Engineering excellence at scale
Raising the engineering bar across a whole company
Shared engineering standards, delivery-flow measurement and an evidence-based maturity report, reaching more than 4,500 developers.
4,500+
developers reached by shared engineering platforms and standards
By 2023, across business units and global hubs.
Context
Philips builds software across many business units and sites, each with its own tools, pipelines and habits. That makes delivery hard to compare and quality problems hard to see early.
Our 2023 data showed the gap. Nearly every team had continuous integration, but fewer than a third could deploy automatically to a test environment. About two thirds had no defined process for technical debt, and most engineers still tested by hand.
The decision
Treat engineering quality as something to measure, not argue about. That meant measuring delivery flow with the DORA metrics and continuous value delivery, and publishing maturity data across the software community before recommending where to invest.
The usual answer was a quality gate at the end of the pipeline, or a report sent up to leadership. I pushed for quality at the developer’s desk instead: automated checks inside everyday builds, backed by data we shared openly with teams. We worked alongside teams to deliver early results rather than mandating from the top.
What I did, and what the team did
I led a continuous value delivery program and oversaw its execution. I built and taught the company approach to measuring delivery flow with DORA and continuous value delivery, and co-authored the company CI/CD reference architecture whitepaper in 2020.
I was lead author of the 2023 State of Software Craftsmanship report, based on data from 138 projects. It showed that teams with weaker engineering practices were more likely to have product faults, and found that about a third of teams were not enforcing quality gates. It set out where to invest next.
I also proposed and led a quality-at-desk program: automated checks on complexity and duplication in developer builds, adopted by several product teams. For a global quality program, I set the KPIs and oversaw implementation.
The craftsmanship assessment program I was accountable for raised the average assessment score from 67 to 78. On one product in the program, customer defects fell from 37 to 0.11 per 100 exams, and its app rating rose from 2.3 to 4.5. Separately, automation work I directed with US teams saved about $300K in infrastructure costs.
Outcome
~60%
shorter release time in the continuous value delivery program I led
Result for the participating teams.
138
projects in the 2023 State of Software Craftsmanship report I lead-authored
Part of a program reaching over 80% of the company software engineers. Published 2024.
37 to 0.11
customer defects per 100 exams in a global quality program
The app rating rose from 2.3 to 4.5 over the same program. I set its KPIs and oversaw implementation.
~$300K
in infrastructure savings
From automation work with US engineering teams.
What I would do again
Measure first, and put the results where engineers work. Quality moved when teams saw their own data in everyday builds, not when it arrived in a report for leadership.