Product for Internal Platforms

For the past 3 years, I have been running a platform engineering organization. Since that term is vague, where I work it means the software side of infrastructure. Compute platforms like kubernetes, storage systems, software development tools, and frameworks for services are part of the mandate. Our customers are other engineers at the company.

I also oversee the product team for this area. Now, I’m not a product manager (which I’ll shorten to PM for the rest of this post, not to be confused with project manager), and I rely on my PM team heavily for their expertise. But that doesn’t mean that I can personally neglect the product side, and indeed I spend a lot of time thinking about the products and strategy for my org as part of my day-to-day work.

These are some things I’ve observed and learned in this process. I share them with you because I think many folks in platform-type teams, especially engineers and those of you on platform teams without formal product managers, might benefit from understanding how to approach these problems.

What’s so hard about product for (internal) platform?

Customer Group Size

Captive Audience

A captive audience leads us to believe that basic metrics for customer adoption are not interesting, which in turn leads us to ignore them, sometimes to our peril. Many platform teams end up with several overlapping half-finished products because they assumed a captive audience would lead to product success.

It’s Hard To Think Like Your Customer

When you are on a platform team, it is easy to lose the feel for what it is like to use your own products, because you are deep in the details. You spend your day living and breathing the ins and outs of git, and your users know the 3 commands they have to memorize and otherwise rely on ohshitgit to get themselves out of trouble. In a perfect world, platform product managers are regularly using the products they support in order to identify pain points and gaps that engineers might miss and users may not complain about. In the real world it’s hard to find time to use your products in anger when you are also dealing with all the other parts of the PM job.

You can observe this dilution of focus and quality in a lot of open source platform products that are created by big companies in order to sell you on their cloud solution or to create an industry standard. It starts to look like software that is building to be built. And you absolutely see it in internal platform projects at companies big and small. When platform teams build to be building, especially when they have grand visions of complex end goals with few intermediary states, you end up with products that are confusing, overengineered, and far from beloved.

So how do you solve this?

Assimilate and Expand

I did this when I built a global service discovery solution long ago. Another team had first identified the problem and created their own version of a solution using ZooKeeper. The solution was fine for their needs, but didn’t solve the general needs of everyone at the company for global scaling. So I took over the idea of the project, and turned it into true platform infrastructure, built for a big company and not just one team therein. There were plenty of product decisions to make as part of that work, but the core identification of the problem as worth solving was done for me. There is a lot of interesting work in taking a solution that is locally-optimized and turning it into something that can be used by a diverse set of applications.

Partner to Prototype

Make a Migration Strategy Early

That is all well and good, but the product work is not “tell everyone you have kubernetes now and they have to use it.” Instead, the product work is to identify different types of customers and figure out what will make it easy for them to migrate. What are the carrots you can provide to get people to do work that they don’t care about doing? Perhaps the carrots are efficiencies in getting access to compute or storage. Perhaps you can offer a higher SLO with the new product. Perhaps it is faster, more secure. But these things don’t just happen. You have to choose which features you are highlighting to your customers, you have to help them understand the offering and advantages, and you have to deliver on those promises.

Despite having captive audiences, platform teams are notorious for creating half-finished product offerings that somehow fail to get adopted. When your platform organization is running three different generations of solutions to the same problem with no clear plan to remove any of them, and your customers are both confused by the offerings and dissatisfied with them, you have a serious product failure on your hands. The migration strategy must be a primary part of the product planning.

You Aren’t Google, So Don’t Build When You Don’t Have To

Instead, start with a clear understanding of the problem, and an accounting of your existing ecosystem and culture, before diving into a technical solution. Your data volume is out of control. You might need to solve this with a better storage solution, or you might need to solve it by identifying the top data producers, and asking whether the data they’re storing is actually valuable. You’ll often find that the data is garbage, or the developers can change their workflow, or a little bit of query performance tuning makes this application scale just fine in a normal RDBMS. Only build when you have exhausted the alternatives.

Summing Up

Whether you are a platform engineer, engineering manager, or PM, it pays to remember that you still need to be customer-focused and strategic about your platform offerings. Without a clear strategy for showing impact and value, you end up overlooked and understaffed, and no amount of cool new technology will solve that problem.

Thanks to my darling product and platform friends for their feedback on drafts, especially Renee and Pete.

Enjoy this post? You might like my book, The Manager’s Path, available on Amazon and Safari Online!

Author, “The Manager’s Path.” http://amzn.to/2FvjeHH Distributed systems, dysfunctional programming. camilletalk.com, elidedbranches.com

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store