SM: Let’s get specific. How did you identify the problem itself? Did you do market research of some sort?
JJ: My market research was to identify CIOs and CTOs. I came up with a list of people I knew or people who I knew second hand. There were about 45 CIOs, CEOs, and CTOs on my list. I bought them breakfasts and just asked them, “If I build this, will you buy it?” and had them rate their answer on a 1 to 10 scale with 1 being you absolutely would not buy, and 10 you will absolutely buy it. I did get concerned that they would just tell me what I wanted to hear. I did not know that there is a term for that phenomenon called survey bias, but I recognized the phenomena and knew I had to remove it.
In that regard, I gave them six different ideas. I did not give them a lot of context because I figured if it was important they would know. Email spam control was just one idea, information discovery was another one. I knew from my time at Interwoven that information discovery was a big problem. There was no good solution because all the search engines were no good. We partnered with 16 different technology search vendors to supply search technology for our repositories. We had five different search engines for five repositories, and we hated all of them.
We also knew that natural language processing did not work. I am partially responsible for wasting $150 million in 2000 to buy a company that generated metadata. We thought that metadata could solve the problem, so we bought MetaCode. We sold hundreds of copies, but nobody could get it to work. Content authors don’t really care about metadata. Metadata is only needed when you search or browse the site. There was a huge disconnect between the benefit and the effort.
I took my list of potential problems to all those various CIOs and CTOs and talked with them. I came back and tallied the results. Spam control got an average of 7 out of 10. The Information Discovery piece, to my surprise, got 10 out of 10. There was even one person who gave it 11 out of 10, and that was the COO of eBay, Maynard Webb.
SM: eBay’s search engine is the worst one anywhere!
JJ: He knew that, too. I went back to him and said, “You ranked this as an 11, you can’t be an innocent bystander. I need you on my board.” He is on the board now.
When I saw the results it hit me. I knew I had to look into that problem instead of a semi-good spam problem. What interested me was not only the problem, but that these folks also gave me hints of the solution. I asked those who I interviewed how they were finding information on a daily basis, and the answer was that they would call people they knew. Days or weeks later they would get an email link back with the content they were looking for.
I learned two things. First, a lot of the information was already out there, but it was in people’s heads. Second, to solve that I had to get into people’s heads, and email was not the answer. These were c-level executives.
SM: What kind of information are we talking about?
JJ: It could be anything. It could be product information, content, or support. A lot of the companies I talked to were B2B tech companies. There was a lot of information in documents.
SM: So it was an enterprise search problem?
JJ: It was a search, browsing, and navigation problem.
SM: And knowledge management. This problem has had many names.
JJ: I started to think that they were right. The search engine and navigation did not work. Every quarter, month or year there was a new agency and site design being worked. Why can’t we get it right the first time? There must be something fundamentally wrong. If people’s brains are making decisions about what is useful and what is not, perhaps the solution is not in the content.
All search engines believe that relevance is in the content itself and that by parsing the content, the grammar and natural language structure, we can make this relevant. That is just not the case. If I take my business card out it might have some relevance to our conversation right now, but if I just threw it on the street nobody would give a dime for it. Why is that? The content did not change.
SM: The context did.
JJ: Exactly.
This segment is part 4 in the series : Simulating The Brain: Baynote CEO Jack Jia
1 2 3 4 5 6 7