Problem
The client wanted a sales forecasting model more accurate and reliable than the one it had, and believed AI could help.
Machine learning is only as effective as the historical data available to train it.
The Proof/ Case study
Is sales forecasting machine learning-ready...or not?
How Acorn assessed whether machine learning could improve a sales CRM vendor's forecasts, and recommended waiting because there was not enough historical data.
- Industry:
- Sales CRM software
- Client:
- A vendor of B2B sales CRM software. Client name not published.
- Functions:
- sales
What Acorn did
The client's sales team provided a breakdown of the key processes and metrics they wanted to automate. Acorn used Bayesian inference to aggregate the historical sales data into a roadmap for a predictive model, including an analysis of critical factors in the sales process and how to implement the model in the client's CRM tool.
Our data scientists also cleaned up the data.
Result
Acorn recommended waiting to deploy a predictive model, because the sales team did not have enough historical data to inform a reliable tool.
We helped the team see the value in waiting while ensuring the model they funded was flexible enough to accommodate future customers.
Who this is for
Sales leaders considering a machine learning forecast who want an honest answer on whether their data can support one yet.