Attribution… or the lack thereof?

ROI models don't reveal everything. There are nuances that may not get captured.

Attribution in marketing is one of the keenly discussed and deliberated topics, more so with the disruption technology has brought into the landscape. While, on the one hand, it has now become simpler to decode the consumer journey into bits and bytes of actual action from the first click to the last, on the other, it still leaves a lot unexplored. And perhaps, unexplained.

As a practitioner of marketing, with the sole purpose of influencing and impacting consumer behaviour, I reckon two pieces still need to be appropriately fit in this puzzle.

First, it is truly established now that consumer journey and decision-making is fluid, non-linear and seamlessly moves between the online and offline world. Hence, in this omni-channel world, there are as many ‘offline clicks’ as there are online.

For a vast majority of consumers, a touch and feel experience at the store, what the retailer chose to recommend, what that one friend talked about and what that one review blogger wrote in his or her last post matter as much as (perhaps even more) every little or big marketing intervention the brands delivered in his/her decision-making journey. In such a scenario, what and where is attribution?

Of course, there are various marketing ROI models that can statistically input variables and can apportion percentage of consumer action (sales or preference or clicks or queries) to each variable with a weightage. Howsoever statistically robust that might be depending on the quality of input variables we put into the model, there are still nuances that may not get captured.

For distribution led brands, channel or trade marketing inputs play a big role towards tertiary sales uplift along with pricing interventions, and a lot of these may not get adequately represented in a consistent manner in the marketing attribution models.

Often, the models tend to skew towards the more readily available data points of GRPs, impressions and clicks. Non-standardisation of input variables remains a challenge still. Also, in categories where consumer experience happens offline and purchase online, or in some cases vice versa, the distinction between the first and last click blurs.

Second, brand choice is always an emotional decision, often, at a very sub-conscious level. It is these emotions that we as marketers try to get insights on and leverage through our propositions. Therefore, persuasion needs to be looked at as a continuum – a continuum of emotions, from weak to strong and something that can fundamentally be altered through targeted marketing interventions.

It is within this paradigm, softer measures like brand love, trust and advocacy play a pivotal role in making consumers switch from one brand to the other without any apparent rational explanation. The point being, the emotional trigger leading to persuasion (and the action) may have happened at a point in the continuum other than the last or the second last click, that perhaps a time decay model may not even capture. How many times has it happened - that a brand added in the shopping cart has been replaced by another at the last moment because of a memory trigger of a conversation at a social do or the kid’s birthday party?

It is this balance of science and art that makes marketing as a practice so fascinating and honestly, as marketers keeps us engaged and challenged. Attribution is necessary for achieving business growth in the most optimised manner.

However, models can only tell us that much, we need to ensure our frameworks are flexible and open enough. More importantly, we need to be clear on how we are going to implement the learnings from the attribution work, that delivers continuous improvement for our brands and business.

(The author is director marketing, South Asia, Signify Innovations, formerly Philips Lighting India).