Accurate revenue forecasting is crucial for Marley Spoon's business. As the company expanded, the business analysis team started reaching the bottleneck of spreadsheets:
Dina, Marley Spoon’s business analysis lead, was looking for a solution that would be flexible enough to support a bespoke and complex revenue model (including cohorts) connected to their data warehouse, while easily supporting multiple versions and scenarios, and being able to present directly to stakeholders.
Marley Spoon’s model is using Causal’s live data integrations to pull 2 years of weekly cohort data for each country from their data warehouse.
Causal automatically refreshes the actuals on a weekly basis, and automatically compares them against different saved versions of Marley Spoon’s revenue forecast.
Our customer success team worked with Marley Spoon to build out their full revenue model in Causal. The model is at a weekly granularity, tracking each weekly cohort separately, spanning 6 years.
The model also spans 8 countries and 2 brands, letting Dina and her team drill down into any of the 12 combinations of 'country' and 'brand', as well as easily see aggregated views across the categories.
The total number of formulas required in the model was reduced from 6.5 million in Excel to a few hundred in Causal (a 10,000x reduction). Dina’s team are tracking multiple versions of the model separately using Causal’s Version Control functionality.
Causal’s modelling functionality and data automation save Dina’s team 30 hours per month. This has freed up their time to focus on the strategic work that adds value to the business — analysing performance and working with business partners to make better decisions.
With drastically fewer formulas and dedicated functionality for scenarios and versions, there’s much less complexity to manage in Causal, making it much less likely to make hidden errors. Causal’s plain-english formulas also make every model auditable by anyone, significantly reducing the chance of formula errors and reducing the time required to double-check work.