Data-Driven Product Teams Are Easier to Empower and Earn Autonomy Faster

Problem

At one of the companies I worked at, I inherited some great PMs, leading some awesome cross-functional teams (about 35 specialists across engineering, design, analytics, user research, content and marketing). However, these PMs and their teams were very unhappy. They felt they had no autonomy and were not responding to customer needs but rather executing what felt to them, like randomly chosen roadmap features. I met with my leadership and others across the firm and it seemed like a lack of shared goals and success metrics was the root cause of this problem. Everyone at the firm wanted to be focused, ‘data driven’ and have self-managing, autonomous teams. They were just struggling to make it happen.

In fact, leadership worried it wouldn’t be safe to empower teams, to respond directly to customer needs in this way. They felt that without agreed goals and shared metrics first, the teams trying to earn their autonomy this way, would just fail. And so, because they weren’t sure how to pick the goals and metrics yet, they’d only set feature goals for the teams. So I researched our business challenges, worked through the customer research and found the right metric targets and offered to do a trial with my teams. I proposed these goals and metrics to leadership, who reviewed and supported them. And so these teams could trial making data-driven product decisions for customers and start to ‘earn’ their autonomy.

Actions taken

First and foremost, data-driven decision making needs solid data infrastructure. One of the key challenges I faced in getting firm-wide support for this trial was that we uncovered fractured tech platforms and clashing metrics formulas. So we took a step back and started to consolidate the platforms, and migrate the key data sets together where possible, and produced shared data dashboards for the new metrics. A lot of the time doing this was spent unifying the metric formulas. This work sped up, as the company hired more analytics leaders and brought in 3rd party specialists to help.

Secondly, I proposed to focus my teams (who managed the platform product) on retention metrics, as finance had shown us, moving this metric would be most impactful to the recurring revenue we needed to scale. And as our subscriptions were annual and charged monthly, I chose monthly active users, as our shared ‘north star’ metric, because analysis showed it historically to be a leading indicator for our annual retention. Plus, customer confusion over the product's monthly purpose had come up in qualitative feedback, and we wanted to work on our ‘monthly’ value. So I pitched this to my teams and leadership and gained support for the new shared goal and metric. Also, as part of wider analytics training, I presented this at the company All-Hands to gain support and encourage radical suggestions on what might most improve our shared metric.

Prior to this trial, the metrics dashboards were siloed to a specialist team for use and poorly understood. As we progressed in our trial, we published our shared dashboards and encouraged feedback and discussion. This built trust in the teams and helped them start to ‘earn’ autonomy for their product choices, for customers and the business. One of my favourite moments was when a team rolling out a key A/B test, hit errors in a troublesome user journey. Prior to the shared goals and metrics trial, it used to take months to debug such problems. However, thanks to the trial, now everyone could see and understand the challenge and the relevant specialists actually stepped forward without being asked to! It was solved in hours this time! And I’m delighted to say, with all these teams pulling together we improved the metric enough, that going forward, as these same PMs proposed new roadmap ideas, they were most often approved.

Lessons learned

• North Star metrics are great for helping to encourage an empowered product team culture. But if you don’t have the right environment yet, give it time - it can require a lot of investment first, to get this setup. Consider that you may need significant time for data engineering, data infrastructure, setting and agreeing shared goals and metrics and data analytics training, before you can even focus your team(s) on targeting the chosen metrics.
• At the same time, metrics are a great way to ‘earn’ autonomy for new teams and move to solving the most impactful customer problems, when you feel you might be more of a random ‘feature factory’. In my own case, as we started sharing ideas on what would move metrics, things changed radically. PMs and teams started making suggestions that were accepted into the roadmaps, with full leadership support. Best of all, not long after we started, the teams were able to move our North Star metric by 200%!