Case Study: Rescuing a 20-Person Dev Team in Financial Services

Summary

Client: Specialist insurance division of a financial services firm.

Team: 20-person development team, grown from 6 in 18 months.

Services: Delivery leadership, agile coaching, Training, Jira overhaul, team restructure, problem solving.

Engagement length: 100 days.

Key results: 50% increase in throughput · project cycle time cut from 9 months to 3 · £1m in licence-renewal costs avoided.

The Challenge

A team of 6 had grown rapidly in under two years, and now comprised 20 skilled developers, responsible for supporting a fast-growing insurance product. Unfortunately, despite the growth of the team, they lacked a delivery lead, scrum master, or project manager, creating a structural gap that had ultimately caught up to them.

As a result, the team found themselves juggling building new features with keeping the platform running. Support was taking up the bulk of their time, and with an admin system replacement project quietly approaching its deadline, they needed to act or face a £1m licence renewal bill.

None of this was due to a lack of ability; with 17 projects running simultaneously, it left nobody to prioritise or push back on new work. From the outside, it looked like a team that had slowed down, but on closer inspection, it became apparent that they were simply spread too thin, without the person needed to see the whole picture.

That's where we came in.

What We Did

We worked closely with the client, providing an experienced pair of hands willing to get stuck in. Our goal was to create solutions to each problem, so the client wouldn't need to continue finding workarounds.

That meant:

• Reorganising a team of 20 into 4 smaller, focused delivery teams. Right-sized, cross-functional, and each able to deliver end to end.

• Building a dedicated support function, separate from the new-build team, with its own portal and process.

• Centralising all work in Jira, including portfolio-level tracking via epics, so for the first time anyone could see who was working on what.

• Replacing 10-week UAT cycles with continuous testing, so defects surfaced during build, not months later.

• Training the team through two targeted courses rather than one generic programme.

Discovery

We started by carrying out a structured audit of the team's Jira data, along with looking at how they worked. We used our findings to confirm what their leadership already suspected and get buy-in from stakeholders before changing anything.

Training

At Add Agility, we have a suite of interactive 1–3-hour courses that can be tailored to each client. This team, and some of their stakeholders, received an introduction to agile, Jira 101, and writing better requirements in the agile of AI.

Rolling Out Agile

We took the existing backlog and restructure it into a simple, low-admin hierarchy. Four focused teams came about, and a stakeholder analysis grew alongside every new epic. This ensured that the right people were in the room at the right time, resulting in fewer delays and cancellations.

Coaching and Mentoring

Delivery leadership was never meant to be permanent, and that shaped how we worked from day one. We conducted one-to-ones, small group coaching, and made ourselves available for making decisions the leads weren't yet confident to make. Our goal was to shift the team towards a state where they could eventually operate independent of us.

Solving Problems

Agile process makes problems visible; it doesn't solve them. To solve these problems, we restructured underperforming workstreams and negotiated additional underwriting time with stakeholders to fix an unworkable UAT process. We also took on decisions the existing management lacked the bandwidth to handle on top of day-to-day delivery.

The Results

By the time we left, the following had been achieved:

• Throughput Increased 50%: Without hiring a single employee, we'd achieved a result roughly equivalent to adding 10 skilled people to a 20-person team. To get there, we freed up capacity in support, addressed underperformance, and implemented testing throughout the build rather than at the end.

• Fall in Work in Progress: Our efforts led to a fall in work in progress from 352 items to just 70. This represents a huge improvement, saving the team from constant context-switching while providing a clear view of the full picture.

• Reduced Project Cycle Time: With a shorter project cycle time, down from 9 months to just 3, value landed sooner. We made changes to process, governance, and how the team engaged with stakeholders, resulting in less risk sat waiting inside long-running projects.

• 21 Epics Delivered: Just six months prior to us coming on board, 17 epics had been running simultaneously, which for a team of 20 was far too many. We delivered all 17 and were even able to take on 4 more, without the backlog creeping back up.

• Avoided Licence Renewal Costs: The client was facing a £1m licence renewal fee that was quietly getting closer. We helped them avoid this entirely, simply by getting a critical admin system replacement project back on track and over the line on time.

• Drop in Support Headcount: The support headcount dropped from 10 to 5, without impacting performance. In fact, they managed to clear a significant backlog of old tickets, while improving response times and cutting complaints. All thanks to a new process and a dedicated, focused team rather than one spread thin doing two jobs at once.

What's Next

This is what happens when a growing team gets an experienced pair of hands, not another framework. If your team also finds itself in a similar position, with good people, invisible bottlenecks, and nobody with the bandwidth to fix what everyone already knows is wrong, then get in touch with Add Agility today.