Using Mixpanel to accelerate data-driven decision making

Launched in 2002, T Map Mobility is Korea’s leading mobility platform with 20 million users and approximately 14 million monthly active users (MAUs). T Map Mobility provides various mobility services based on extensive real-time traffic information and big data. Among them, T Map Daeri is a drive-home valet service that maximizes user convenience and can be used directly from T Map Navigation without installing a separate application or signing up. With the vision of providing a safe and secure service for both customers and drivers, T Map Daeri has grown 30 times in the two years since its launch in 2021 and has been on a steep J-curve since July 2023, becoming Korea’s leading drive-home valet service.

The T Map Daeri team struggled with low utilization rates after the service was launched. The initial response to the anticipated service was much lower than expected, which came as a significant shock to the team. Data analysts were required to help gather new data on Logbox and Tableau, which slowed down decision-making at critical moments. The T Map Daeri team wanted to accelerate their data-driven decision-making process and increase utilization.

The team adopted an agile process to enable faster innovation, as well as partnering with Mixpanel, a data analytics tool to enable faster decision-making. With Mixpanel, they were able to monitor important data in real time without the help of data analysts. Product owners, product managers, and marketers can now proactively analyze data, conduct experiments, and quickly take action to improve metrics. As a result of these changes, the team was able to make focused and impactful decisions within tight timeframes, contributing to the rapid growth of T Map Daeri.

• Reduced human resources and time required for data extraction and dashboard building
• Data democratization that enables data access and analysis across all levels of the organization
• Rapid growth due to accelerated data-driven decision-making

Did you expect T Map Daeri to grow as much as it has?

Honestly, we assumed that just by announcing T Map Daeri’s services, customers would naturally start using them. In reality, we faced very disappointing results for the first 6 months. We were deeply shocked as T Map has a user base of 20 million, so we thought that just by launching the service, it would do very well. We didn’t even think marketing would be necessary.

So when the team who had such high expectations launched the service and found that it was not well adopted, they all felt pretty hopeless with the situation. As a result, the team morale at that time was very low, and the entire business, development, and planning teams had to reset. We started building a new team from scratch. That’s when I joined T Map Daeri as the Product Manager for the customer app, and then a year later, I took over as the leader of the planning team.

What changes did you make to overcome the crisis of T Map Daeri?

We started T Map Daeri more than five years later than our competitors, so we needed to move faster. Therefore, we tried various changes in our working methods, including adopting the Agile methodology. Despite already using Logbox and Tableau, the decision to introduce Mixpanel was also aimed at accelerating data-driven decision-making.

What was the most noticeable change you felt after introducing Mixpanel?

T Map is an app service, while T Map Daeri is a web service. Apps and websites have different structures and different data storage spaces. Understanding this structure allows for cross-analysis, such as comparing actions taken by users on the T Map app to those on the web. Therefore, this analysis was within the domain of data analysts. For product managers to delve into such detailed analyses seemed physically impossible. For example, in our Logbox, there were over 100 databases and over 100 tables within each database. 

After the introduction of Mixpanel, these tasks became much simpler. We no longer need to worry about other things; we just define user behavior actions, and that’s all we need to know. This aspect seems to be a very powerful and innovative point of Mixpanel.

What was the key to the growth of T Map Daeri’s J-Curve?

One of the biggest concerns at that time was that customers coming into the service were not being retained well. Thus, we analyzed how to retain customers from many different angles and discovered a key insight: using the service three times locks in retention. We found that the retention rate suddenly spiked for user cohorts that used the service three or more times. Once we came to this realization, we focused all our energies on how to make new users use the service three times or more by optimizing both the marketing and the product itself.

Tell us more about T Map Daeri’s loyalty program and how it helped improve retention.

When people hear about our points-based loyalty program, they might think, “Isn’t that just a common tactic that everyone uses?” However, from a management perspective, points entail significant financial burdens, which is why other similar services don’t usually introduce such a system. The reason we were able to boldly introduce a points system is because of the key insight we discovered that was mentioned earlier, “If a customer uses the service three or more times, they retain very well.” Additionally, after examining various data on retention and despite the financial burden, we were able to persuade management that introducing a points system was necessary for our growth.

What has been your favorite moment since implementing Mixpanel?

Recently, we launched a multi-request service. This service allows a customer to make multiple requests for drivers after making an initial request. Shortly after the service launch, we utilized Mixpanel to create 16 reports within a week, including reports that analyzed the paths customers took to make additional requests, how they accessed them, and so on.

During the retrospective with the team, we quickly realized areas where customers were using the service as expected and areas where they were using it differently than we anticipated. This allowed us to quickly reflect and learn lessons.

Previously, it wasn’t so much that it was impossible, but rather that the effort required to create dashboards for monitoring using tools like Tableau was too significant so we were not able to create dashboards for everything that the product managers wanted. However, after the introduction of Mixpanel, being able to freely create reports for everything we wanted to monitor whenever we wanted to was a huge improvement.

What teams in T Map Daeri are using Mixpanel the most?

One of the primary goals of introducing Mixpanel was not only for product managers but also for marketers to easily and quickly extract and analyze the data they need. Recently, we added a “share” feature to our pop-ups. While the product team initially thought, “Who would share a pop-up ad with friends?”, the marketing team believed, “If there’s a ‘share’ feature, customers will voluntarily share beneficial ad content with their friends.” There was a difference of opinion.

At that time, we decided to launch it and roll the dice. The day after the launch, the marketing team checked the performance of the “share” feature using Mixpanel, and it turned out they were correct – users were using the “share” feature. Seeing the team share their results with the entire organization made us feel like we achieved our goal of introducing Mixpanel, which was to provide our marketing team with a tool to easily analyze and validate data on their own.

What are your core KPIs and how do you analyze them?

As with any service, the key KPI for T Map Daeri is undoubtedly growth, and the metric representing growth is the “number of driver requests made per day.”

However, the process involves complex steps: requesting a driver, the driver accepting the request, starting the trip, and trip completion. This process typically takes about an hour. Therefore, in addition to the number of requests, we also track conversion rates for each step of the funnel mentioned above. For each of the conversion steps, we have a log of the various factors that we believe can influence conversion rates.

We analyze through Mixpanel how conversion rates at each step are influenced by various factors. For example, whether the customer pays by card or cash can significantly affect usage patterns and conversion rates. Similarly, there can be significant differences in customer behavior patterns based on whether the departure point is in Gangnam or a rural area.

Therefore, instead of simply analyzing “What was the overall conversion rate?” we utilize Mixpanel’s Breakdown feature to analyze “What differences exist by region, payment method, or customer experience?” This allows for a more nuanced analysis of the data.

How would you describe Mixpanel to others?

One sentence that I would use to describe Mixpanel is: “Introducing Mixpanel completely changes the speed of data-driven decision-making.” 

For instance, during a recent planning meeting, a developer suddenly raised a critical question about what happens if a driver fails to process the completion of a trip after completing the trip. Concerned about potential side effects, I quickly used Mixpanel to extract the overall distribution of trip durations until completion. I discovered that only 0.35% of trips remained incomplete after 2 hours. Armed with this data, we made an informed decision right there in the meeting that it was not a significant enough problem to solve. 

With Mixpanel, obtaining necessary data is as easy and quick as pressing a button on a vending machine. This tool significantly accelerates the speed of planning, decision-making, and agile processes. It’s a game-changer. It’s about having an environment where you can use data instantly for decision-making, rather than relying on a process where you have to wait a long time for data before coming back to making decisions. I believe there’s a huge difference between the two.

What is the vision for T Map Daeri?

The biggest concern for our customers is the anxiety of entrusting their cars to unfamiliar drivers. T Map Daeri aims to completely alleviate the worries of such customers. Just as people think of T Map when navigating, we want them to think of T Map Daeri when they need drive-home valet services.

Read the interview in Korean here.