How we automated Costa Coffee's demand planning
As Costa Coffee expanded across multiple regions, manual demand planning slowed the team down. We helped them automate the process, cut hours of work, and gain reliable insights at scale.
The challenge: Time-consuming demand forecasting
Costa Coffee had been using Anaplan, a cloud-based connected platform for demand and supply planning, for about two years. However, most of their processes related to Anaplan were still manual. This was particularly time-consuming for their demand planners, especially as the company was rapidly expanding across Europe, the UK, and MENAT countries.
They needed clear, reliable insights to spot trends and plan inventory effectively, without having to spend hours in spreadsheets every day.
The key challenge was to organise and automate the way they collect, prepare, and process data related to demand planning.
What we did: Transforming data planning through automation
We kicked things off by supporting Costa’s existing demand planning app and laying the groundwork for automation. First came a deep dive into their data – a full discovery phase with the client’s team to understand what they had, what they needed, and how to bridge the gap.
Then we got to work:
- Mapped and defined the necessary data transformations
- Built a rock-solid pipeline for automatically pushing data into Anaplan
- Cleaned, organised, and prepped everything to reduce friction down the line
From there, we moved into implementation mode – automating processes that used to take hours and cutting out the manual work entirely. Using Azure Data Factory and Databricks, we built scalable, future-ready data flows, while Power BI dashboards gave everyone a clear view of what was happening behind the scenes.
Today, most of the data is already live and flowing directly into Anaplan. The rest is in testing or being fine-tuned for rollout.
The result: Hours of manual work reduced to minutes
The biggest improvement? Time.
- Demand planners now spend significantly less time on manual work
- Tasks that once took several hours are now done in under an hour
- Data is automatically pushed into Anaplan, reducing human error
The number of mistakes has dropped thanks to consistent, automated workflows
The tech stack
- Azure data factory
- Databricks
- SQL
- Pyspark
- PowerBI
