This client, a leading UK credit reference agency, approached us in 2019 to help with a specific challenge as part of an existing coaching contract.
The Client Need
The client had a specific function developing data tools for their customers. These tools are complex, bringing together massive financial transaction data sets to give behavioural insight related to credit profiles and scores. Each product was developed as a unique project calling on whichever delivery teams had capacity at the time. The client felt that product lead times were too slow, which was restricting growth, and they asked for advice on improving this.
Identifying the Challenge
Based on previous experience coaching the development teams, we knew that some teams were short lived, with individuals swapping teams to meet new project needs. This is known as “taking the team to the work” and is typically inefficient because 1) the team members need time to learn how to work together and 2) the team generally needs time to learn the technologies associated with each project.
We also discovered that the quality of the technology base was deteriorating as project teams applied workarounds and quick-fixes to meet their time goals: this in turn made future development take longer, increasing the problem in a vicious cycle.
The Solution
Working with the data tooling function, we identified that the bespoke client offerings were really combinations of the same discrete services, packaged together in different ways with different data sets as bespoke offerings.
This opened up the possibility of switching from project-based to product-based development. In product-based development, fixed teams are associated with specific products, removing the need to keep re-learning how to work together while also building and maintaining technical knowledge. Any work related to that product then comes to the same team, which is known as “taking the work to the team”. The team knows they will continue to work on the same technology so they have a vested interest in maintaining its quality. These three effects combine over time to accelerate development and reduce lead times.
The client agreed to a product-development experiment: we trained two subject matter experts in product ownership and gave them two commonly used services to look after, with fixed teams to support them. When any work was requested on those services, it would always go through those product owners and teams. Success would be measured using two simple metrics: lead time and team happiness.
The Outcome
The outcome of this experiment was stark. Within weeks, anecdotal evidence from the team was that they were less stressed and able to work much more efficiently without context-switching. Other teams and project managers were so impressed they adopted a similar approach.
At a two-month review, tech leads reported code was becoming cleaner and lead times in the experiment teams were demonstrably reducing. This allowed the teams involved to plan in more work, sowing the seeds for sustainable growth.