Accelerating data maturity and enabling eCommerce growth
We built an ecommerce stack on top of Snowplow and Shopify, and delivered a highly sophisticated attribution model so that the Eve team would know exactly what marketing approaches were working well.
Eve Sleep is a UK-based sleep wellness company. Founded in 2015, Eve offers a full suite of sleep-optimising furniture and bedding through their direct to consumer (DTC) ecommerce shop. Only two years after its founding, Eve raised £30mn through its IPO on the London Stock Exchange.
Eve had scaled rapidly and had reached a pivotal moment of change: the company was looking to grow and expand their proposition in an increasingly competitive market. The leadership team had a clear vision for the future of data in the company, but their data infrastructure could not support the next phase.
As a company, Eve described themselves as data-rich and organisationally poor: data was generally available, but it wasn’t in shape to drive decision making. Marketing analytics were available via Google Analytics and while operational data was housed in Netsuite, but both were disconnected and insufficient. Acquisition data was disconnected from product and profitability metrics, hindering the team’s ability to make informed strategic decisions. The team needed to update its data stack to put insight and profitability into focus. Beyond execution alone, their data leaders needed an interim team to build a foundation and multi-year roadmap which could be implemented long after the engagement ended.
- Build a data stack with an integrated marketing platform that enables profit-based optimisation decisions
- Update data collection to support more granular insights into conversion and product usage
- Implement Looker as a self-service visualisation platform
- Accelerate the maturity of Eve’s data infrastructure and shape their multi-year data strategy
Data strategy & roadmap
During Sprint Zero we met with the broader stakeholder set (i.e., Marketing, Product, Data, and Leadership teams) to catalogue the range of potential use cases. In addition to scoping out our own work, we build a multi-year data strategy that Eve could implement after the engagement ended.
data collection & etl
The first stage of accelerating the stack was to ensure the quality and reliability of data inputs. We refreshed data collection from over 10 tools, including writing custom scripts to extract affiliate acquisition data. Using Fivetran for ETL, we aggregated this data into dbt for modelling.
Modelling with dbt
dbt served as Eve’s centralised, unified source of truth. First, we modelled out the business’ data ecosystem into a conceptual domain model that served as the foundation for data architecture. This allowed us to create and deploy two funnel models that allow Product and Marketing teams to identify key user journeys, with granular breakdowns by channel and user type, to optimise conversion rates.
self-service BI with Looker
We created a suite of marketing dashboards and explores that permit marketing users and agencies to dig further into the performance of each channel based on standard and custom attribution models. Each channel model reports on marketing stats like impression share all the way down to contribution margins (CM2). We worked with the marketing team to define the level of granularity and the type of dashboard required for rapid insights and optimisations in platform.
We delivered the data stack that Eve needed as a scalable foundation for the team to own, including:
- Fully functional, updated data stack reliant upon Fivetran, dbt, and Looker
- Single source of truth housed in clear, documented, centralised data models in dbt and Git
- Prototypes of core marketing dashboards that will allow Marketing and Product teams to generate conversion and product insights.