Logo y42
Valentina Perezalonso
Valentina Perezalonso
Content Marketing Manager
26 Jan, 2022 · 6 min read

A playground for data nerds: Oatsome and their data journey

Learn how Oatsome built their own marketing attribution model and customer prediction platform from scratch with Y42.



Food & Beverages


Frankfurt, Germany

  • Y42 allowed Oatsome to accelerate daily business with reduced engineering and loading times. 

  • Through the implementation of data reports and dashboards with Y42, Oatsome instilled a data-driven culture and allowed for a better understanding of the business and performance across the board.

  • Y42 brought team members closer to the topic of data and enabled transparency and clarity within the company.

The company

Oatsome is a German startup that went into business in 2017 with the goal of making breakfast healthy, quick, and tasty. Their products range from smoothie bowls and toppings to bliss balls — foods for both the breakfast and snacking markets.

As a DTC e-commerce, Oatsome works with data on all fronts: data on the behavior of customers on their website, the performance of their social media ad campaigns, the revenue generated through affiliate marketing, how products are performing in the market and which products to roll out next, the logistics behind shipments and providers, and much more. All this information needs to be processed and made digestible for business insights, and that’s where the Data and Analytics Team at Oatsome comes in.

Business challenge

Like many startups, Oatsome had one ambitious goal in mind: make the whole company data-driven. But who would be responsible for this daunting task? Well, the small but ambitious team who loved data, of course. Despite having a data scientist in the team, Oatsome wanted to have a business-user-friendly tool that could be implemented and deployed by an analytics manager. The idea behind this was to allow the data scientist to focus on more complex tasks.

They started with the basics and went for Google Analytics, but problems quickly started to arise — namely, how to incorporate third-party tools. So a change was made and they started building their own data pipelines which went into Google Sheets. Soon enough, they had 3,000 sheets, and nobody had any overview of where everything was, while the cell limits and API connectors were being exceeded.

The solution for this spreadsheet hell? Implementing a data warehouse. But the challenges didn’t end there. Oatsome wanted to build their own marketing attribution model (MAM), customer data platform (CPD), and customer prediction platform (CPP), then take this data, learn from it, and trigger automations.

At that moment, they understood that the data warehouse would be the basis for this data infrastructure, but a core building block was missing: a tool to put all these models together.

Why they chose Y42

The road to Oatsome’s ideal data tool was clearly outlined. They understood that millions of companies offered data pipelines that connected to BigQuery from anywhere; they also knew that another bunch of companies offered attribution models, while others could create customer data platforms and customer prediction models as well as trigger automations. But there was a catch: Oatsome wanted to have everything in one place. They also wanted an overview of their data lineage and the ability to understand and control what happens between stages, where data comes from, where it goes, and how it’s transformed.

In a twist of fate, Y42 reached out to them and Oatsome met their perfect match. In Y42 they found a data playground where the models and concepts of their dreams could be built and brought to life.

Till Körner
Till Körner
Data Analyst at Oatsome

"My favorite things about Y42 are the flexibility and the user interface. It enables us to easily show other team members how models work and what’s happening inside of them, making data accessible to anybody, regardless of how deep into the topic they are."


Influenced by their past experiences with performance marketing and the pain points they were aware of, Oatsome’s first use case with Y42 was a marketing dashboard that encompassed all their marketing data points like Facebook, Google, Pinterest, and Taboola. With this dashboard, Oatsome wanted to understand and visualize the company’s influence on the market as well as marketing costs and revenue, while making this information accessible to everyone.

With the first dashboard in place, the marketing team’s next step was to build their own MAM from scratch in order to allocate each customer purchase to the exact channel they came from. For this, they needed to pull data from Google Analytics on their customer visits, from all their marketing campaigns on social media and Google Ads, and lastly, from their client base on Shopify. The sales data was then attributed to the customer visits data, and each individual visit was quantified based on micro-transactions. The marketing costs were then added and the model was completed.

With the MAM ready, a CDP (a place where all past client data could be stored) was in the works. These two models would then be the basis for a CPP (a space to predict future customers’ behavior and turn these insights into actions). This prediction platform is currently being developed by Oatsome’s data scientist.


With data work being initiated on the marketing side of the company, soon enough, dashboards and reports were being expanded across Oatsome. Y42 enabled every team to obtain answers to questions that only data and visualizations could resolve.

This has meant that everyone in the company, from the CEO to the last customer support intern, has access to data reports, allowing them to see what’s happening in different departments and how things are developing, which creates a lot of transparency. For Oatsome, the aftermath of Y42 is a much better understanding of how their business works as well as the ability to grasp which KPIs impact the business. In turn, this has given the team foresight of changes or problems and the ability to prepare for tackling them in advance.

Now, even those team members who have historically been far away from the topic of data can use it to structure their work — from the product team asking what their net promoter score will be before rolling out a new product and what effect the launch could have on other parts of the company, to the marketing team getting a better ROAS and shifting their budget allocation thanks to their attribution model and learning how to spend money smarter. Even the logistics team has gained a better understanding of an order’s lifecycle from placement to delivery.

Another great benefit for Oatsome has been a significant decrease in engineering time. Before Y42, it would take days to build custom API features for Google Sheets and work on Facebook integrations, but now all these things are just one click away. Load times have also been a game changer — before, they would have five million rows and would have to wait at least five minutes for them to load into Data Studio. Now they can load 25 million rows from BigQuery in a matter of five seconds.

Despite having encountered some initial problems with certain naming conventions in Y42 and the fact that team members are now breathing down the data team’s neck wanting to optimize everything because the reports have made them aware of every tiny issue, Oatsome have significantly improved their daily business. This is because they recognized where they wanted the company to go, how they wanted their teams to work, and took matters into their own hands by implementing the necessary solutions and the right tool to achieve their objectives.

Share this on

About Y42

Y42 is an Integrated Data Development Environment (IDDE) purpose-built for analytics engineers. It helps companies easily design production-ready data pipelines (integrate, model, orchestrate) on top of their Google BigQuery or Snowflake cloud data warehouse. Next to interactive, end-to-end lineage and embedded, dynamic documentation, DataOps best practices such as Virtual Data Builds are baked in to ensure true pipeline scalability.

It's the perfect choice for experienced data professionals that want to reduce their tooling overhead, collaborate with junior data staff, or (re)think their data stack from scratch.

Leave us a comment

This field is required
This field is required
This field is required