Our Services
Data Science Capabilities
Our data science experts build and deliver predictive and optimisation models tied to a business outcome, and put them reliably into production. We build the capability inside your team as we go, so the work continues long after we leave.
The problem we see
Forgotten value of decision-making
Focus on generative AI forgets the value of data science
Models in a notebook never reach the business
No dependable way to run, monitor and improve models in production
How we do it
Predictive & Optimisation Models
We invest time in identifying a specific business outcome rather than a vanity metric and build models that prove that outcome.
Experimentation
We test ideas quickly and honestly, so you invest behind the ones that work.
Model operations
We put the pipelines, monitoring and controls in place to keep models dependable in production.
Upskilling your data team
We work alongside your people throughout, so the capability stays with them when we leave.
How we deliver
We deliver with, not to
Shape Our approach is shaped with your people, so the thinking is yours from day 1
Build We build alongside your team rather than in a back room, so the know-how transfers as we go.
Hand over We step back and you move forward, owning the technology, the capability and the confidence to carry it on.
What you leave with
Models tied to measurable business value
A production and monitoring setup your own team can run
A data team more capable than before FOIL arrived
Case study | Data Science Automated Sales Forecasting
Luxury Outlet Retailer
FOIL transitioned quarterly sales forecasting with an automated, machine-learning tool that produces weekly forecasts and scenario plans. Cutting forecast cycles from weeks to minutes.
What we did
FOIL designed and built an automated sales and footfall forecasting solution for the client's budgeting and re-merchandising teams, replacing a manual spreadsheet process. A machine-learning model forecasts total footfall and per-brand, per-unit sales at daily granularity, learning from historical trends, seasonality, campaigns and moving events such as holidays. A web application surfaces these forecasts and gives teams a scenario builder, where they can add or amend leases, introduce brands, merge or split units, and see the forecast update in real time. Every change is auditable and exportable. FOIL built the whole solution inside the client's existing Databricks environment, so their data never leaves their own warehouses.
The impact
The tool gives the client faster, more frequent and genuinely data-driven forecasting in place of a process that consumed over 100 hours per cycle. Work that previously took teams weeks of spreadsheet effort now takes minutes, and forecasts run weekly rather than quarterly, with the capability to run daily. Teams spend less time crunching numbers and more on strategy: brand mix, re-merchandising and acting earlier on emerging trends. Scenario planning lets them test decisions before committing. FOIL expects this to support a sales uplift over time, where even a small gain is material at the client's scale, though that benefit is projected and still to be proven.
— Forecasting effort reduced from 100 hours to minutes
— Forecasting now weekly runs with daily granularity, in comparison to quarterly reviews
"It’s such a fundamental shift... Going from the spreadsheet you do once a year for a month on end to something that you can literally do in a minute."
Rob Davies
Strategist, FOIL Delivery Team
