Prospector

Mining 2.5 million Companies House records for high-probability leads

Prospector

The Opportunity

With large open source company datasets available, generating B2B leads shouldn’t be difficult for established businesses.

The advent of machine learning has created the opportunity to apply the rigour and intelligence that marketing teams expect of lookalike audiences, to untapped sources of accounting information from the UK Companies House dataset.

What we did

✔︎ Created an algorithm to intelligently match each of our client's top-tier customers to its official Companies House record.
✔︎ Built a bespoke machine learning model that learns the defining characteristics of those customers from accountancy and company features.
✔︎ Applied the model to the full Companies House dataset to predict a prioritised list of high-value active accounts.

The Results

With a list of your existing top-tier customers, the Prospector machine learning model predicts which of the 2.5 million active accounts within the Companies House data have the highest probability of becoming top-tier customers for your business.

The output from Prospector is a detailed report of high-potential accounts to target and which existing customer each most closely resembles.

Prospector gives your company a personalised view of which UK businesses are mostly likely to become your next top-tier customers.

2.5M

active company accounts analysed

How we did it

To get started, you only need to provide us with a representative sample of account names and postcodes for each of your top-tier accounts. For example, an Excel file transferred to us through our secure file transfer portal.

The Prospector algorithms work with this data to match each account to its official Companies House record. It can handle poor quality data (e.g. duplicates, spelling mistakes) and the model is designed to learn only from accounts where it finds a match with high confidence.

Prospector then trains a cutting-edge machine learning algorithm to learn the defining characteristics of your high-value accounts, based on features generated from the matched Companies House data. Once the model is trained to a suitable level of accuracy, we score the entire Companies Houses dataset, returning the prioritised list of high potential accounts that are currently not part of your customer base.

We then fine-tune the model based on automatically generated Prospector reports of the matching and model building process, to incorporate features that are unique to your customer base and requirements. The output is a list of high-probability, high-potential accounts, and accompanying detailed report from the data exploration phase.

Brochure

Download the guide

Contact

Start a conversation

Take the first step by speaking with one of our data experts today.

Flux
Anti-Money Laundering