doctors, their friends and families, but their search behaviour reveals what they’re really interested in. In his 2018 TED talk, he describes Google search as “digital truth serum”, with incredibly reliable predictive value.3 Google, indeed, describes itself as the “world’s biggest database of human intentions”.
The challenges with share of voice and sentiment analysis
Share of voice has been the primary tool to drive and predict market growth, however digital media present specific challenges. As digital is increasingly personalised and behind walled gardens, tracking competitive and category spend has become difficult. This makes calculating share of voice almost impossible, which requires knowing a brand’s ad spend as a proportion of the total competitive set. In some categories, such as technology, it has also become harder for brands to know what their competitive set actually is. As such, the metric has become less effective at predicting share of market and so less useful to brands.
While sentiment analysis has many useful applications, it is an unreliable alternative to share of voice. The brands people talk about on social media and other online forums tend to be those that they’re least likely to buy. In automotive, for instance, all the chatter is about Porsche, Ferrari, and Lotus, but not the best-selling brands, Ford, Kia, and Hyundai. Moreover, the tonality of the discussion surrounding a brand is a poor predictor of share of market, as the brands that sell best are also often the brands that attract the most negative sentiment. This is in no small measure because popular brands have more users in total and so more negative experiences overall. New analytic techniques, however, are emerging that use machine learning to pick up on the subtleties of online discussions better. These approaches have significant application in understanding how communications activities, particularly those aimed at generating consumer dialogue, are impacting consumer sentiment.
Building on share of search
One notable limitation to share of search is that when consumers are searching for brands in the wake of bad news – for an automotive brand after a product recall, for instance – share of search goes up while market share may well be falling. Share of search is unable to tell you why people search for a product or brand, nor whether they’re searching for positive or negative reasons. An important next step will be to combine share of search with other indicators – including a more context-sensitive variant of sentiment analysis – which should help refine the data and also create stronger links to specific marketing activations.
Reception for share of search in the marketing community and in specialist media has been warm to the point of enthusiastic – a rare point of good news in an otherwise dismal year. Binet is experienced and modest enough to know – and to have declared – that share of search is “by no means a silver bullet” for predicting market share. Even though it is not a perfect predictor, it does accommodate and get around many of the challenges CMOs have faced historically in understanding the drivers of market share with advanced warning.
As brands look to recover and re-establish growth in 2021 and beyond, it’s likely that share of search will grow in popularity. Its application will prove particularly popular given the budgetary pressures all CMOs now face. In the recessionary, post-pandemic market, it is those CMOs who speak the language of the C-suite – who can demonstrate the financial impact of their media and marketing investments on business performance – who will most successfully bridge the CMO-CFO divide. Share of search should be embraced as a welcome tool that helps them do precisely that.
To the Person
Christian Polman is Managing Director UK at Looping Group with expertise in corporate strategy, digital transformation, and growth enablement. His career has spanned senior strategy roles, including most recently as Chief Strategy Officer of independent global marketing analytics consultancy Ebiquity plc, as a consultant at Bain & Company, in Strategy & Operations at Google, and in the Strategy & Analysis unit of Digitas. He has worked with leading companies such as Unilever, PepsiCo, TJX, Royal Bank of Scotland, ITV, Volkswagen, and General Motors. He holds a B.S. and an M.B.A. from Cornell University in New York.
1 “It’s time to replace ‘share of voice’ with ‘share of search’”, Marketing Week, 24.09.20, https://bit.ly/2HxKBGP
2 “Share of search: The most important metric you’ve never heard of”, WARC, August 2020, https://bit.ly/35tXlq8
3 ‘‘The secrets in our Google searches”, TEDxWarwick, https://www.youtube.com/watch?v=IaFQVNuAA3w