Remove the Bottleneck and Get the Most Out of Data with Self-Serve Analytics

Martí Colominas

Head of Data at Kahoot!

people reacted
6 -minute Read,

Posted on September 10, 2021

In a company with more than 170 employees, more than 150 use Amplitude to get statistical data and insights on how customers use Kahoot!

Kahoot self-serve analytics

Founded in 2012 out of a research project at the Norwegian University of Science and Technology, Kahoot! has grown to serve more than 550,000 paying users. In 2020, more than 250 million games were played on Kahoot!, with over 1.5 billion participating players in 200 countries.

Kahoot! is an all-encompassing platform where users can create, share, and play learning games or trivia quizzes. The company is headquartered in Oslo, with offices in the U.S. and throughout Europe.

Over the past nine years, Kahoot has experienced phenomenal growth. Amplitude has been a critical partner, helping us get the most out of our product usage data as we have scaled.

The Steps We Took to Improve Data Governance

I began working at Kahoot! in 2018 as a data scientist. Before Kahoot!, I worked at a well-known Norwegian media company called Schibsted. We used Amplitude at Schibsted for global components in the Marketplaces division. We set up a system that leveraged Kafka alongside the internal event tracker to feed Amplitude with these events. I knew we could make similar strides with Amplitude at Kahoot!.

Kahoot! had already adopted Amplitude, but we were also still fairly small as an organization, with around 50 people. Kahoot! had started monetizing by that point with paid subscriptions, but as far as Amplitude was concerned, we were only using a few functionalities and tracking minimal events.

Everyone talks about big data, machine learning, AI, etc., but as the industry matures, we can’t forget about the basics. People preach about the importance of data governance, but at the same time, organizations still suffer from that classic bottleneck where all requests go through the data analysts. A large part of our success in how we’ve scaled with Amplitude is that we put a lot of time into creating self-service analytics that allow people across the company to use product data for analysis by themselves. Today, our core team comprises around 170 employees. Of that, about 150 have an Amplitude account, with 100 of those being monthly active users.

To set yourself up for success, you need to put in the effort upfront on data governance. We began by setting clear naming conventions, making unnecessary information invisible for the end user, and considering which properties had different values and needed to be merged. This is simple, even boring, stuff, but it makes life so much easier. It also ensures you can trust your product analytics data and that your end user doesn’t need a background in data science to understand it all.

So many companies want to skip ahead to the data insights, but if you don’t have a good data governance framework, the insights aren’t as useful. Click To Tweet

It’s important to understand what questions you want to answer, and then build from there. Make tracking plans. Again, it’s not flashy, but it’s so critical, because it’s the foundation to any future data work you do. So many companies want to skip ahead to the data insights, but if you don’t have good quality data—as well as data governance—the insights aren’t as useful. If you take the time to plan everything well, then the data becomes much more powerful.

When People Understand Data, They Drive Growth

Once we’d shored up our data governance framework and people saw the value of having aggregated usage data available to them, they were much more eager to use the platform. But helping everyone in the organization realize that value took some coordination.

Within the almost three years I’ve been with Kahoot!, I’ve seen the company grow from being a quiz and social tool to being a full-blown learning platform. Half of all teachers and students in the United States hosted or played a Kahoot! game throughout the pandemic, and 97% of Fortune 500 companies use Kahoot! for training, onboarding, presentations, and events. We’ve gone from tracking a small number of interactions to working with a massive, super-rich metadata set. It’s quickly become imperative that everyone in the organization can access and use that data with Amplitude in order to understand customer behaviors and drive growth, loyalty, and adoption throughout their verticals.

Amplitude allows many more people in the company to be self served when it comes to analytics. That means that engineers, PMs, and marketers can understand how our users use the product or how the features we release are performing without having to rely on the data team for everything. Product managers can now make product decisions based on what they see happening in the product via Amplitude. Many functionalities have been improved, such as moving buttons or making them more visible, and these decisions were made quickly.

With Kahoot! experiencing such explosive growth, we’ve incorporated Amplitude into training for new hires—and we use our own platform to make that training fun. We provide ad hoc individual support to ensure that everyone understands how to use Amplitude. It might seem counterintuitive to allocate additional resources to individuals when we want to have more of a hands-off approach, but we’ve found that it pays off for everyone at Kahoot! to have a good start with Amplitude within two weeks of joining the company.

Today, Kahoot! has over 170 employees, which swells to more than 500 employees when you include our recent acquisitions. Because of our initial work on data governance, we see the importance of data analysis embedded into the company culture. Amplitude is now part of Kahoot!’s DNA.

Fast Insights Make Way for Data-Informed Decisions

Amplitude crunches data super fast, which is important because we have a lot of traffic and we need to get to those insights quickly. It’s also very easy to use, allowing users to build funnels and segmentations, and it syncs well with Iterable, which has made Amplitude our hub for behavioral analysis.

We also leverage Amplitude’s monitoring and alerts to tell us when something’s broken. Amplitude monitors transitions from one place to another that other monitoring platforms don’t catch, such as if customers get redirected from a pricing page. We’ve caught some bugs that way that we wouldn’t have seen otherwise.

Our main users of Amplitude at Kahoot! are product managers and engineers, who monitor metrics around their releases. The second big group consists of our marketers, who monitor pricing pages and upgrade flows inside the product to ascertain whether any changes contribute to higher conversion rates. They also monitor for general operational and performance issues and run A/B tests.

It’s impossible to overemphasize the importance of self-service analytics. By doing so, you can remove the classic data bottleneck. Click To Tweet

It’s impossible to overemphasize the importance of self-service analytics in all of these cases. We’ve removed that classic data bottleneck you see at so many companies, where it takes two weeks for someone to receive a report. The immediacy of insights sparks further curiosity. People want to take a deeper dive into the results, which means our Amplitude users become more competent in the field of data analysis.

Now, whether it’s the product managers or the engineers, the marketers or the customer support team, they feel they need Amplitude to carry out their jobs. And because product decisions are informed with data, that data has become a critical part of product management. Aggregated product data about usage and activity is no longer thought of as the cherry on top of the cake—it’s the entire cake.

Time to Take on New Projects

Since implementing Amplitude, the requests to our data team are down by about 80%. This is because the self-service nature of Amplitude means that my colleagues can get answers to their questions right away themselves, without having to submit a request to the data science team. People can see for themselves the number of users who clicked on a particular button or viewed a particular page or how the conversion rate changed between option A and option B.

Because we don’t have to do that anymore, our data analysts have time to focus on the other 20% of requests, which are much more complex queries requiring us to write SQL and transform the data. One of the projects we took on in our newly freed time was to create a churn model and test it in Amplitude. That model has resulted in a 20% drop in churn.

The quality of Amplitude’s UI means that now our production managers can see that they need to move a button, for example, or they need to make it more visible. Before, they wouldn’t have understood that without the help of a data scientist. Amplitude’s UI makes that easy, and our product managers also now understand the reasoning behind these decisions the way a data scientist would. We now use Amplitude to monitor every single release we push out to market.

The past 12 months have been a year of change for everyone, with people using Kahoot! to stay in touch and stay engaged in classrooms and living rooms worldwide. Amplitude has given Kahoot! an edge when it comes to making learning awesome for our users.


Learn more about self-service analytics by signing up for AmpliTour, the hands-on Amplitude workshop.

Martí Colominas

Martí Colominas heads the data team at Kahoot!, based out of Norway. Previously he worked at Schibsted as an insights data analyst for more than 20 different Marketplaces across Europe, Africa and America.

More from Martí

Inside Amplitude