Research and Effectiveness

The origin of this brain-dump stemmed from a recent post Prof Mark Ritson, sharing an earlier talk on The 10 Key Factors Driving Advertising Effectiveness - Media, Marketing & Effectiveness. (link - https://vimeo.com/373969781). As usual, Prof Ritson in his distinctive drawl, delivers some very insightful observations, which are definitely worth listening to and more importantly, to then reflect on what needs to change in how we do our respective roles in Marketing everyday. Of all the 10 drivers, I want to focus on just the one, that relates to Research.

There are details in the talk on how this chart was arrived at, using the Effies database - suffice to say, there is a relationship between Research and Effectiveness.

And to quote Prof Ritson "..Thank F***, there is one!" :-)

Not that I ever doubted it, but getting another validation that research is relevant, is always exciting. But with great power, comes great responsibility. As with all database related analyses that show a relationship between 2 factors, it is never a given every single time. There is a high degree of co-relation - so, all the planets still need to align perfectly, meaning there are a lot of underlying factors which still need to be considered. We still need to define for ourselves WHY this is happening and HOW to make sure we make it happen more often.

The WHY is easy - research leads to more informed business decision making, plain and simple. Either by re-framing what we already know or by unlocking new insight, the business decisions taken will be more effective. That is why we exist.

The HOW is more complex - how can we ensure our research is and continues to be a driver of effectiveness? In the chart above, there is also this notion of "too much research", that does not add to anything. So, what does the "right amount" of research look like? What behaviors embraced by Research practitioners, will result in effective research?

As with all things, there will be a constant need to re-appraise how we do what we do, to continue to be effective. I thought it would be interesting to use the construct of Traditional Programming versus Machine Learning, to define how we need to approach Research today, to ensure it continues to be effective.

A lot of what we do in Insights today can be likened to the Traditional Programming approach, embodied by the below equation -

INPUT is any “data” we use – either quantitative data from surveys or qualitative data which we observe from focus-groups, in-depth interviews and the like.

RULES are the sum-total of our existing/expert knowledge – statistical analysis, relationship between different data-points, cultural learning, behavioral-sciences understanding, Neuroscience understanding of how our brains process information and how humans make decisions etc. 

OUTPUT is the insights / implications we derive from applying the RULES to the INPUT and use this to drive business decisions. 

The biggest challenge with this approach, is the constraint imposed by what we know (RULES). And when we live in an era where what we know is constantly being challenged / updated with new information, we need a new operating model.

And we can take inspiration from how Machine Learning managed to leap frog ahead in its adoption and development. So, the pivot to how we do Research / Insights needs to look like the equation below -

INPUT will be hypotheses we need to start all research discussions with.

OUTPUT will be the constant stream of data in ever increasing amounts coming to us in almost real-time.

RULES will be the new learning we will get by either validating or disproving the hypotheses we started with.

And it is not a linear equation with a start to end – rather it is a loop of self-sustaining new learning that constantly feeds as the INPUT for the future. There is a need for constant re-appraisal of what we learnt from before and how new data coming in, changes the reality as we know it.  

It is not easy to affect this shift, because it is not the way we have done things for the longest time and it will require a concerted effort to drive new behaviors like being comfortable with an approach that is forever-beta, based on hypotheses-led learning and more importantly to learn-by-doing. So, what can help with this?

First is a ruthless focus on the business impact achieved by Research. With the scale and speed of data available now, it is even more important that we become super-focused on the fact that all data exists to fuel decision-making and drive business action, and all data needs to be looked through the lens of – what is it in aid of? An article in the HBR, published in May 1985 by Alan R. Andreasen, he talks about "backward market research". (https://hbr.org/1985/05/backward-market-research) - implying that we need to be crystal clear on the destination, ie. business outcome we are looking to impact, and everything should stem from there.

Second, we need to adapt our thinking. There are two types of Human Intelligence, defined in Psychology – Crystallized Intelligence and Fluid Intelligence.

(Source : https://www.thoughtco.com/fluid-crystallized-intelligence-4172807)

Crystallized Intelligence refers to the accumulation of knowledge and skills we learn, throughout our life. Fluid Intelligence involves our ability to identify patterns and relationships and help us think logically and react to novel, un-experienced situations. The Crystallized Intelligence kind of thinking, which is based on rules, learned behavior, quantitative relationships can only take us so far. We increasingly need to dial up the Fluid Intelligence kind of thinking, which relies on an approach that is hypotheses-led, strong focus on experimentation, qualitative inspiration and creativity. 

STAY SAFE, STAY RELEVANT