We just recently won a major industry award for our Brilliant at the Basics programme: Computing’s DevOps Awards 2021 included a special award for Most Successful Cultural Transformation. Creating the submission allowed me to reflect on what has been a two year programme. We learned an incredible amount, we bonded as a team but, most importantly, I remember this as being fantastically fun. It’s definitely one of the best things I’ve done in my career.
I’m very proud of what we achieved with this programme. And the award is great recognition. Below is the submission which explains the programme.
The goal: To get 2,000 people to brilliance in their agile and DevOps practices, to do so consistently, to show true servant-leadership in the process and have fun along the way.
The answer: a two-year programme called Brilliant at the Basics.
The programme comprises:
- The most senior leaders in the organisation rolling up their sleeves and teaching the teams.
- A rating scorecard against 15 practices.
- Peer review and best practice sharing across teams.
- A drumbeat of one practice every two weeks with everyone participating.
- Crowd-sourced material from inside and outside the organisation contributed by the teams
- Competitions and hackathons to drive innovation and deep focus on the difficult areas like test driven development.
- Regular reviews at team and department level of progress.
- Open sharing of teams’ ratings.
- Connection to outcome measures: we used work-in-progress, mean-time-to-recovery and defect counts in particular.
- Fun – each practice had an informal competition to present in a new, different way.
At the start we had no teams that were brilliant and many had no awareness of their deficiencies. Currently 80% of our teams are brilliant at 12 of the 15 practices and newly formed teams are gaining brilliance in less than three months.
The culture of an organisation starts at the top; innovation is driven bottom-up. The best leaders are servant-leaders.
The leaders of IBM’s CIO Sales and Marketing Systems department took these premises to heart. If they were to accelerate their teams’ progression towards consistently true agile and DevOps practices, they would have to lead from the front. The challenge was to do that at scale: the department is 2,000 individuals and 285 teams across the world.
The impetus for the programme came from a leadership meeting in Bangalore in Feb 2019. We recognised that our teams had stagnated in their agile practices.
Our goal as an organisation is to become a world-class software engineering organisation. To do that we needed to have a way to measure ourselves.
We identified 15 agile practices and using research from across industry came up with a set of scorings for each that would allow teams to self-assess. That gave us a baseline.
Then we needed to get better.
Every two weeks during 2019 and early 2020, we focused on a practice and two of the top leaders in the organisation led the education sessions. This drove partnership between leaders who sometimes didn’t work together often and set an example that they were willing to learn before they could teach. Each fortnight we would have events, materials to read, games and a final self-assessment in each team. That would culminate in an action plan for improvement.
We used Slack channels extensively to answer questions and to share best practices among the teams.
Metrics reviews, at a team level, of where we sit in our agile and DevOps practices are a reviewed monthly. Team leaders are encouraged to share their improvement plans.
Self-assessment and peer review were very important components. Teams could self-assess but that was supplemented by peer review with iteration managers and first line leaders from different teams. Peer assessment drove consistency across all 285 teams as well as the opportunity to share best practices – the reviewers brought what they saw back to their own departments.
A key element to the learning sessions was fun. A sort of informal competition started to take place where the teams presented information in new, fun ways, whether that was from role-playing videos to online games.
We also found that the teams were very active in contributing information. Each focus fortnight had pre-prepared materials gathered by the leaders of the practice. But teams quickly supplemented that with their own findings. We were keen to learn from outside so many of the materials came from external sources. Thoughtworks’ blog was a particularly rich seam of information.
Information sharing has also worked well. Teams can see how their peers are doing. That allows a team that might be struggling with one practice to find a team that excels. We’ve encouraged shameless copying of what works from great teams. A number of shared solutions and frameworks have emerged from CI/CD pipelines to application penetration testing to security secrets storage.
We got 80% of our teams to brilliance over the course of the programme. But we’re operating in an environment where attrition is high. An equally important measure is how quickly we can get new teams formed and brilliant – that is now three months.
We inherited an organisation that did quarterly deployments; now our time from taking an item off the backlog to delivery (work-in-progress) is, on average, 4.5 days across all technologies from mainframe to cloud containers.
Our mean-time-to-recovery for Severity 1 incidents has plunged from greater than 20 hours to 5 hours.
The program is well liked by our teams. We run an NPS survey each year on whether employees would recommend their team as a great place to work to a colleague or a friend. At the start, our NPS was 37; halfway through the programme, it was 53. At the end of the programme it stood at 57 and many of the write-in comments mentioned the Brilliant at the Basics programme as a differentiator.
We completed the first phase of the programme in mid-2020. Since then, we have had more intensive education around some of the harder practices. We ran a six-week in-depth set of workshops for the entire department on user-centricity. We taught developers the skills to interview users and incorporate their feedback. We instrumented nearly all of our applications with Usabilla to get continuous user feedback
We had special focus on Test Driven Development with a competition across the entire organisation and cash prizes for the winners.
We have taken the same pedagogical principles into new areas: our second half 2020 Hackathon for example was on implementing artificial intelligence into our solutions. As a result, we delivered 20 significant new enhancements to our systems using AI.
Once you’ve worked on a high-performing agile team using the best engineering practices it’s hard to go back. That memory of how a practice is done well stays with you forever. We’ve seen this over and over again in the formation of new teams and how we interact with teams outside our organisation. Experienced team members bring that knowledge of how things are done well to their new teams – inside and outside IBM.