Risky, but rewarding… a cross-functional team’s discovery project during uncertain times

This piece was co-written by Aoife Spengeman and Kristine Nielsen.

At the Wellcome Trust, we fund many different research projects across different teams and in using different funding models. As part of this, we want to improve data usage for informed and unbiased decision making. We want to share our reflections on a team’s journey to understanding data problems in Wellcome.

1. The project & the team

A small group from across the organisation collaborated on learning about internal data usage to help our decision-making around funding and inform a future data strategy. The group was cross-functional, spanning across the digital team, grants team and strategy team. We applied a variety of methods including an agile approach, human-centered design and data mapping. We had a mix of skills with a user researcher, an agile delivery manager, data specialists, funding business analyst, and this was the first time we worked together.

2. Delivering in phases

An agile approach is useful for projects that are of high complexity and a high degree of novelty. This project was open-ended and exploratory, and we therefore used an agile approach. As a first step, we split our work into phases. This allowed us to focus on the most pressing questions in the first phase, and get feedback from stakeholders, which would then inform our next phase. Getting feedback early and often is a classic agile practice and we did this through presentations/demos at the end of each increment. The phases looked like this:

First phase: Research

• Audit of existing analysis done in organisation
• Initial pilot interviews of relevant people, early-on indicators of where to focus in

Second phase: problem definition

• Scoping
• Analysis
• Desk research/fact finding
• Qualitative data; interviews with relevant people
• Data mapping/process mapping
• Turning analysis into insights

Third phase: recommendations / foundation for strategy

• Defining what a base level system for data analysis should look like and what should be focused on in the future to further advance that to a more complex set-up
• Service blueprinting
• Surfacing opportunities and making actionable recommendations

3. How we worked

When we received the project brief from our stakeholders, we were trusted to refine it. We resisted the urge to to try to to solve problems early on, but still set boundaries on the overall outcome for the team. This allowed team members to self-organise and decide how to solve the problems themselves.

Staying in the problem space before going into solutions

We spent time asking questions such as ‘What is the current landscape?’ and ‘What are the range of possible problems at play?’. We focused on iteratively developing our understanding of the problem space week by week based on new information we would find. Although sometimes nerve-wracking, we didn’t define the problems at hand until halfway through. We recognised that data problems are often related to human problems and we needed to first understand these.

Frequent team communication & feedback

We met weekly and used a stand-up style format. We would each talk through our work from the past week and identify any areas of overlap or blockers. We kept our agenda flexible by allowing for expansive conversations, but would reign them in to clear outcomes and tasks for the next week. This was also a space for team members to request help from others, bounce ideas off each other, and get feedback on their work.

Integrity of approach

When we presented the user needs, success criteria, and tools necessary to set up a base level data system, we emphasised that they needed iteration and future refinement. As we finished our work, we wanted to make sure an agile approach continued beyond the project, and made clear our findings only made sense if they are actioned with the same approach.

4. Reflections

Looking back at the project, we have the following reflections:

Working in an ambiguous problem space is rewarding, but can feel unnatural

In this project, every time we scratched on a problem, it unravelled a larger problem. It was hard to keep our focus on delivery with so much ambiguity. It veered upon being frustrating at times. But, we soon accepted that the problem space was too big for us to ever learn everything about. We therefore focused on unpacking the problem in a holistic way while staying aware of what we didn’t know. The ability to do this was underpinned by the support of senior stakeholders, who helped us develop working assumptions that filled evidence gaps with their insights. Our outputs were framed and treated as hypotheses as a result.

Awareness of the environment is crucial

Data usage has been a long-term issue in Wellcome. We were apprehensive about exploring it during a pandemic and when Wellcome’s strategy was changing so fundamentally. There was an opportunity for positive influence, but we risked not being able to gather an accurate understanding during this time of instability. We kept this tension at the front of our minds and aimed to be sensitive and transparent throughout.

For example, while collecting data we were sure to gather informed consent and present this data as anonymous as possible.

Cross-functional team is useful when looking at an organisation-wide challenge

Working cross-functionally may come with some friction as diverse perspectives come together. Yet, we were able to work to our strengths and bring different perspectives to the table. We think we were able to do this because;

• We were a team of only 5 people, and we communicated efficiently around these broad areas of discussion;
• All team members had flexible and open mindsets;
• Each were aware of own skills/strengths and when to use them;
• We recognised the efficiency of having network knowledge and domain knowledge and we were motivated to maintain this.

This project gained a lot by the wealth of insights brought by the team members. This insight helped formulate useful research questions quickly, interpret findings, find contacts to speak with, and how best to frame our results.

An informal team can be liberating, given the right mindset

In Wellcome Digital product teams, most team members have specific responsibilities, and this is how we are used to working.

For this project, team members had different skills, but we did not set out responsibilities from the beginning. Instead we responded week by week to what we needed to do and learn next. This lack of structure was liberating. Team member’s willingness to be open and pick up tasks outside of their normal remit underpinned this success. Similarly, each team member was motivated to learn about our data issues in a holistic way for long-term improvement.

5. Final thoughts

We shared the outputs with senior stakeholders in the organisation. It also inspired follow-on projects to zoom in on certain aspects of our analysis, as well as contributing to founding new product teams to solve the problems we helped to define. We are immensely proud of this work, particularly because the team learned to work together effectively in such a time of complexity and instability.