top of page
  • Mal McCallion

Will Rightmove plug its data leak?

Updated: Dec 11, 2023

Ever since it started, the property portal has not been great at stopping others using its valuable data. As data becomes more valuable due to Generative AI, will it find a way?

Zoom tried to do it (then reverse-ferreted and looked terrible in the process).

Is Rightmove about to follow?

This is the current frontier in the Generational AI wars - just where are OpenAI, Meta, Google and the gang going to get the next datasets to train their precocious children? Suddenly all of the significant sources - including Reddit and StackOverflow - have realised the value of what they have, and are only willing to share it if there's a cut of future profits for them too.

Similarly, organisations that have hitherto been quite blasé about the vats of information that they have collected over the years, are suddenly starting to look at them as a potentially welcome revenue line on their annual accounts.

For Rightmove the issue is made rather complicated by the fact that this isn't, actually, their data. It is collected via feeds from their estate agent customers. However, that has never ever stopped them from threatening (15.7.1 here, since you ask) others from using this data for their own commercial ends.

Arguably, this is unenforceable as the data is publicly available via thousands of estate agent sites and they're licensed to use it themselves. However, this is guaranteed to be an area that new CEO, Johan Svanstrom, has at the very least got a report on his desk about.

I wonder if it might be too tempting to give it a go?

Made with TRUST_AI - see the Charter:

12 views0 comments

Recent Posts

See All


bottom of page