Data Platforms Often Aren't a Single Product

A cursory search of LinkedIn suggests there are tens of thousands of people, open roles and companies invested in data platforms. The concept of a data platform product manager has gained a lot of traction and interest in the last 2 years, to the point that it is not uncommon to see a “platform product manager” role, many of which are specific to data platform product. This brings up one of the most confusing things in technology - defining exactly what a platform is (and is not).

Think about your own company for a second (or any you’ve worked at). I bet the word platform is used fairly often to describe data storage, transformation, metric definitions, experimentation and a host of other “capabilities” that businesses use to operate. The word platform makes things seem simpler, more put together and cleaner than they would if we actually had to describe what was going on in such a platform.

As I was working on a job description for a company that asked for help in writing a platform product role definition, I began to suspect I had fallen into the same “simplicity” trap that many do when they use the word platform. This platform product manager role was not in fact managing one product, but instead the 4-5 separate products that comprised the platform. It was only once I dug into that level of detail that I felt I could fully describe the role.

Bringing this back to our day to day - why is it helpful to imagine that a platform is in most cases a set of sub-products? I think its because we often define a platform at the level we define a capability. Businesses need to experiment, they need to manage metrics, they need to provide data visualization, and they need data governance. So we make platforms that support each of those capabilities. The reality is that those capabilities only come from have multiple products that support multiple audiences within a business.

Here’s why I find this helpful. To support the customers within a business we can’t just say “here’s a platform”. We need to enable the customer to use a specific product (e.g. the ability to define or investigate a metric) even if we think about the capability (manage metrics) as a platform. As a data platform product manager, it is really critical that we not fall into the trap of thinking about our platform as a single product. How do we do that? It starts with the customer, rather than the product.

Let’s take an analytics platform as an example. Customer across the business - analysts, marketers, data engineers, executives, and many more have “analytics needs” which ostensibly are served by an analytics platform. Even if you are reading this saying “that’s in theory only”, hang with me for a second. As we define roadmaps, priorities and investments, it doesn’t work to describe the evolution of an analytics platform as a whole. That’s because individual customers have a harder time understanding what they get out of it. Instead, I think it is productive to describe to each customer a roadmap for changes/updates to the sub-product with which they interact. In an analytics platform, this could vary from SQL code to metrics, to data visualization to data monitoring.

The idea of a platform is really attractive but it can mask the complexity of the underlying investments and make it harder for the customer to understand how they benefit from said “platform”. If you have these data platform product roles, are thinking about investing in them, or are simply trying to wrap your head around the idea, I think its worth asking “if we couldn’t use the word platform here, how would we describe the products?” It can help add clarity and specificity and help make priorities and planning clearer.

Thanks so much for reading and as always, I deeply appreciate you taking the time to share if you find it worthwhile.

---