Get the Inside Track of Tracking Consumer Behavior
The journey to a data-driven organization is a long and ongoing one.
And we could, of course, keep you in the dark of our progress, but where’s the fun in that? That’s why some of our Tracking squad members let you in on their daily work. Keep track!
There is no fast track to working in data. Just like all roads lead to Rome, all roads could also lead to the Tracking squad. Developer Alejandro Ivanez studied ship design, Digital Tracking & Analytics Specialist Haroun Benhalima held various marketing and sales consulting positions, and Product Owner Karine Caimo studied biology. And now they’re all enabling data-driven decision-making through tracking consumer behavior.
What the… Track?!
The team (13 people in total) tracks consumer behavior for almost all DPG Media brands. Over the years, they have harmonized their approach to tracking more and more. Many brands now share platforms and modules, and nearly all tracking occurs via Google Analytics and Snowplow, a behavioral data platform.
With the behavioral data, other DPG Media teams can optimize the user experience and take data-driven decisions. Like advertising teams, for example, they can build audiences to personalize ads. “We can measure almost everything,” says Haroun, “but measuring everything just because you can, isn’t very useful. So for each stakeholder, we jointly figure out what to measure based on their KPIs.
Take an editorial team, for example. They want to know how many people have read an article and how much time they spend on the page. But a product analytics team intends to optimize the websites and apps and is more interested in the performance of a button or the location of elements on a page, so we support them by setting up tracking so they can build their dashboards for AB-testing.”
Get the Data, Stat!
The team consists of both Tracking Specialists and Data Engineers, as they don’t ‘just’ track behavior; they also work on Google Analytics and Snowplow implementations, Kafka and Kinesis real-time data streaming, coupled with dbt and Snowflake integrations, and so on. Haroun: “The Tracking Specialists are a bit less technical and a bit more functional. They operate between business and IT as they have to translate the KPIs of the business into tracking implementations. The engineers bring the data in real-time to modern endpoints (where the data can be used by stakeholders) like Kafka and Snowflake. And they help the business create dashboards with the data.”
Most team members work on all platforms but usually specialize in one platform. Depending on the requests, they take up the work. And there are many requests! The Tracking squad prioritizes and adds them to their sprints based on each platform's roadmap. But of course, there are always additional requests - so that means more prioritizing and harmonizing to do! “When people see a drop or can’t find their data, they are quick to find us,” says Karine. “Often, it’s a human error, so we try to optimize our process to prevent these errors. To keep us from firefighting,” she laughs. It could also be a regression issue when new code got pushed to production without adequately checking the impact on the tracking setup. So proper regression testing is essential, and hopefully, one day, part of the definition of done.
Increasing Computing Power for Max Verstappen
The more traffic on our digital products, the more data. So when something big happens, like Max Verstappen winning Formula 1, the number of events (like clicks or page views) can easily double for a short while or several days. The Tracking squad needs to ensure that they still deliver real-time data during those spikes. So they automatically scale the computing power up or down to accommodate the variable load. Alejandro: “We open up more highways, so to speak, and increase the speed from 130 km/h to 250 km/h. To illustrate, normally we have about 15,000 events per second, with these big happenings, it’s 30,000.”
How to Access Data Responsibly
All information is stored in Snowflake, but accessing the data isn't easy, with around 700 million events a day. Alejandro: “We help our stakeholders navigate Snowflake and educate them on how to access data. Basically, the smart way is cheap; the easy way is expensive. People often pull unnecessary data, and that costs money. So we are working on improving the technical capabilities of the users by refactoring their moves and, for example, helping them with SQL and the Snowflake set-up to keep control of the costs.”
Karine quickly adds to the challenge: “In a data-driven organization, everyone needs access to data and wants dashboards. So the sky is the limit. But we do have to halt the enthusiasm now and then.” Haroun, too, recognizes the struggle: “Big data is a nice buzzword that costs a lot of money. But we don’t want too much data; we want relevant data.”
We help our stakeholders navigate Snowflake and educate them on how to access data.
Keep the Quality on Track
Working in the Tracking team is thus not purely technical, the trio agrees. Alejandro: “Once the pipeline is up and running, the job becomes more educational and supportive of people using the data. We explain how it works, how they can maintain quality, et cetera.” Karine sums up different meeting types: technical meetings on implementations and harmonization, stakeholder meetings on ambitions, and regular meetings such as refinements and retros.
“But it’s always a good mix,” says Alejandro. “In the coming weeks, we’ll be working on a sort of translator box between stakeholder requirements and Snowflake. The ‘box’ will be a mix of dbt, Docker containers and Kubernetes cluster running on Datafly, for Airflow scheduling. In addition, we’ll use DataHub to document everything properly. And obviously, we’ll make sure that we pull data the smart way instead of the easy way.”
It’s these types of challenges that make Karine, Alejandro, and Haroun happy with their job. Haroun: “We have a little bit of everything. On the one hand, we have to understand the business, but we also need to translate business insights into something technical. So you learn a lot about data, pipelines, data structures, data design. This is very challenging but very nice at the same time.”
Both Alejandro and Karine like that the work is highly visible and offers immediate results. Karine explains: “You immediately see what your data is used for. I saw a big evolution within DPG Media over the years; we’ve become more and more data-driven. Sometimes with giant leaps, other times with small steps - and a hiccup here and there. It’s super challenging.”
So, if you’ve read the Tracking squad’s story and now think: hmmm, this might be a good fit for me… You’re probably right. Haroun: “It’s challenging, but we have a super safe and nice environment. There’s an enormous amount of psychological safety in our team. Everyone is super approachable, and you can ask everything. Everyone will always make time for you, which helps you learn quickly. As a consultant, I've experienced different environments, and it’s super amazing to find a team like this.”