In my last post, we talked about the growing importance of two concepts that are radically changing consumer marketing practices: The Choice Equation and the Influence Mix.
The influence mix
{ D = f(P+M+O)}, captures the forces behind a consumer purchase decision by looking at her prior preferences and experiences (P), information from marketers (M) and inputs from other people (O).
The Choice Equation on the other hand
{Choice = f(Taste, Influence, Context and Behaviour)} puts the Influence Mix to work, by measuring consumers’ personal tastes and preferences, the influence of others, ,the current context and the past behaviour.
Let’s dig a little deeper into these two great concepts.
There are two big challenges in making these a reality.
- Access to ingredients (data) and
- Availability of a recipe (algorithm)
Take the first challenge. Access to data.
Surprisingly, finding data doesn’t seem to be the challenge. Today, the amount of data available on each of the above factors is mind-boggling.
For instance:
- Our tastes are captured quite well in the vast digital footprint each of us leaves behind. We rate and review everything in our lives, from food, hotels and electronics, to movies and books, even taxi drivers! We leave a trail of crumbs from every slice of our behaviour, on sites like Amazon, Trip Advisor, Yelp, etc, not to mention new review sites that keep popping up.
- Likewise, the influences are measured in the currency of likes and retweets, on social media platforms like Twitter and Facebook. These are increasingly becoming the source of consumer discovery and propagation of consumer sentiment.
- Context refers to the effect of immediate events on our decisions. This includes location, day of week, time of day, weather, etc. The astonishing growth of sensors and their increasing fidelity means that we can track each of these context variables down, to micro-levels of distance and time.
- Finally, data on our behaviour, both online and offline, is enormous. It is readily available to both online giants and the vast number of traditional enterprises for whom we are consumers (banks, retailers, hotels, telecom operators, restaurants, health care providers, etc.)
But here’s the catch. Data on each aspect of choice is scattered over the internet. The process of collating all this data, and making sense of it, is entirely up to the consumer.
Example: Research shows that we make an average of 10 searches and read 6-12 reviews before we can choose a hotel.
That we spend an average of 27 minutes to find an offer or deal and spend 3.5 hours a month searching for content (say, movies) to watch.
Each decision, each choice takes an average of 45 minutes. In every category.
For example :
If I want to go to Bali for the weekend, I have thousands of travel sites swamping me with options for flights to take and places to stay, hundreds of events, attractions, restaurants, places to shop at, and thousands of reviews and recommendations on each, many conflicting. Asking friends is fun, but everyone has their own deeply passionate and completely divergent opinions.The process of choosing what to do on a holiday turns into misery.
So, it’s not a data problem at all. In fact, quite the opposite. There is too much data everywhere today. And in fact, the abundance of data hinders consumers by providing too many choices.
To extend the cooking metaphor, we can safely say that we have too many types of ingredients and far too many variations of each ingredient.
The second and bigger challenge we face when confronted with so many choices, is the lack of a recipe.
Here is an illustration of a choice sequence that will sound familiar to many travelers.
What we need is a recipe, something that can put together all these discrete data points and come up with an answer that suits us as an individual. In other words, a recipe (in the form of an algorithm).
The choice equation is exactly that: a recipe.
It allows us to create a mix of ingredients to determine what drives choice for each category; and within each category, for each consumer.
Let’s take a practical example of the Choice Equation at work. The choice of a movie.
Sometimes, I choose what to watch based on:
- My tastes (e.g., I love Quentin Tarantino, I find Gwyneth Paltrow annoying, Iranian directors make sensitive movies)
- On other’s views (IMDB and Rotten Tomato both rate this highly)
- Behaviour (I saw the latest Avengers with my kids last week, I want a break from action heroes)
- Context (it’s Friday evening, it’s been a tough week, and I want to relax)
The Choice Equation is a great recipe that allows us to combine all these ingredients.
Now that we have these two great ideas, imagine if we had a system that could:
- Bring together all these data;
- Map the connections and affinities between any two choices;
- Weigh the relative importance of each element appropriately;
- Adjust this for each individual’s preferences; and
- Continually learn about these preferences and fine-tune them based on context
We could transform the misery of choosing into the magic of choice. We could do to choice, what Google did to search.
This is nirvana for consumers.
It shifts the cognitive load onto an algorithm designed specifically to understand one’s choice. It plots all the factors we’ve discussed onto a personalised taste fabric, predicts one’s likely consideration set of choices using this fabric, and filters all of these based on one’s past behaviour and current context.
The millions of searches, the sorting through 100s of conflicting reviews, the asking around from friends, will become one ‘personalised choice query’. That gives you 4-6 choices, personalised to your tastes. 45 minutes of misery becomes 5 mins of magic.
The consumer may still choose to spend more time on those 4-6 choices. But it will be spent adding her intuition and gut feel. Not on resolving conflicts in data. Or in trying to figure out the relative weight of each opinion on her decision. In other words, she might explore the magic of choice, not thrash around in its misery.
It is also nirvana for marketers.
Armed with this equation, marketers need not rely on limited data sets, gut feel, use of price promotions, etc. They can predict more accurately and more holistically, a small range of choices that any consumer might prefer.
All this, of course, is extremely complex. Both, from the perspective of the consumer and from a technological perspective. In my next post, I will detail a solution that works at scale and in real-time. In my 4th post, I will talk about what we have learnt about preference ladders and choices. In the final one, I will talk about why the marketer’s obsessions with offers are distracting us from the real problem of solving choices.
So stay tuned!
