In 2009, I had the idea of building a “Kayak for Everything.” Kayak was a great example of the power of a vertical search engine (VSE). While Kayak was great for travel, I felt the world needed VSEs for choosing schools, ski resorts, smartphones, dogs, and hundreds of other highly considered purchases. The Internet was (and still is) polluted with “the best of” sites that are simply fronts for making money. What consumers needed were VSEs that allowed them to discover what was best for THEM by filtering and comparing products and services with unbiased data.
Since Kayak, there have been some successful VSEs, including Zillow and Yelp. We believed that all VSEs were essentially the same; the data was different. Scott Leonard, a former partner of mine at DoubleClick, began building the platform while I was collecting the data. We outsourced data collection (ski resorts were our first topic) or had new hires gather datasets found online (school data was available from the Department of Education). We called our company FindTheBest.com. We used TeamStormIt techniques to brainstorm new topics and features.
We soon discovered that we could ingest almost any dataset into our VSE platform to make it accessible to consumers. For example, we acquired the Social Security Death Records database (yes, it’s called that), which was the basis of our ancestor site, MooseRoots. We added features like visualizations that publishers could embed. We changed our name to Graphiq since FindTheBest no longer fit the company.
It took years, but we became one of the largest collections of sites with hundreds of millions of indexed pages and helping 30 million users each month to make informed decisions. One day, Google announced their “knowledge graph (KG)” for their search product and Personal Assistant (PA). Though we didn’t call our large, interconnected datasets a KG, we realized WE had the largest KG. As PAs like Google, Siri, and Alexa grew in popularity, we knew we already had the most answers to any question.
We failed to get any of the personal assistant companies to believe we could help them. We had a hackathon one Friday where Dylan Wenzlau “ripped” out Alexa’s brain and inserted the Graphiq brain. In just a day, Graphiq was better than all of the PAs out there. Ivan Bercovich filmed the demo and emailed to Jeff Bezos. A few months later Amazon acquired us. Today, the Graphiq team powers virtually all of the Q&A on Alexa and other Amazon products.
However, after the acquisition by Amazon, we jointly decided to shut down the Graphiq sites in 2019. I still receive email from people begging for the sites to come back. Well, we have some good news. Scott Leonard is back to building a new and improved VSE platform while Kian O’Connor, my son, is acquiring datasets and building topics. TheDataProject.ai is their new company. Soon, consumers and researchers will have better access to more of the world’s structured data. Stay tuned.
Today, TDP launched it’s first topic: Mental Health Facilities. Mental Health Facilities allows users to locate mental health services in their city. Since substance abuse is often a comorbidity with mental health, users can select facilities that provide help with drug abuse as well.