Semmx

What if Facebook and Wikipedia had baby, but built on machine learning?


Background

The technical founders and CEO initially engaged me as a consultant to help them design a product based on their information retrieval software. After a few rounds superficial redesigns and fixing glaring usability issues I convinced them the problem was more than skin deep, we needed to find and distill what the true value proposition was to consumers as information retrieval was a solved problem in the eyes of the consumer. At that point I joined full time as the VP of User Experience.

Problem

While the prototype was quite rough from a consumer perspective the founders had hit on an interesting problem and had begun to solve it with an impressive technology. Search is great for finding information when you know what you don't know. It's great for point knowledge (answers). However it has limited persistence and context. Search, and by proxy Google, is not so great at helping you find the answers when you don't know what you don't know. Semmx had the technology to solve that problem on many levels, but had the interesting problem of solving a problem most people don't realize they have!

responsive

This is the original engineer PUI, information retrieval focused "MVP". I was originally engaged to redesign this interface, but convinced the founders to rethink the value proposition.

Solution

Working with the CEO, utilizing secondary research, we identified an opportunity space in between existing search engines and existing social networks. Social networks are great for keeping up with social obligations (photos, jokes, events) but less useful for nurturing lifelong passions and interests. By adding a social layer to Semmx's always up to date topical index of the web, we could create the stickiness necessary to bring value to the consumer. We reframed the offering from a "better alternative to Google" to platform for following your lifelong passions and the people who share those passions.

TLDR: What if Facebook and Wikipedia had baby but built on machine learning? That's Semmx's mission.

responsive

To test the social theory, we built out a handful of separate vertical interest sites around broad topics: wine, spirituality, sports. These topics would normally be too broad to be useful from a social perspective, but because of our ability to generate topic areas computationally we could have a site that served people just starting to learn about wine, and orient them with other users with similar knowledge, while also allowing for more specialized spaces where advanced connoisseurs could follow trends and eventually create their own content.

responsive

After proving concept was possible we collected the subject areas under the InterestPlace brand and generated similar spaces for hundreds of other topic areas.


Back to all projects