How Nationwide Scaled DevOps Across 200+ Teams: Lessons from the DevOps Handbook

Background

Nationwide is a U.S. insurance company that ranked 68th on the 2017 Fortune 500 list with $40,074.1 million in revenue.

DevOps Case Study Overview

Nationwide is featured in the *DevOps Handbook* as a high‑maturity example, frequently speaking at DevOpsDays and invited by tool vendors to share its experience.

Presentation Source

The analysis is based on a November 2017 San Francisco DevOps Enterprise Summit talk titled “DevOps Handbook Experiments in Accelerating Delivery – Nationwide”, transcribed from a YouTube video also hosted on Tencent Video.

Organizational Scale

IT organization >5,100 staff, >2,600 developers and testers.

200+ product development teams serving 23 business units.

Complex Dev, QA, Ops structure; cross‑BU pipelines often require CIO‑level coordination.

Strategic Context

IT pursued four strategic goals: Agile (adopted 10 years ago), DevOps (speed, efficiency, risk reduction), Lean IT (enterprise‑scale adoption), and CMMI (process maturity).

DevOps Journey Timeline

2015 – “DevOps Cabin”: early framework, IBM UrbanCode as a unified deployment tool.

2016 – Organizational focus on development model, lean IT, and supply‑chain management.

2017 – “Dual‑mode experiments”: minimal marketable product (MMP) focus, programmer‑experience experiments drawn from the Handbook.

2018 – Scaling tools to 200+ teams, adopting Google SRE model, treating pipelines as products.

Team Interaction Model

Three guiding principles: speed, practitioner‑driven improvement, and end‑to‑end value‑stream scope. Experiments are run by two “A/B” teams, overseen by a DevOps Leadership Team (Cindy and Jim). A platform team builds enterprise‑grade APIs; a governance team of CIOs reviews each iteration.

Epic Work Items

Work items (EPICs) are selected directly from the *DevOps Handbook*; teams verify each item against the book’s chapters and page numbers.

Developer Experience Improvements

Automation of test environments reduced test‑bed size and cycle time; pairing developers with testers eliminated duplicate tests; performance‑test turnaround dropped from 90 days to 2 hours. Introduction of Rocket.Chat and a deployment chatbot let developers trigger deploy via chat, replacing 10‑20 manual clicks in UrbanCode.

Guidance “Climbing the Mountain” Metaphor

A one‑page “DevOps climbing guide” maps three phases (base camp, north‑face camp, summit) to practice maturity, using “lead time” as the sole metric. The guide is used internally to onboard new teams.

Lead‑Time Metric

Lead time can be measured across 200+ teams, is understandable at organization, department, and team levels, and can be tracked daily while maintaining zero‑downtime deployments.

Support Model

“Sherpa” teams coach and mentor lower‑maturity teams; monthly DevOps Dojos provide hands‑on guidance; pipelines are treated as products requiring dedicated support.

First Success Story

During Hurricane Harvey, a surge in flood‑damage claims prompted a business request to streamline the vehicle‑inundation insurance claim process. The IT team eliminated 40 % of steps and delivered the change within the same business day, reducing a typical 90‑day cycle to hours.

Conclusion

The case demonstrates that a large, regulated financial services organization can achieve rapid, low‑risk delivery by iteratively experimenting, measuring lead time, and scaling support structures.