Do you need a Data Science Product Manager in your team?

I mentioned in a debrief from the latest Data Leaders Summit, the rise of the Product Manager role within Data Science teams.

This was one of a couple of themes that took me by surprise. Over recent years I’ve become used to hearing about need for more Data Engineers or Analysts to complement Data Scientists. But the focus on Product Managers & product development life-cycles was a new one on me.

However, this was not an isolated incident from only a speaker or two. Both in presentation & personal conversations, many leaders confirmed that had Product Manager roles. So, what is going on?

In this post, I will share a combination of my initial thoughts & resources I have discovered as I looked into this. I hope it helps  you decide whether or not Product Managers are needed in your team.

What is a Data Science Product Manager?

To start, let’s define what is meant by this new job title.

Talking with leaders at events, it became clear that some Data Science teams have matured beyond offering advice. Their output was no longer “decision support” analysis, providing models or insight to influence leaders.

Increasingly, these teams were producing products. These could be deployable models (decisioning, optimisation, categorisation) or even entire automated processes. Developing and deploying these into “live” business operation requires some additional skills. It is to meet that need that Product Manager roles have evolved.

Taken from the historical role of Product Managers in operational or marketing teams, these roles own a lifecycle. From initial innovation (e.g. insight generation sessions), through design & development into deployment.

This article, from the IoT for All blog, helps bring the Data Science Product Manager role to life. It is not prescriptive (as frankly the role is still evolving), but highlights some of the key skills needed:

https://www.iotforall.com/new-role-data-science-product-manager/

Was that helpful? The references to facilitation & communication skills reminded me of the need for softer skills. Those matter across so many Data Science or Analytics roles. But, the Product Manager skillset also reminded me of how I used to define Analytics Business Partners. One key difference is the judgement & knowledge needed to manage a production line & pilots.

How do you develop Data Science Products?

Given the popularity of this new role, focussed on a production line, lets explore that further. How do Product Managers & others develop Data Science products?

A number of skills are needed & the most appropriate development methodology will vary by business. But, listening to Product Managers present, there are some common themes.

That thought also struck me as I listened to the presenters at a recent FinTech event. There speakers were drawing on influences from Analytics, Systems Thinking, Agile Working & Design Thinking. But, at least one Product Manager also stressed the role of their product development workflow.

In this article from Harvard Business Review, Emily Glassberg Sands, shares a high-level view of how to do this. How to build great “data products“.

How to Build Great Data Products

Executive Summary Products fueled by data and machine learning can be a powerful way to solve users’ needs and stave off competition. Classic examples include Google search and Amazon product recommendations, but the opportunity extends far beyond the tech giants.

Lots of those tips will be familiar to Data Science & even Analytics leaders. What struck me was the perspective of seeing this work as product development. Plus, the publication of this view in such a mainstream leadership journal. This does appear to be becoming accepted best practice in these circles.

Is this an opportunity for other Product Managers?

Given this emphasis upon product management skills, does this role represent an opportunity for others? I mean for Product Managers currently working outside any Data or Analytics field.

My own experience, with business partner (translator + relationship manager) roles is mixed. I saw more success from those with a background as an analyst or modeller and an aptitude for softer skills. They succeeded in this role better than business partners from other business areas who sought to learn data or insight skills.

However, the Data Science Product Manager may be a different case. reading the above articles, the mastery of product development & management skills appears key.

So, it was interesting to discover this blog post from Cohort Plus. It reads as if aimed at Product Managers in the technology space, but is still a useful introduction for others in such a role.

If you are a Product Manager and interested in making the move into a Data Science team, this introduction should help (apologies but posts from Medium will not display snippets).

Do you need a Data Science Product Manager role?

Hope you found those musings & links useful. Have you decided whether you need one?

I would love to hear your views or experience. It would be great to have comments or feedback from both those who see the value of a Product Manager role in these teams & those who think it’s a fad.

I’m sure more roles will evolve as these teams mature. Customer Insight Leader blog will keep a weather eye on ones that matter. Keeping our readers informed on best practice & resources to help you develop.