Some thoughts on Internal Platforms

Jason Yip
3 min readFeb 10, 2023

The first question to ask for an internal platform is “how does this enable our product strategy?”

The first question that should be asked for an internal platform is NOT “what are the key activities?” nor “what are the key metrics?” but rather “how does this enable our product strategy?”

Understanding how to enable product strategy requires understanding product life cycles and the different goals at each stage.

Diagram showing the product life cycle curve and different goals at each stage. Market development: “Run lots of experiments”. Growth:”Iterate to grow”. Maturity: “Extract $$$”. Decline: “Reduce operating costs and/or exit”
Different goals at each stage of the product life cycle

For product capabilities intended to discover or create new markets, innovation teams need to run a lot of experiments. An internal experimentation platform can reduce the effort required to run experiments, analyse them, and share results. There is no advantage for each innovation team to re-build their own experimentation infrastructure, negotiate incompatible formats, etc.

For product capabilities intended to grow an existing market, feature teams need to iterate quickly. An internal platform can reduce the effort required to build new features on top of existing core capabilities. Feature teams should be spending most of their time focused on differentiating features and minimize their time having to deal with commodity capabilities.

For product capabilities that are commodities, platform teams should be trying to reduce maintenance effort and otherwise minimize the total cost of ownership. Reducing the operating costs of commodity capabilities should not come at the expense of developing differentiating or innovating capabilities.

Internal platforms need to enable ways of working, not just provide technology.

ThoughtWorks has a digital platform capability model with 5 categories:

  • “Delivery infrastructure”: make it easier to continuously deploy and monitor;
  • “API and architecture remediation”: make it easier to evolve product capabilities independently;
  • “Self-service data”: make it easier to access and use shared data;
  • “Experimental infrastructure and telemetry”: make it easier to setup, monitor, and analyse experiments;
  • “Customer touch point technology”

--

--

Jason Yip

Senior Manager Product Engineering at Grainger. Extreme Programming, Agile, Lean guy. Ex-Spotify, ex-ThoughtWorks, ex-CruiseControl