I first talked about my definition of a Data Product in Episode 105 of my podcast, because I didn't want to be guilty of talking about a concept for years without actually defining what I mean.

I consider my definition the "producty" one, one which focuses less on the technical output and asset being made, and rather, on the innate value and benefits the output enables. Myself, and folks in our DPLC community, tend to also believe that producty data product salso leverage the methods of software product management and UX design as well—and that this is what enables them to create value worth paying for.

Data Product: Defined but Evolving

Like with code, sometimes seeing the evolution and change log is helpful. Products are also about getting started, shipping it, getting feedback. So, I'm having fun doing that with this evolving definition. Most recently, in v3, I felt that while one could argue many GenAI solutions are still about decision support, I don't think it fully captures the spirit of what GenAI can do. So, I've included "idea generation" here. Notably, I also reverted back to my v1 definition as the basis; that's based on feedback from others, and what I find myself saying when people ask.  I also mention "exchange of value" because in internal data products, literal money may not be exchanged. This is because the status quo may seem least risky to users of internal data products. That hypothetical sales or marketing team may prefer doing things "the old way" without an AI tool that the internal data science team thinks is "good for them." When those business teams are willing to let go of their old way, and take a chance on a new application/UX, they are giving up something of value: the safe status quo.

Here's the latest definition:

v3. 1/1/24 A data product is a data-driven, end-to-end, human-in-the-loop decision support or idea generation solution that’s so valuable that customers would pay for it or exchange something of value to use it.

v2. 7/21/23 A data product is a managed, end-to-end software solution that organizes, refines, or transforms data to solve a problem that's so important customers would pay for it or exchange something of value to use it.

v1. 11/29/22 A data product is a data-driven end-to-end, human-in-the-loop decision support solution that’s so valuable, users will pay to use it if they have to.”*

*Version 1 of these definitions was first contemplated in my deep-dive solo podcast episode #105.

Data products vs. data-as-a-product

Data as a product is much more about a partially refined data asset that in its raw form, may not be particularly useful to an actual end user or stakeholder—but when refined into a product or UX, adds value. An example would be something like buying a mailing list of qualified "prospects." The raw CSV file that might be the delivery format is not particularly useful. It only becomes useful once the buyer can take the data and use it in the desired way to get the benefit they seek. I typically think of "data-as-a-product" as a bundle of information that is for sale, in a mostly raw or partially refined form. It is, however, typically not very complex, and may not need the "product management" treatment particularly if it is delivered in a "once and done" form.

Photo by Suresh Designer on Unsplash