Turn and Face the Strange Data Changes

It’s now a cliche that data changes every field it touches. But how? It might be that it just makes the business better, maybe more scalable. But there are lots of cognitive and cultural effects. Here are some basic ones that I've observed over many years in the field:

1. We’re not as smart as we think

The powerful and limiting effects of heuristics and biases are now clear (https://en.wikipedia.org/wiki/List_of_cognitive_biases). We all make up stories, tend to see evidence that confirms a view and operate as though our ideas are proven. But they aren’t. Who adapted better in world war 2, boys from the farm or boys from the city? Conventional thought usually selects the former though data showed the opposite. Data shows that our thinking can be really flawed, wrong, even completely false. Every industry has its beliefs about what is and what is possible. Plans and ideas like in war often do not survive the first encounter with reality.

2. The world often may not be what we think it is

Galileo’s use of the telescope quickly confounded long standing views about the universe. Far from being at earthly scale and perfect, the universe was revealed as vast and complicated. It’s also worth bearing in mind that Galileo had to defend what he saw. The data weren’t just accepted. It wasn’t as some claimed from the imperfections of the device or distortions from something meteorological. People love to talk about how open they are to new ideas. But that’s hardly the typical case. Data can unmask old views and this may not be appreciated.

3. Little things can matter

In the early days of consumer websites, developers started noticing that seemingly small changes could have big effects on consumer behavior. Sometimes this could be a change to a process (like checkout) and sometimes it could be even more superficial, like changing a color scheme. The effects could be quite sharp and quite fast. The fact that we now had detailed data about consumer experience meant that we could see what effects even small changes had. We found out that we could be really surprised.

4. And we might not know why

And stuck without an explanation. We want the world to be governed by clear, clean rules and patterns to be explainable. Data can reveal a world to us that isn’t so cleanly parsed. Sometimes it’s just the way things are. We can make up explanations and rationalize ex post facto but sometimes we don’t know and, worse, can’t know. Data does not necessarily provide an automatic answer. Blue just worked better....

5. The world might be simpler than you think

The sad path to consumer credit default is often experienced as a singular contingent fate. Looking at many instances shows that these paths actually consist of a few types and the patterns in the payments are very clear tells. The pain is personal but from the emotion-free data perspective the picture isn’t complicated.

6. On the other hand, the world may be more complicated (and interesting) than you think

This was a Galilean insight. The story gets repeated down here on earth. Start collecting data and you might find that there are more species than you had ever imagined. Start collecting data and you might find all kinds of different uses of your product or service and different interests driving those. Sometimes these might be of epic importance. Many years ago, a bunch of quirky folks showed up at Schwinn with some requests to help them with their sport. They had been using a particular Schwinn bike to do a kind of riding in the hills of Marin California. They were dismissed and so they went home to found the mountain biking industry. Schwinn missed that the world of bicycling was much more varied than they could imagine.

7. People aren’t as evidenced-based as they think they are (nor as optimal) They are biased about themselves.

Most doctors would say that their practices are driven by the science. But studies show all kinds of variation driven by a number of other factors. Doctors don’t practice like robots, outputting treatments based on data. Sometimes the data doesn’t even exist. Do national databases with every treatment and outcome for a given condition exist? They are starting to as interest grows in personalized medicine. It’s often surprising how much people rely on ideas as opposed to data. They celebrate intuition and then get upended by honest empirical analysis. Data exposes and this might not always be welcome (like in Moneyball).

8. Data can lead to new ways of doing things

Pharmaceutical companies were early in the game of trying to figure out doctor prescription behavior. They discovered that while reps had a strong effect (and spent a lot of money optimizing that), they discovered that patients drove some of the behavior. The direct-to-consumer advertising industry, for good or for ill, was born. Push turned into pull. Sometimes the data perspective creates the actual product. Pandora learned individual tastes based on a famous model of music, and then packaged that as personalized radio. Data can mean new products/services and business models.

9. Data affects authority

In the pre-data era, company decisions were dominated by position. Truth moved down through the hierarchy. Data can make fools of would-be kings. It turns out that while thinking is good, even incredibly smart people can be wrong, or don’t have the idea that matters. Data can be a blow to human ego.

10. Data affects the culture

Adopting the data transformation forces a new culture. From relying on the insights of great individuals we shift to a focus on experiments. Failure is not profane. Insight evolves. We think and interact in smaller increments. The shift to agile is a significant cultural shift.

Recent Posts

See All

We’re going to use this space for short takes on current issues in AI that we believe are relevant to our customers. AI is a big, complicated topic and the term has become confusing. AI is much mo

The COVID pandemic has exposed and exacerbated deep social issues, including our failure to teach algebra. But what does algebra have to do with pandemics? The answer lies in the difficulties of getti

As AI techniques for modeling big data (especially deep learning) have exploded in interest and use, concerns over bias and negative impact have been raised. There are many articles and many general