Empowering Teams To Create Impactful Products

Hello, I’m Audrey Tsang. I’m the Chief Product Officer at Clue. We’re a Berlin based menstrual tracking app helping over 12 million monthly active users globally live healthier lives. Given that I work in female health—fun fact—at the time of this recording, I’m in my follicular phase. So, that’s between that my period and ovulation, and I often experienced this time as the more energetic, more social, and hopefully for you and this talk, more charismatic phase of my cycle. So now that I’ve overshared that, some background on me, I’ve been building consumer tech products for over 15 years now, mostly in Silicon Valley. Before coming to Clue two years ago, I led product teams at Pinterest, HotelTonight, Yelp, and at my own startup. So, I’ve seen startups through many different stages and through many different challenges. What’s been true throughout my career though, is that I’ve always been drawn to diverse teams and products that improve lives. And that’s also what drew me to Clue.

Today, I’m going to tell you a story about how a values-driven team began delivering performance-oriented results. In other words, this is a story about doing good for our users while building a sustainable business so we can keep doing good. I’ll go over lessons we learned while relaunching our subscription product, but first 60 seconds of context about Clue and why we do what we do. First, what user problem are we solving? There remains today, still, an acute lack of information and access to essential care in female health globally. And there is still so much we don’t know. And missing scientific research and under-funding, together with cultural stigma, continuously provide disappointing outcomes and create suffering and lost economic opportunity.

We want to change this. So, no matter where they are from or how they identify, our vision is for women and people with cycles to live full lives, in tune with their biology, not in spite of it. Our mission is to empower users with the science, data, and technology needed to make informed choices about their lives. We value trust, care, and science, and we’ve built these into a period tracking app that is science-based, data-driven, and emotionally-intelligent. Clue was founded 10 years ago and has since grown to a 70-person team. And as CPO, I manage product, design, data science, and monetization.

I’ll go over three lessons today about what we’ve learned from re-launching our subscription product, and about how to empower teams deliver results while living our values. But first, I want to give you some background on the mistakes we’ve made, and why we chose to change tack over the last few months.

Clue Plus, our subscription product, launched in its first rendition almost two years ago. It was priced at 99 cents per month to make it affordable to as many people as possible. It included premium features like cycle alerts, which for example, would alert you of irregularities in your cycle. And like many newly launched products, Clue Plus V1 saw some initial success because of buzz and loyal users subscribing. And as can sometimes happen, with early success, product and engineering teams moved on to build other—maybe more exciting things, and marketing efforts were shifted to focus on user growth. So a few months later, revenue growth and conversion, of course, slowed.

This continued until August 2019, when as a business, revenue emerged as a much more important proof point than user growth, and we refocused our efforts on Clue Plus. We formed a revenue team consisting of CRM, Content, BI, Product folks, and they pulled in other resources from science brand, partnerships, and growth. This team set out to tackle three main problems. One, low awareness, as most users didn’t even know there was a premium offering. Two, poor economics, 99 cents a month was much too low of a price to make this a sustaining business. And three, an unsatisfying feature set. We knew from cancellation surveys and support feedback that many subscribers didn’t understand or benefit from the premium functionality offered in V1. What differentiated Clue and make users love it was not at all linked to the subscription offering.

So in September 2019, we re-launched Clue Plus as a contribution model with the hypothesis that an altruistic model would better highlight Clue’s differentiating features of science, trust, and data privacy, while at the same time making all of Clue freely accessible to the most people. We removed most in-app premium features from behind the paywall, we changed our messaging, and we increased our prices to prove this hypothesis. And the team did this all without dedicated engineering resources. And despite that, the teams still managed to meaningfully increase revenue and learn, running over 50 experiments in three months. You can see our most successful campaign on the left here, “We’re not like other period trackers”, which we aren’t. And some of the new premium features on the right, like period product discounts and special contents, like puppies for PMS that we positioned as a “Thank you for contributing.”

While we were learning and making progress, revenue growth again slowed in February 2020, after the typical holiday surge in new devices, and the “New Year, New You” seasonal pop that health app C. By this time, the monetization team was burnt out. CRM campaigns stumbled on deliverability issues. We were unable to hit aggressive targets, and we were blocked on launching much-needed subscription tooling quickly enough.

Then in March of this year, we were under even greater pressure to grow revenue and conserve cash. As you all experienced globally, businesses are coping with the pandemic and its economic consequences. So, we assembled a new revenue task force, but this time we approached things differently. Spoiler, we’ve now had five record months of revenue, and the team is confident that we’ll continue to hit new aggressive targets.

So while it’s still early in our journey, I’ll share what we’ve done differently this time and what we’ve learned about how to empower teams to achieve these results while still staying true to our values. So, let’s look at these lessons learned. The first lesson is about focus. We dedicated a cross-functional team to revenue and nothing else this time. Here’s our lovely team—actually just a part of it—but you can see that in contrast to V2, we have added engineering and design to the revenue task force. And this team coordinates the efforts of three of our four cross-functional product development teams. All also dedicated its revenue initiatives.

This full stack focus setup is empowering because it allows the team to: one, broaden the solution space by drawing from more diverse perspectives; two, quickly scope potential solutions and make informed prioritization decisions based on data around impact, effort, and risk; three, agilely go from hypothesis to design to test to iteration; and four, have more fun. So, setup like this, the team has launched something new every sprint and we’ve hit revenue targets every month for the last five months. So maybe the need for a cross-functional team can seem obvious, but so often it does not happen. Setup like this, the team has launched something new every sprint and we’ve hit revenue targets every month for the last five months. So maybe the need for a cross-functional team can seem obvious, but so often it doesn’t happen. It takes ruthless prioritization and consciously not doing some things, all of which will seem very, very important. So in our case, while it wasn’t easy, it did pay off.

The second lesson is about giving the team autonomy with top-down vision, not top-down solutions. So, most of us have at some point been given a management directive from a founder, the board, a manager, spanning everything from hit XYZ metric to build XYZ feature. As I’ve learned, starting with a solution like this instead of the problem, almost always results in worse solutions and less motivated teams. Instead, providing a top-down vision means that management provides the “why”, the constraints and the principles that allow the team, the autonomy to set the “what”, the targets and the roadmap.

So let’s talk about goal setting first, empowering the team to set good metric targets starts with communicating the “why”. In this case, the “why” was about to helping Clue make enough money so that we can sustainably continue to make a difference in female health. For the first time, I shared our monthly burden with the team and the impact revenue would have an offsetting that burden. I provided a high level 2020 revenue number as guidance, and more importantly, tied it to our runway. While this amount of transparency can feel scary, the team responded to it with resilience. They set their own aggressive monthly targets shown here, in absolute and percent growth. And this served as an important motivator for the team, because they could see real progress towards challenging goals and then celebrate when they hit the goal. And they felt confident that what they were doing made a difference to our business and to our users.

That was goal setting. Now let’s talk about roadmap setting. Empowering the team to set a roadmap is about communicating clear constraints and principles. In this case, the constraint was time. I shared a time horizon tied to business needs. The principles are around risk and user values, primarily how aggressive should we be with asking our users to subscribe. In addition to estimating impact and effort, the team then could also consider values, alignment, and risks to user satisfaction as part of the prioritization and roadmap setting exercise. You can see how this plays out with the prioritization and scoping of our premium content feature. At first—I’ll admit sweat up and a desire to move fast—I told the team, “Just build a content tab.” This would turn out to overestimate impact, underestimate effort, and lose sight of the user problem. By framing this instead in terms of constraints and our principles, the team came up with a more nuanced solution, “Science says” that plays to Clue’s strengths, its scientific expertise and values, and being emotionally intelligent and caring.

The third lesson is about boundaries. In V3 of Clue Plus, we are much more aggressive about asking users to subscribe. We have many more entry points to subscribe and we have paywalled some of our previously free features, in addition to adding new ones. While being more aggressive, we’ve still managed upholds our user-centric values. One way we’re able to walk that line is by defining boundaries for being aggressive and then walking right up to those boundaries. So, what do I mean, how do we define those boundaries?

First, we challenged existing assumptions about what we think is too aggressive or too annoying to the user. And second, we defined a measurable boundary for negative user feedback, which in our case was a greater than 0.1 decrease in our average app store rating. We can see on the left that we’ve added entry points to subscribe on all of our main screens, which is pretty aggressive, especially compared to before. The team debated how much this might anger or annoy our users. But instead of dwelling on discussion, the team quickly launched the experiment with the boundaries that we had defined. What made this possible was a commitment to carefully monitor, not just conversion, but also app store reviews and support requests, and base decisions on data rather than debate. In this case, we ended up adding an entry point to our cycle view, and then adding and then removing the entry point from the calendar view. All while increasing conversion and staying under acceptable decreases in our average app store rating.

A second way to be more aggressive is to be selectively aggressive. When considering whether to put an existing feature behind the paywall for V3, the team completed what we called a Subscriber Feature Segmentation. This analysis revealed that certain features were disproportionately more used by subscribers than non-subscribers. And these features were a great place to start experimenting with taking away what were previously free features. This analysis led us to move our symptom details feature behind the paywall, because it’s 61% more likely to be used by subscribers than non-subscribers. And now that we’ve shipped it, this feature contributes to almost 10% of our subscription conversion.

And finally, staying values-driven while being aggressive, requires the entire team to be clear on those boundaries and aware of the process. And while making bold moves like this, it’s important to welcome debate and communicate early and often into the broader team to keep everyone bought in. Here’s an example of how this plays out during a series of experiments on our new user subscription information screen. In iteration one of the screen, users could dismiss the prompt by tapping outside the in-app message or on the “X” in the top right corner. In iteration two, the team wants to experiment quickly with driving more traffic to the purchase page, which came after this. And so, we’re using code and design from iteration one. We did this quickly by just removing the “X”, making it less obvious how to dismiss the screen. Our UX researcher and one of our Android engineers pointed out this dark design pattern, and they were upset that we would mislead users in such a way. So, what started as a time-saving experiment by the revenue sprint team turned into divisive miscommunication with the broader team. This was only resolved once the revenue team had a chance to communicate to the broader team, the backgrounds, the intention, and the principles for decision making and iteration moving forward. This communication was incredibly important. With a shared understanding and greater now trust for each other the team could define new iterations to test moving forward. And ultimately, what they found was that it didn’t matter whether the “X” was there or not. Simply renaming the button “Next” delivered the largest conversion gains.

These experiments were some of the first efforts towards improving our new user conversion rate, which is important. With one of the best opportunities for converting a user—being when the user is new—having this shared understanding, this ability to collaborate, and quickly being able to experiment with efforts like this, have all greatly contributed to growing subscriptions. And as you can see, we can see the benefits in our new user conversion rate, which has increased on both platforms over 6X since beginning.

To sum it all up, we’ve learned a lot by doing things differently this time around. First, we should empower the team with focus and diversity to make better, quicker, cross-functional, data-driven decisions. Second, we should empower the team with the autonomy to define their own targets, roadmap, and solutions so that they have more motivating targets and stronger solutions. And third, we should empower the team with clear boundaries and the values behind them so that they can experiment quickly and with conviction amidst, always difficult trade-offs. These are challenging times to say the least, but it has forced us to approach things differently. And as a result, learn and grow tremendously.