How Our Data Failed Us, And What We Should Do To Fix It.
Last week, I discussed how our world’s data is for all intents and purposes…dead. The business world, the sports world, social science, economics, and just about every other category you can think of that makes decisions based on an understanding of what has happened in the past coupled with an understanding of what is happening at the current time. These two sets of inputs are then used to make a prediction of what will happen in the near future. As I argued, the better the inputs and understanding of these two data categories, the better the predictions of future events will be.
Right now neither of these inputs exist in any meaningful way, so we are left adrift without any reliable way to understand even the immediate future. The demise of data has been hastened by the COVID-19 public health and economic crisis, but the hard truth is it was already on its last legs.
The world is vastly more complex today than it was even a decade ago. As the internet has connected nearly all of us to news and information, opinion, shopping, education, and more, our options have grown exponentially. At the same time, we have isolated ourselves even more and the vast majority live in a bubble of like-minded people and organizations. These bubbles are not just political in nature, but also consumer-based, generational, as well as non-political in thinking. Meanwhile, we are providing more personal data at a level never seen before.
The most useful, predictive, and valuable data generated from this extraordinary complexity is generally proprietary — owned and utilized by the private companies we give it to in exchange for access to our online world. Facebook, Google, Apple, Amazon, cell phone companies, and other big data firms (that are not household names) know more about us than even we as individuals could know about ourselves. For example, in the context of the COVID-19 crisis, there’s evidence that what and how we search for symptoms is not only predictive of outbreaks well before they happen, but of the prevalence of some symptoms before there’s enough public health data for the medical professionals to understand them as such.
Not all that long ago, the consumer world was a tight-knit captive audience. We had three TV stations to watch, a newspaper or two to choose from, and if a car commute was part of the routine then radio ads. But all of these points of contact happened at very specific times throughout the day and were very predictable. If Company X ran a television ad campaign in a market and sales went up in that market compared to a similar market with no ads, it was easy to figure out the effectiveness of the campaign.
Clearly, the world today is different in almost every way. Information is consumed constantly, across more channels than we can count, at a pace that was previously unimaginable, and across networks that are not entirely visible even to the massive tech companies with access to the best, fastest, and most complete data sets.
Consumer behavior now changes day-to-day, even hour-to-hour, in response to the constant barrage of incoming ads, news, social sharing, etc. The old (and too often still in effect) data model of Company X advertises, the consumer experiences the ads, the consumer buys, then let’s ask the consumer why they made that choice just doesn’t make sense in our current world.
Consider examples like being told at the grocery store that you can take a survey to save $20 on your next visit, or getting a survey two days after your hotel stay asking if you are likely to come again or recommend the hotel to a friend. You probably don’t have to think hard about those examples because they’re so commonplace — this is how we do consumer data. Still.
In a world where everything can change over the course of 48 hours, this traditional approach to consumer data makes no sense — it puts the consumer in the position of fitting a brand’s narrative and timeline, rather than giving consumers opportunities to provide new insights. The traditional point of purchase approach is anchored by a single point in time response, at a single point in the consumer relationship *after* the consumer has made decisions. Our world has changed but shockingly, our consumer research and understanding are based on a world of the past. This is mind-boggling to us.
Beyond the pacing challenges, the traditional approach fails us in another significant way as we navigate the global COVID-19 crisis: whole swaths of consumer behavior have changed dramatically and quickly. Whether you look at what people are buying (i.e. sales of baking yeast were up by 457% in the last week of March) or how they’re buying it (increasingly online, for delivery) the patterns have changed, in ways both expected and unexpected, and with unknowable staying power. The “point of purchase” relationship is transformed — possibly permanently.
Companies have lost their main avenue for receiving feedback and this is a hindrance to every employee within any company trying to make sound decisions in these difficult times.
Our approach to data should not be based on proving ROI. Our data should be focused on learning what we can do to improve ROI. This means creating a relationship with the consumer and placing them at the center, sharing information out. If you do not have a meaningful relationship with your customer that is based on sharing and understanding, you are sitting around right now with little to no information. And if we’re being honest, you always were.
This relationship needs to be built over time and that means engaging with the customer at multiple points, not just when they have given you money. This not only deepens understanding and trust but is also an insurance policy for the next time our data dies.
Companies should have adapted to our new world long before this pandemic, but companies who do not adapt now are setting themselves up for failure and are not interacting with their customers in a way that is mutually beneficial or helpful — and the data that results is highly unlikely to be useful.
There’s no time like today to do what we should have done last year: update our systems for collecting, understanding, and using data to reflect the reality of how our modern world works, during and after this crisis.