Insights/Action/Outcome: Unmind had a chatbot that was designed to answer questions within their app. Amplitude Analytics helped them discover that when the bot was involved in Unmind courses, users were more likely to drop off. This data contributed to the team’s decision to remove the bot from the app, and they saw an uplift in course completions.
Data is power. But that power is not always enough to ingrain data into everyday processes. Jumping that hurdle was where our product analytics journey began to become more data-informed. The process has improved how we build products, engage customers, and create user experiences.
The quest to turn Unmind into a data-led company
is a holistic workplace wellbeing platform on a mission to create mentally healthy workplaces, where everyone can flourish. Today, we have approximately 180 employees or “Unminders.”
Our journey to data maturity began a few years ago. We always valued data, and we recognized the importance of making cross-functional decisions that were data-driven.
Yet, we were not handling data-informed decision-making in a scalable way. The data team was the classic bottleneck, we were approached frequently with questions and requests, which we’d answer by querying the database or running Python scripts. But the requests outpaced the bandwidth of the data team, and even though we prioritised those requests, there was always a trade-off.
This bottleneck meant we sacrificed velocity, since it impeded other teams. In addition, addressing data questions prevented us from working on other self-service solutions that would allow stakeholders to get the required answers. We wanted to break the cycle.
The use-case-driven search for a scalable solution
Querying a database to answer targeted questions and sharing in spreadsheets can work well for small companies beginning their data journey. However, as Unmind grew along with our product, features, and client base, it was no longer sustainable to manually perform these tasks. What began as a manageable bottleneck became a pain point that grew along with us.
We set out to find a solution, and created a strong evaluation process to prevent us from getting distracted by unnecessary features in various platforms. We’d recommend using a strictly use-case-driven process:
- Start by looking at your desired use cases
- Decide on must-have features, such as A/B testing
- Judge the platforms and features against your set criteria
We reviewed the leading providers, including Heap, Mixpanel, and Amplitude, and evaluated these platforms through demonstrations. After rating them, Amplitude provided the best fit for our use cases at the time. They offered a rich and robust product set and had a reputation as industry leaders. We were confident our investment would pay off.
Amplitude provided the best fit for our use cases at the time. They offered a rich and robust product set and had a reputation as industry leaders.
Enablement efforts begin with the data team
Analytics has a wide breadth of functionalities, and we wanted to ensure colleagues across the business understood how to properly utilise the platform. We created various internal resources, including an onboarding guide, where we included links to our video tutorials and suggestions on where to start. This proved valuable; Amplitude has a wealth of great online documentation, but it is not always the first place new users look.
Not even the best video tutorials and documentation can replace being hands-on with stakeholders, taking the time to sit down with them and demonstrate how to find what they need instead of doing it for them or pointing them towards another resource. The data team made a concerted effort to support people during onboarding, hosting calls and creating a dedicated Slack channel for Analytics questions. We held weekly drop-in sessions where people could obtain Analytics help without having to book a meeting.
Powerful tools can be intimidating, and some people who had never used a self-serve platform were worried they might break the system. Our extra effort ensured everyone—especially our product managers—knew how to use Analytics with ease.
There is not a single approach to guarantee a successful implementation; many different small actions helped change people’s habits.
Tips for a successful launch
While there is not a single approach to guarantee a successful Analytics implementation, I can recommend many different small actions that helped us to change people’s habits.
- Be approachable and make time as a data team to help colleagues. Our drop-in sessions had a huge impact. People grew to feel comfortable approaching us with their questions, and we bonded as team members—all of which made the teaching and learning process easier.
- Continue to share Analytics charts and encourage people to use the platform to uncover data-driven answers to a question. Sharing data and insights before implementation can offer people a glimpse at the possibilities. Once it is widely available, employees are more eager to access the data.
- Speak to people in sales, product, and business units to learn the specific type of data they need to achieve their respective goals. There is a tendency to want to track everything, and it is a common belief that more data is always better. However, a use-case-driven approach is more effective. Ascertain what issues people are trying to solve and then decide what data to track.
- Be patient. Changing habits takes time, and people often need a number of initiatives and reminders to become data-reliant and effectively use data insights. It will happen, and the results are worth investing the extra time, kindness, and patience.
There is a tendency to want to track everything, and it is a common belief that more data is always better. However, a use-case-driven approach is more effective.
Unlocking the possibilities at Unmind
Analytics has been valuable to Unmind in several ways. One of those is monitoring the metrics that the product development teams aim to improve. For instance, we have a chart that tracks stickiness: our daily active users (DAU) over monthly active users (MAU) ratio. The product team monitors that metric weekly and uses it as a North Star when releasing new features.
One use case that is particularly popular at Unmind is the funnel view of conversions. We can see if when a user clicks on item A, they also click on item B. Event segmentation is used widely, and we also use the platform for A/B testing. Viewing tests in Analytics makes it easier to see results and compare groups. These results inform our teams which feature or style is more impactful so we can make critical product development decisions.
We have made big changes to our products based on insights from Analytics. For example, we used to have a bot in the app that acted like a companion and would answer user questions. Analytics showed us that when the bot was involved in our courses, people were more likely to drop off. This data contributed to our decision to remove the bot from the app.
Another example of an actionable insight is our examination of predictors of retention. We saw in Amplitude that people who used our Wellbeing Tracker (a scientifically validated questionnaire that helps them understand their current state of well-being) tended to return to the app. We validated this finding using advanced analytics and this was a breakthrough moment for us at the time — it meant users understood the value of our product, and we should take extra efforts to highlight that feature. As a result, we began encouraging the user to complete the Wellbeing Tracker during their initial Welcome Tour of the app. We also changed communication in our onboarding email to include the Wellbeing Tracker.
Where team members were once hesitant to use Amplitude, they are now proud to use it and share insights. Sometimes people will notice a trend in Amplitude, ask us for our opinions, and learn how they can dig deeper. Amplitude integrates with Slack, which allows users to share charts, and people can engage with it in the channel. It is highly collaborative, and that enthusiasm spreads quickly and encourages others.
Changing habits, changing culture
We have seen data-informed decision-making emerge as a vital part of our company culture. We can now more easily understand the impact of a product change, and when we can see the impact quickly, we can move to more meaningful discussions faster. Data helps drive new product development and other proposed changes, making it simpler for everyone to move in the same direction.
The number of data requests we received has decreased since our Amplitude rollout. This decrease freed up the data team to do more advanced analytics, improving both our velocity and impact.
Ultimately, Amplitude helped us progress in defining our North Star metric, and we now take a more bottom-up approach to product development. Being able to explore the data and discover trends allows us to look for opportunities to improve our product in ways that will resonate with our users and empower them to lead more fulfilling lives.