Got a lot of data? Fantastic!
Got a good idea for a product, too?
Many companies like the idea of making use of the data they have. However, without a clear application in mind, data is just thin air.
One of the stories that I hear a lot in business is this:
Company X: „We already have a ton of data. Now we just have to use it properly, and then we can…
- make tons of money
- know better where we are headed as a company / predict the future
- finally, find out why one sock always goes missing in the laundry
Now, while it is generally commendable wanting to make more out of the data one already has, it is helpful to keep one key thing in mind:
1: You need a good idea for a product or service that people find valuable
Many people seem to be under the impression that data in and by itself is valuable. That is certainly not the case. Without a clear application, data is just thin air. So, without a strong idea for a product or a service that is clearly understood as valuable by the targeted group of customers, all the data in the world doesn’t help one bit.
Often I have encountered situations where the hypothesis for the envisioned product or service is weak at best and plenty of resources go into “making the data useful”. I can only suggest turning this process around. Much effort should go into thinking about what one wants to offer before getting started on the time and money intensive part of wrangling the data into shape.
That being said, another issue I frequently encounter is the idea that the data that already exists as a byproduct of business processes can be turned into gold with very little effort. That is wishful thinking as far as I am concerned.
It is similar to a carpenter who builds a beautiful table over the course of a week. At the end of the week, he or she collects all the sawdust, scoops it up, and stuffs it in bags to be sold as raw material for wood oven pellets. Now, he or she will certainly get some money for the sawdust, the byproduct of the work done, but nobody would assume it would net more than the table when he or she sells it.
With data, it is quite similar. If it is the byproduct of an existing process, there is a chance of scooping it up and making some use out of it. But for significant value, data needs to be made — just like the table. So, another thing to keep in mind is:
2: To create extraordinary value from data, you need to create the necessary data first
While it is possible to use data for other purposes than it was intended for originally (like using cell phone data to get an idea of how much people are moving around during the Covid Crisis), nothing ever beats the value of data that is specifically generated for a particular purpose — even if the significance of proxy data for “off-label-use” is oftentimes astonishing. But when it comes to the difference between good and great, specifically created data makes the difference.
Tesla is a good example. Famously working on a self-driving car, the company was often criticised for not using data that already exists, such as map data and navigation system data. But Tesla’s product vision is that the car can navigate in any kind of circumstances, regardless of whether it received previous information about its surroundings or not.
Even the most accurate maps or navigation system data will not reflect all the things that can happen on a road. Let’s say a car 5km ahead of you loses a tire. There is no way it will show up in any of these data sources — yet it needs dealing with when the car reaches the point where the tire is lying.
This is precisely the reason why Tesla cars first create the data they need to interact with their surroundings, analyse it, and take action based on it (steer around the tire or break). Any kind of proxy data would not allow the same performance as the self-driving system.
But even for applications that are less complex and progressive than self-driving, both factors outlined hold: you always need a solid idea and some well-fitted data to build something significant.