Summary

After I got exposure to the LeadGen business with ColdFire agency I got in touch with Alex from CyberLeads. I really liked his business model (recurring and a productized service, meaning relatively low-tech). For another venture (LifeSketch) I was looking for leads but couldn't quite narrow down his database of prospects because he had no Filter option for how outdated a website (design) was. This would be very helpful as I (as the owner of LifeSketch, a design agency) was interested in B2B companies with outdated (and ugly) websites. This sparked the idea to build a web scraper that could somewhat objectively quantity the beauty of a website. The business model would then be: sell the ugliest websites on the web (with KDM contact info) to design agencies as high-potential leads.

The idea
The idea was to score a website 0 - 100 (100 being like Apple.com) and derive that score from a bunch of variables. Some inputs were visual, some were grabbed from snippets in the HTML and some were based on plain website copy. This was before the launch of GPT and other sophisticated AI models, so we ran various tests using our hard-coded algorithm.

With mixed results we paused the venture as we were unable to sell any subscription. With AI models however I'm sure the quality of the Design Score could be much higher and this data could still be sold as a signal to businesses or get sold as a stand-alone attribute via API to large database.