“How can I explain more from the baseline? How can I compare apples and apples in my modelling? How can I see the consumer reaction to my activity?” These are some of the questions I‘ve heard clients asking about modelling over the years.
When we first set up MESH Experience, we believed that capturing every single brand encounter that someone experienced would help us to unpick exactly how brand perceptions were formed. Jeremy Bullmore has a wonderful quote: “People build brands, like birds build nests, from scraps and straws they chance upon.” If we could capture the “scraps and straws” and understand people’s brand perceptions before and after the experience capture, we could see the experiences that drive brand metrics.
To solve this challenge, we created an approach we call Real-time Experience Tracking (RET), where participants report every time they see, hear or experience a brand or one of its competitors in a mobile diary. By measuring experiences, the whole customer experience, in real-time, in the context of the real world, we’ve been able to help our clients to make quicker and more informed decisions about their marketing investment.
Professor Hugh Wilson and Professor Emma Macdonald spotted the power of the approach over 10 years ago and worked with us to create an analytics suite. Their work was published in Harvard Business Review Review – read here.
Every touchpoint counts in marketing mix modelling
Now I am delighted to reveal how the use of MESH experience data improves the insight from advanced marketing mix modelling, as demonstrated in a recent publication by Dr Peter Cain of Marketscience in the International Journal of Research in Marketing (IJRM) – read here.
With the advent of digital media, modern marketing mix modelling typically focuses on consumer journey models of demand, where consumers follow a sequence of touchpoints to off and online product purchase. It is important to understand every single touchpoint along the journey, from seeing a TV ad to spotting the brand in-store or speaking to a friend about it. However, it is difficult to find a data source that includes every touchpoint. This is where the RET data comes in!
Working with Peter on a marketing mix model for a long-standing client where we had years of continuous real-time experience tracking data available to us, Peter demonstrated how this data source could turbo-charge the model.
How MESH’s experience data can turbocharge marketing mix modelling
- Creates consistent media metrics for better comparisons
Experience data enables the creation of GRP-type metrics for marketing variables expressed in differing units, thereby allowing for more robust comparisons. For example, TV advertising is typically expressed in GRPs, whereas paid search is expressed in clicks or impressions. This complicates an apples-to-apples comparison. Experience data can help solve this problem. We know how many people have “noticed” an in-store display in the same way as they have “noticed” a TV ad or a poster. We also know how positive, persuasive and relevant the encounter was. Furthermore, we can look at the share of TV experiences, the share of store experiences or consumption experiences or any other experiences, to gauge performance relative to competitors. We have found overall Share of Experience to correlate more strongly than Share of Voice with Market Share. Therefore, we can measure quantity and quality metrics across every single experience (100+) for the client and their competitors.
- Facilitates measurement of marketing tactics
Experience data allows us to adjust the quality and reach (noticeability) data to enable sufficient variation for robust time series modelling. For example, the role of in-store product display is often a constant presence lasting many months creating a clear measurement challenge in marketing response analysis. In-store customer experience can provide necessary variation, where constant presence is combined with the quality of the encounter and the degree of customer interaction. For instance, we might find that in one retailer the store display is noticed much more in April than in October and that the response is more positive. We might also see that a competitor received a better response over a summer promotional period. With each experience recorded, participants provide a comment describing the experience and why it elicited the response it did from them. In some instances, they provide a photo of the experience too. This can help us to explain the “why” behind the “what”. We can understand why the competitor’s activity has performed so well over the summer months and why our client’s activity was working in Spring and Fall.
- Helps to explain long-term purchase patterns
A key part of the marketing mix model is an estimate of base sales. This represents the long-run or trend component of the data, reflecting any persistent changes in consumer brand loyalty. Conventional mix models typically assume a fixed or deterministic baseline. As such, estimation of customer loyalty and the long-term impact of marketing is precluded by construction. In the Marketscience approach, base sales are modelled as an evolving time-varying process in tandem with short-term effects. This enables quantification of the drivers of base sales, providing a deeper understanding of the mechanics of brand-building. Customer product experience can play a critical role in this process, helping to explain baseline evolution as part of the earned media experience. The experience data can help to explain more of the activity that would typically be left in the baseline. For example, if we have a detailed consumer response to in-store activity or PR activity, this helps us to understand its impact in the model both short and longer-term.
In today’s world, we have a dizzying array of data. We have data lakes and data dumps. But not all data are created equal! We need to mine for the gems among the debris and to include these gems in our modelling. Looking through the customer’s eyes to experience our brands as they do, in the real world, provides a rich data source that can help to explain the behaviors and outcomes we observe. A big thank you to Peter for taking the time to write up this study and to explain how real-time experience tracking data can form such a valuable input to the modelling process.
Contributors: Fiona Blades, President & Chief Experience Officer, MESH Experience