Rob Kleiman

• Written by Rob Kleiman

Giving Amazon a Voice in the Developer Productivity Conversation

Amazon has long invested in tools that help teams build, deploy and operate safely at scale. But measuring the business value of a better developer experience had stayed an unsolved problem. Meanwhile, frameworks like DORA and SPACE, both backed by Google, shaped how the industry talked about developer productivity, and Amazon was visibly absent from that conversation.

The challenge

I led the content strategy and execution for Amazon’s new cost to serve software framework (CTS-SW), which measures what it costs to deliver a unit of software, like a deployment, to customers: positioning it as a credible, data-backed alternative and giving Amazon a voice in a conversation it had been missing from.

What I did

  • Went to the source. I met directly with the VP leading the methodology to understand the framework deeply, then shaped its positioning and narrative.
  • Found the headline. In 2024, Amazon’s developer experience improvements cut the cost to serve software by 15.9% and raised weekly production deployments per builder by 18.3%, year over year. Automation cut human interventions per deployment by 30.4%. To make it concrete for other companies, the story used an example: a bank with 1,000 developers getting a similar 15% improvement would avoid about $20 million in cost.
  • Built a publishing machine, not a one-off. A content series and runbook tied to operational health meetings gave the team a repeatable process it could own.
  • Kept it moving. I ran weekly standups and sprint planning across stakeholders to hold publication timelines.
  • Published where the audience was. Executive pieces ran on the AWS and Amazon Science blogs, LinkedIn and at re:Invent:

Jim Haughwout on measuring developer experience with cost to serve (AWS, November 2025):

Measuring Developer Experience Impact with Cost to Serve

What changed

The framework drew inbound interest in building on AWS, fed into whitepaper development and was reshared by Amazon’s VP of Agentic AI. It established Amazon’s voice in the developer productivity conversation and created new funnel and leads for AWS solutions.

What I took away

Thought leadership needs a number people can repeat and a process that keeps publishing after the launch buzz fades.