Purplebricks & endjin

Streamlining business workflows through disciplined data science.

Purplebricks are a high growth 24/7 online estate agency offering to sell customer properties for a low cost flat fee. They floated on the stock market with a market valuation of £240m just 19 months after the company launched.

They also had a very unique meeting room decor.

Purplebricks Meeting Room Unicorn

They wanted to explore how machine learning could help the organization scale efficiently, as demand for scarce business resources was rapidly increasing.

By applying our structured approach to experimentation, and using Azure Machine Learning to rapidly develop and train predictive models, endjin were able forecast customer propensity with a high enough level of confidence for the business to invest further.

In doing so, we were also able to disprove a long held business belief about their data. Being able to rapidly prove or disprove business hypotheses enables organizations to zero in on the best solutions, avoiding unnecessary effort and cost of chasing hunches.

Endjin armed Purplebricks with the insight needed to effectively allocate resources, increase business efficiency, and help solve one of Purplebricks' most critical business challenges.

Machine Learning - the process is the science

James Broome02/03/2016

What do machine learning and data science actually mean? This post digs into the detail behind the endjin approach to structured experimentation, arguing that the "science" is really all about following the process, allowing you to iterate to insights quickly when there are no guarantees of success.

Machine Learning - mad science or a pragmatic process?

James Broome19/02/2016

This post looks at what machine learning really is (and isn't), dispelling some of the myths and hype that have emerged as the interest in data science, predictive analytics and machine learning has grown. Without any hard guarantees of success, it argues that machine learning as a discipline is simply trial and error at scale – proving or disproving statistical scenarios through structured experimentation.