Data Is a Resource, Not a Product
Block of text extracted from here:
If I’m selling a product, I don’t care what people are using my product for, and how often they are using it, I just care if they buy it (or not). The market will tell me how valuable it is.
But this isn’t true for data - there is no market for determining value. What people say they want, and how they are using data, are very different from the most valuable use of the data.
Right there is basically my entire point, so if you’re satisfied with that, go ahead and do something else like play tennis or read a book or something.
I wanted an expansion. So here we are.
The point if it wasn’t clear from the above, which I guess it wasn’t sufficiently otherwise you’d be playing tennis right now: data is a resource, not a product.
The first part is about how the product analogy is wrong.
Products exist within markets. Free markets (sort of). Market participants are free to decide what to produce and consume.
Producers create products that they guess consumers will want, preferably in a cheap enough way to generate profit for themselves. Consumers buy products they think are valuable, more valuable than the value of the £££ they are willing to pay to acquire the products. End Introductory Microeconomics Lecture 1 Slides 4-6.
The reason that this system works is that we have a definite way of measuring value. GBP. Producers know (via the Smithian invisible hand) what to produce: things consumers will buy. This is all they need to concern themselves with.
Isn’t the same true of data “products”?
Just give the consumers of the data throughout the business what they want (at a reasonable cost), right?
Turns out this isn’t true. Because: there is no market.
The way in which markets discover what products should come into existence is not present, which is sort of the whole point of markets and the raison d’être of producers.
Without these market dynamics, we require an alternative way of determining value. We can’t just listen to individual data consumers because they have fuzzy ideas of value and are not its ultimate arbiter. That’s the business itself: we should be doing things that create value for the business, not these individual consumers.
This is why it matters how our “consumers” are using our “product”, unlike in real markets. They must be using data effectively, otherwise they can actually destroy biz value.
A better way to conceptualise data at an org is as a resource.
Like oil or lithium or wood.
We have to put in effort to extract/collect data. We then have to process it, and make useful things with it.
Here’s a resource pipeline and the data equivalents:
- Find resources –> find data sources
- Extract resources –> gather data
- Ensure quality of resources –> data quality
- Transport to storage –> extract data and store
- Sort, organise –> transform data
- Standardise –> create organised representations
- Build something –> analyse the data
- Production line –> create a model
This isn’t a particualrly new idea.
You can collect all the wood you want, but unless you do something valuable with it, it’s kind of pointless. Yes, you can still sell the raw material (you can sell data, too) but the real value is in doing something creative and useful and valuable with it.
On a societal level (a company is a mini society), that is. We don’t collect resources unless we have some use for them, unless we can transform them into something valuable.
Analysis is like using wood to build a cool house. You’ve created value out of (processed) raw materials. Well done.
But what’s more valuable is this type of process scaled up and automated. Factories. This is what models are. They take input data (raw materials) and use it to produce valuable output (products) for us all to use. Over and over again.