Data: the “so what” is in the “for what”
One of the difficulties in evaluating data projects is in understanding the value they bring. Other IT endeavors seem much more tangible: the return on investment on new infrastructure, or integrated CRM, is significantly easier to comprehend.
Data projects can suffer from the “build it and they will come” storyline. Unless you have a visionary leader who can see through to the results, it’s difficult to swallow when the investment is so large. Indeed, as I wrote four years ago the business plan can look something like:
1. Get data
Working alongside colleagues at SVDS in the last four years, I’ve tried to figure out how to explain what step 2 is to people, and debug those question marks.
You might be tempted to shrug this off as an early adopter problem. Once the value of data is pervasively understood, there’ll be no problem getting data investments green-lit. Of course we need our own AI!
However, keying investment off mainstream acceptance is still a bad idea. It doesn’t explain “step 2” for you. As a result, we see a sad raft of big data projects that now sit around idle, as they don’t have a corresponding use in the business.
Data isn’t software
It seems obvious that data isn’t software. But to both the technical outsider and insider it’s easy to confuse. After all, you need to deploy software to manipulate data. The outsider thinks “it’s all computers”, and the insider thinks “oh, that’s another software project.” But the software isn’t the asset here. For example, a clothing manufacturer is not judged by the asset value of its factories.
As a raw resource, data is about as compelling as looking at a rock of ore. The only way you understand its value is to have figured out how to refine it and put it to use. As a consequence, whenever you ask a question about data, you need to add “for what” at the end.
- How much sales data should we retain? … for what?
- Is real time data access needed? … for what?
- Is our data clean enough? … for what?
And of course, the kicker, “how much is our data worth?”
For recommending new purchases? For reducing our warehoused inventory? For equipping our field technicians?
So, there’s really an internal market for data, where its value is based on the potential application. All of which means that the right place to start isn’t the data itself, but the “for whats”, and then figure out what you need to invest in to make those happen. The value to the business is in achieving those goals. Assessing the value of your data and associated infrastructure can best be understood in those terms: it doesn’t have a lot of relevance as a standalone figure.
This is why I spend a lot of time talking in public on the topic of data strategy, a structured way to figure all this out. The most important takeaway from the data strategy tutorials we teach is always this: figure out your for-whats before getting started with anything else.