How Holistics uses Snowplow data to optimize their data modeling and self-service BI platform
Self-serve data is the holy grail for many organizations. But achieving it is challenging. Making data readily available for teams and individuals to ‘help themselves’ to relevant segments of data comes with a host of complex challenges – not least of these is building the right data models so other teams have access to data that makes sense for them.
For Holistics, a BI platform designed to make self-serve data a reality for other organizations, the task of democratizing data was no less challenging within their own team.
To get there, the team at Holistics went on a journey to build a self-service platform that would enable their product and marketing teams to ‘fish for themselves’, gathering the insights they needed to develop Holistics into the best-in-class platform for data modeling and BI.
Holistics at a glance: Holistics is a data modeling & self-service BI platform that enables non-technical users to get access to analytics without bothering data teams. Once raw data is stored in a data warehouse, data teams use Holistics to transform, model the data and prepare well-defined datasets and reports for key data consumers.Holistics serves data teams in over 30 countries, with offices with offices based across Singapore, Ho Chi Minh City, and Jakarta.. One of the key innovations of Holistics is their SQL-based data modeling layer, where it allows data users to use familiar and powerful language (SQL) to define a central definition of metrics. These are then exposed to business users in a clean, simple interface to use, enabling them to get accurate analytics data without the need of an analyst. |
The challenge: tracking across product, app and website
Holistics is a platform that empowers internal teams to discover data themselves, without relying on analysts to build dashboards for them. Holistics makes it possible to dive deep and explore data, gaining insights they can use on the fly.
At Holistics internally, there’s a lot of behavioral data that needs to be captured, modeled and analyzed. With two main properties that need to be set up for tracking: the core Holistics product and the Holistics website, plus a third application called dbdiagram, the team needed a way to understand how users interact across these platforms and be able to serve data to the product and marketing teams respectively.
Initially, the Holistics team began with basic tracking with Google Analytics. But after deciding to track user behavior within dbdiagram, they wanted to explore alternative tracking solutions to give them better oversight of interactions within the product.
First, they weighed the benefits of building their own tracking infrastructure with solutions like Segment and Snowplow. Being open source, the team at Holistics were attracted to the idea of testing out Snowplow themselves without a financial commitment or budget approval from management.
Taking control of their data
To reduce risks during the transition from GA, Holistics decided to experiment with Snowplow on dbdiagram before potentially rolling out tracking to their main product and the marketing website.
Snowplow was already in place when Ha Pham joined as the company’s first full time analyst. He found there were major discrepancies in the data provided by GA compared to the data the team were tracking with Snowplow.
“Our bounce rate was very different between GA and Snowplow. GA was showing a very high bounce rate of around 60%. Recalculating the bounce rate using events from Snowplow, where we could really control the definition of what a ‘bounce session’ really means. The bounce rate from this calculation was reduced to between 30-40%!”
Ha Pham – Data Analyst, Holistics
Diving deeper into why these discrepancies existed, Ha first modeled the data into page views but later found it made more sense to aggregate the data further into the session level and then into the user level. One area where the gap was particularly stark was the bounce rate, where GA showed the bounce rate to be as high 60% but Snowplow showed a rate of 30-40%. With the ability to define this himself, Ha could calculate the bounce rate much more accurately, reduced down to almost half their original number.
Considering Holistics is a B2B SaaS website, Ha expected the typical user interaction to be very different to that of an Ecommerce website, so GA’s precalculated methods of measuring bounce rates were less likely to be reliable. In this instance, it made a lot more sense to work with Snowplow for web analytics. Snowplow now serves as the main tracking solution across Holistics’ marketing websites, where access to raw, granular data makes it possible for internal teams such as product and marketing to self-serve using Holistics datasets.
Refining the Holistics product with behavioral data
Tracking and analyzing website data is only part of the story for Holistics. As a data software company, they need a way to keep on top of their customer-facing product to identify what product features resonate with users, which don’t, and which need tweaking.
For the Holistics team, building a game-changing product is as much about eliminating unnecessary features as it is about building new ones. Nowhere is this more true than in the Holistics UI – the heart of the Holistics experience that makes self-serving data a breeze.
With Holistics, once your foundational data models are in place, it’s possible to drag and drop different datasets to build a view of certain metrics, geographies, usage and more without having to write complex SQL queries. This is especially powerful for marketing and product teams who can fish data for themselves.
You wouldn’t know it, but hidden behind their clean UI, a great deal of behind-the-scenes work has gone into understanding how users interact, what they click on, what they don’t click on, and how they generally progress on their analytics journey through the Holistics platform. These insights have allowed Holistic’s product team to ensure the platform stays minimal and doesn’t suffer ‘feature bloat’.
“We checked for what pages people visit the most within the product, and what buttons or other elements people click the most (or least!) and used this information to eliminate unnecessary elements within the UI to revamp and introduce more useful functions.”
Ha Pham – Data analyst, Holistics
To get these insights, the Holistics team needed access to raw data that they could model and serve to the product team so they could turn those insights into a better experience for users. With help from Ha, the team could dive into Snowplow data to identify the least-clicked buttons and product elements by users, enabling them to inform their approach to streamlining the UI.
After these enhancements, the change in the look of the UI was striking. Moving away from having lots of buttons and different functionalities, the current version is a lot cleaner and less cluttered which makes for a smoother user experience overall. Using their own platform in tandem with Snowplow’s well structured data, the product team now has what they need to continually optimize the Holistics experience.
“One of the advantages of Snowplow is that it gives us a lot of raw data, but organizing and structuring that data in a way that makes it easy for the rest of the company to consume is where the synergies between the two products come into their own.”
Evan Tan, Chief of Staff, Holistics
The future for the data team at Holistics
For Ha and the Holistics team, there is a bright future ahead of expanding their data capabilities and experimenting with new approaches. One new avenue Ha would like to explore is implicit tracking – capturing some data automatically to see if additional elements and events will lead to more insights.
“When someone asks me a question, I write down the hypothesis, the user flow, I follow the flow in the events, and I identify the elements that are interlaced with that event and I track them explicitly.”
Ha Pham – Data Analyst, Holistics
As for Holistics, there continues to be a strong data culture where product teams and data analysts like Ha work together to build frictionless data exploration experiences for organizations around the world.
Snowplow and Holistics: a powerful combination
Snowplow and Holistics can make for a powerful combination. With Snowplow, you have access to rich events made up of over 130 data points, captured in a way that gives you total control over how you use it.
To learn more about using Holistics and Snowplow together, check out their example of modeling Snowplow data in their Analytics Setup Guidebook.
Click here to learn more.
Working in tandem with Holistics’ intuitive self-service platform, organizations can deliver granular behavioral data across the business, where it can be put to work – making an impact or allowing them to explore to uncover key insights.
“There is something to say about the ability to persist data back to the data warehouse, which means that using the Snowplow raw data with the Holistics modeling layer, it’s possible to create a calculated field or pre-aggregated metric, based on the raw data. This allows you to work with ‘calculated’ raw data into a certain metric to make sense of the data. This flexibility to create new definitions, adjust them and update them in a specific place so it can be socialized across the company.”
Evan Tan, Chief of Staff, Holistics
That’s where Holistics technology shines: giving independent teams the ability to dive into and explore data further until they stumble upon that ‘aha’ moment that can be instrumental to innovation and business success. And while providing a seamless ‘self-serve’ data experience is still a logistical challenge, solutions like Holistics and Snowplow can go a long way to help data teams turn data democratization into reality.
