The Future of Brand Health Tracking (See also the video)

(And what it means for people working in Insights)

Patrick Bruin, IKEA, and Sjoerd Koornstra, The House of Insights

Brand Health Tracking is currently under pressure in (some) companies. The current Covid crisis has worsened this situation. Some companies are spending less or hardly at all on marketing activities and are slowing down on the measurement as well. Since our sales are under pressure, we are hardly communicating, so why should we measure the brand health? The fact that more data comes available, structured and unstructured, has had an impact cutting back on Brand Health Measurement as well.  In this article we will explain why we think Brand Health Tracking will be important in the future. Based upon our broad experience within several companies as well as in depth desk research we will paint the framework and conditions for a sound Brand Health Tracking system.  We will also discuss some misunderstandings around Brand Health Tracking which we have seen in practice. This is related to scope of measurement and how it should be used within companies. We will then also indicate the consequences for the Company Insights managers.

Strong Brands are becoming more important than ever before!

The spectrum of sales channels is expanding. Sales via online platforms as well direct will increase. These sales are supported by personalized marketing and promotion campaigns.  Consumers can be influenced at more moments and via different means. Social media influencing will increase but also will digital / bot assistants. Virtual shops, in which the real shopping experience is simulated, will have an impact on purchase decisions. Competitive brands are very active to steal your consumer. Consumers are more often than before bombarded with competitive messages.  The stronger your brand, the less the consumer is susceptive for temptation. And how do you know that your brand is strong in the mind of the consumer? By measuring on a frequent basis.

When to do and not to do Brand Health Tracking 

When should you conduct a Brand Health Tracking and when not? The main reason is that the brand should be an important factor in the purchase decision process.  A brand can give a consumer emotional and functional reasons to buy. A brand can stand for quality or safety, e.g., in the auto industry. But it can also stand for bonding, a family car, or premiumness, a luxury car or courage, high speed. If a brand is only an obliged labelling like a private label brand in a discounter, and for example price is the main purchasing factor, then a brand health tracking is less useful. The brand will not be built to play an important factor in the purchasing process. More tactical research can be sufficient, for example on shelf attention. 

A brand should also have a minimum sized position in order to find sufficient consumers for valid feedback. If the brand is in the development stage, low trial rates, in combination with category building, then a launch tracker is to be preferred. This tool has a focus on category entry points (drivers / barriers) and for brand level on awareness and purchase intention. 

Another factor is that management should be willing to take decisions based upon the outcomes. In other words, the brand health tracking should be used as a management tool. If the management utilizes its own experience and gut feeling as the main source for decision making then a brand heath tracking is not necessary. In this case tactical research as well as sales monitoring (e.g., retail audit tracking) can be sufficient.

The foundation of a good Brand Health Tracking system

A Brand Health Tracking should cover elements to manage the brand. Brand Health Tracking will identify the points of attention for your brand. These might be weak points or thresholds due to the fact that a higher level has been achieved, e.g., sufficient trial has been generated and now the focus should be on more frequent consumption. These points of attentions will lead to define the jobs to be done. A Brand Health Tracking does not measure the performance of your communication in detail. It does not tell which touchpoint is working better and if the communication message has come through. But a Brand Health Tracker tells you if your credentials have been developed, if your brand is sufficient aspirational and if you are well connecting. This will lead to prioritization of your message streams. Brand Health Tracking will also tell you if your advertising is noticed, appreciated and if the brand is developing in the direction of the objectives of your communication campaign. It is like a health check-up at your GP. The doctor can detect if there is a problem with your hip since he noticed it is not so flexible anymore. He will not be able to tell you what is driving this problem. For this you have to go to a specialist. If you want to understand why the ad is not working, you can execute ad hoc research. But there is misunderstanding about this. Heard a few years ago at the Annual Esomar Congress the Global Director of a FMCG Company saying: ”We killed all the Brand Health Trackings, since we have never taken decisions based upon the Brand Health Tracking”. Knowing the boundaries within and how to operate is a prerequisite for a well-functioning Brand Health Tracking.

A Brand Health Tracking covers the following aspects:

• Do consumers know me?
• Have consumers seen me?
• Do consumers like me?
• How often do consumers use me?
• What do consumers think about me?
• What is my overall fitness score (brand health / equity)? 
• Do consumers prefer me over other brands?

These aspects are quite generic. The following are more specific and not always taken into account, especially not if a from-the-shelf agency tracking system is implemented:

• The measurement framework should be consistent with decision making about the way brands are managed. Companies have brand bibles or how to build brand frameworks, which are usually uniformly applied
• The frequency should be in line with the ability to react by the organization. It does not make sense to measure quarterly if your annual budget is fixed and you have hardly means to adapt during the year. We have seen companies measuring on monthly level whereas (semi) annually would be sufficient given their actionability radius. But the management wanted to have a monthly dashboard and for this reason spending an unnecessary amount of money on Brand Health Tracking
• The measurement should be comparable in time as well between different markets / categories. Often comparability is difficult due to a different base of brands in time or between markets. This requires a well thought out solution in advance
• The KPIs should be absolute with clear references and not relative. The fact that the brand scores 105 versus the main competitor does not say anything. The main competitor might be very weak or strong. This is the same as the top 10 best restaurants in town. If we know that several Michelin stars are part of it then the picture is completely different. The Michelin star measurement is so strong and recognized due to the fact that the level of a one-star restaurant is in every country the same. It is based upon the same measurement, rules and guidelines. We have seen brand managers arguing that they still do a very good job since their brand was on par with the main competitor, meanwhile the absolute scores of brands were decreasing. Repairment was needed but was denied
• The KPIs should be simple and easy to understand. Our experience is that constructed KPIs based upon statistical modelling are difficult to understand and are also conveying a seemingly scientific approach
• The KPIs should be linked with the brand development stage. For different brands and especially brand sizes (established brand versus upcoming new brand) focus could be on different aspects, e.g. on awareness for smaller brands, whereas recall and associations are more important for long-serving A-brands.

Brand Health is not the same as communication evaluation

Brand Health is a reflection of all the brand signals in the market. Brand messages are communicated via different marketing activities (advertising, sponsoring etc.)  but also via retailers. The last decade a lot of FMCG brands have become victims of retailers by more frequent and increasing price promotions. These price promotions have a diluting impact on the Brand Health. A marketer was once disputing the validity a Brand Health Tracking. For one of the flagship brands the company had never invested as much as in that specific year (amongst others football championship, Olympic Games) and all the marketing activities (primarily ads) perform very well in testing. Then the question how consumers perceive all these intense and frequent price promotions in retail could not be answered.

Brand Health is not the same as in-market performance

A strong brand, with a high overall fitness score, has compared with competition two advantages. The brand can potentially demand a higher price and attract more consumers. This will generate more revenue for the company. This is the case if the situation differs not too much from competition. We were asked by senior management after a market visit: Our premium brand has the highest brand health score but is not gaining in volume anymore, meanwhile our second brand is still growing. There must be something wrong with the Brand Heath Score measurement. The reason was that the pricing of the premium brand was only affordable for a rather small group of consumers.

Brand Health in FMCG can in general not be measured from social media

A few years ago, a Director Procurement said: “Why should we spend money on brand health tracking if everything is available for free online?”. We did a pilot in the beverages industry.  The results were:

• The category is not a key topic to be discussed in general
• The category and brands are mentioned as part of the social experience. We are in this bar and just drank this brand
• It is mentioned more often when consumers have a negative experience 
• Brands are more often mentioned when referred to the respective marketing activities
• Awareness, usage and preference are too difficult to measure via social media
• For limited brands results could be measured for image or personality items and could potentially be used as an early warning system.

The above will differ per category, but overall we think you cannot measure Brand Health from social media. Social media are very useful, though, to obtain conversations about new topics related to the category (brands) not known upfront.

Brand Health Tracking should be used as a management tool

Brands have to be managed and a tracking can be supportive. Managing means analysing the situation and diagnose the weak and strong points. This analysis should also be done detailed on every brand, but also holistic from a portfolio point of view. The latter could imply that the roles the brands are playing have to be adjusted.  Anyhow, this process should lead to prioritizing marketing activities for the next period (year) and linking them with measurable KPIs in the brand health tracking. Consecutively the situation is repetitive. This way you can see if really progress has been made and which elements are lagging behind. In other words: where a good job has been done and where there is still a job to be done.  Key metrics can be used as input for the yearly appraisal of Brand Managers. Really good brand managers dare to be vulnerable. But we have seen too often that managers try to avoid that their careers are hindered by an unpleasant metric from the Brand Health Tracking. We have seen Brand Managers cherry picking as well as presenting high relative growth figures on small metrics, like 400% growth in awareness. This will probably never change, but should be avoided.

Brand Health Tracking should be focused; not a receptacle for all kinds of initiatives

As mentioned in the previous paragraph brand health tracking should be used as a management tool. This means that it should measure what is needed to manage brands. This measurement should be focused on the information which is needed to take decisions.  We have seen too often that over the years a Brand Health Tracking has become a vehicle in which various kinds of adjacent initiatives are thrown. In our experience that is driven primarily by two factors:

1. Companies include as much as possible, also about category understanding. If you want to know drivers and barriers from your category versus competitive categories, as well moments and occasions then it might be better to conduct a focused category understanding study
2. Brand managers want to measure also project related elements. This can be detailed consumption or perceptions of some new marketing activities

Consecutively, these temporary questions are not deleted anymore and stay in the questionnaire for ages. We have seen situations that nobody knew anymore why and for whom these questions had been incorporated. 

On the other side, if you have one or two simple questions and approaching the consumer base separately is very costly, why not include if this is feasible? But too often there is no more space available and yet these questions are added. According to us you can very easy solve this. Reserve in advance space for only a few ad hoc questions but with a clear limit. This gives the opportunity to include some ad hoc question and the limitation is aligned with the rest of the organization.

Brand Health should be integrated with other data sources and reporting using augmented analytics techniques

Brand Health tracking is often analysed in isolation, whereas brand health is the output of marketing and sales activities. A holistic view on the brand makes integration of several sources like retail audit tracing, household panels, media touchpoint measures like reach, GRPs, frequency, expenses, clicks, purchase funnel metrics, conversion rates and social media indispensable. This enables to analyse the Brand Health in sales and marketing context.

Another phenomenon we see is the increased usage of augmented analytics techniques the coming years. Automated analysis will be possible by making use of AI (Artificial Intelligence) and ML (Machine Learning) techniques.  Combined with natural language processing (NLP) dynamic and autogenerated data stories can be delivered.  This will generate enormous speed and enables faster acting on the outcomes. Natural Language Query (NLQ) technologies, asking a question such as “could you give an overview of our brand X versus competitive brand Y”, will facilitate a much easier and faster navigation.

What are the implications for the skills and capabilities for future Insights Managers?

The above has repercussions for the skills and capabilities of an Insights Manager. 

1. Firstly, he/she has to be a business partner for the decision makers, usually marketing. 

1. The Insights Manager has to steer the ICT department and in collaboration find the most effective way to integrate Brand Health Tracking into the internal BI platform. This means that the Brand Health Tracking has to be accessible in an internal BI software platform. However, to integrate survey data, including Brand Health Tracking data, is often underestimated – especially by ICT.  You have to take into account the specific character of survey and marketing information data sources. This data is different from the usual enterprise data like financial or sales data. The completeness of data is not a given nor controllable. As opposed to financial or sales data, where figures are always available.  For specific survey data this has more complications:

• The data is based upon answers by respondent within a sample. Sample data usually have to be weighted to present a representative picture. This implies that every respondent receives based upon his characteristics a certain weight factor. Often a specific group has been “boosted” which has repercussions for the weighting
• Respondents are not always answering all the questions. This implies you have to deal with so called ‘missing values’
• The same surveys with a continuous character conducted in multiple countries very often have different periods of measurement.  E.g. a multi-national company wants to know how their brands are evolving on certain consumer metrics (like awareness, usage and image perception) and is running a global brand health tracking program. One market will do this on weekly level, another monthly, another quarterly and smaller markets yearly. However, you need to have one dashboard in which the global view of brands is presented. This means you have to deal with different time periods. In these dashboards you want easily navigate and quickly drill down to do some deep dives. This implies that your data from different markets, with different time periods with different user groups has be pre-structured already
• The integration of survey data (based upon respondents) with different marketing sources which have a more aggregated level (like media expenses, GRPs, retails audit sales) is another complicating factor. This is even more striking when you have to integrate different marketing information suppliers with deviating measurement and user group definitions from different markets.

This means the Insights Manager has to steer the ICT department and in collaboration find the most effective way to integrate Brand Health Tracking into the internal BI platform. We have seen several problems by treating Brand Health Tracking in the same way as other data sources  and to access the data with a BI tool without taking into account all the necessary data preparations. 

1. Insights managers have to become analytics translators and master sufficiently machine learning algorithms in order to steer data scientists. The increased usage of augmented analytics techniques the coming years has also implications for the skills set of the Insights Manager. Usually, the knowledge about machine learning within companies’ Insights Departments is limited.  Insights managers do not know the possibilities of machine learning and have difficulties in understanding the results. The majority of these analyses are performed with machine learning tools by data scientists. Data scientists are needed to build the analysis models. Data scientists can analyse but are usually less strong in the interpretation and the link with the business context. Insights managers are not able to steer and challenge these data scientists.

McKinsey has also recognized this and according to them there is a huge need for analytics translators, the link between data science and the business. This is described by McKinsey in the article “Analytics translator: The new must-have role” for the Harvard Business Review. According McKinsey translators work closely with the business leaders to understand the business issues and prioritization. Translators then use their knowledge of AI and analytics to translate the business issue into a briefing for the data professionals. These data professionals produce models and solutions. The translators then ensure that insights are distilled from these models and solutions that the business can understand and the business can act on. 

Domain knowledge is the most important skill that a translator must have. They must be experts in their sector / discipline and company to place the value of AI and analytics in the business context. They do not necessarily have to be able to build quantitative models themselves. They have to know which model variants are available (e.g. deep learning versus logistic regression) and for which business problem they can be applied. Translators must be able to interpret the results, discover potential model errors, such as overfitting, and challenge the data scientists. 

In other words: Insights managers have to become analytics translators and master sufficiently machine learning algorithms in order to steer data scientists.

What are the pitfalls of failure of a future Brand Health Tracking System?

Some of the pitfalls will not be different from current situations.  

The resistance of local Insights Managers can be huge, in the case of a multi-national approach. We have oftentimes seen that in a set-up stage these Insights Managers want to keep their own ‘babies’. My country is different from the rest, are the words we’ve heard quite often. But also when the system is running, local Insights Managers might want to change some metrics or a way of measurement. They might use more senior management to push their ideas. This requires that the responsible Brand Health Tracking manager needs to have an antenna and sufficient alignment to detect this at an early stage as well to build on relations with the local managers (stakeholder management).

Marketing managers want to show good results to the rest of the organisation. If the results are not in line with their expectations and might even declassify their work, then they might try to kill the system. We have given examples before.

In our view, the biggest pitfall is if the Insights Managers are not prepared to acquire both sufficient BI tools knowledge and the necessary AI/ML capabilities. The software platform will not live up to expectations and users may walk away. In the event that the augmented analysis does not get off the ground, users may start complaining about the speed of the analysis. Especially if this is properly arranged by other disciplines.

Conclusion

We think that strong brands are becoming much more important. In the future a sound Brand Health Tracking will be indispensable to steer these brands. A sound Brand Health Tracking needs a proper set up and usage. Company Insights Managers need to steer this process. They ought to think thoroughly about the set up and design the best framework for their business situation.

But it requires also that the Insights Manager is able to partner with ICT concerning the integration of data sources. In order to embed augmented analytics techniques, the insight manager will need to obtain an understanding of data science / machine learning. Should the Insight Manager not be able to cope with the ICT / machine learning aspects, others will take over. 

The foundation of a Brand Health Tracking should cover elements to manage the brand. Generic aspects are described and concrete guidelines are given. Brand Health Trackings should be used as management tools and ought to lead to prioritizing marketing activities and linking them with measurable KPIs.

Last, but not least: Brand Health Tracking is not the same as:

• Communication evaluation
• In-market performance

And can, generally, not be measured from social media. 

REFERENCES

Burggrave, C.R. (2018), Marketing is Finance is Business. Milton Keynes, Lightning Source

Henke, N., Levine, J.,  McInerney, P. (2018), You Don’t Have to Be a Data Scientist to Fill This Must-Have Analytics Role. Boston, Harvard Business Review

Koornstra, S.O. (2020), Software Review: Data integration tools with data analytics, visualization and dashboarding. Amsterdam, MOA publication