The discovery problem

BloomReach was started to solve the problem of discovery on the web.  BloomReach cofounder, Ashutosh Garg, had worked in the Google search quality team, and had realized that in order to improve the quality of search experience on the web, you needed to not only continuously improve ranking algorithms but also to help improve the relevancy of the pages being created on the web.  Meanwhile, Raj De Datta had been talking to online marketing executives about their challenges getting all of their content discovered.  He had found that companies that focused on technology but didn’t sell products or services directly, or those with large internal advanced technology groups clearly had the advantage over companies who were primarily focused on selling products and services in retail, travel, education, etc.

BloomReach was created to make web sites more relevant and help level the technology field for companies selling real products and services.  Our goal is to improve the user experience on the web and improve our customer’s revenue and the cost effectiveness of their acquisition marketing programs. This is why we created our Big Data marketing applications powered by our web relevance engine.

While we initially focused on organic search, we believed from the beginning that discovery is a problem that exists across all marketing channels. One simple way to illustrate this problem is to look at the percentage of pages on a website that are getting visits via a channel. If each of your pages is unique and in demand, then you should receive some visitors to these pages. If you do not then you have a discovery problem.

We looked at this metric for some of our customers across three channels: organic search, paid search and social. On average, we found that prior to connecting to BloomReach, only 25% of their pages were visited via organic search. The number was even lower for paid search (about 15%) and social (<5%). While the symptoms of the discovery problem are similar, the nature of the problem differs for each channel.  We’ll discuss these differences and how each channel needs different, but related approaches in future articles.