An exclusive with William Tunstall-Pedoe, CEO and Founder of TrueKnowledge.com.
What is TrueKnowledge.com, and what is the business/ revenue model behind it?
TrueKnowledge.com (www.trueknowledge.com) has a technology which automatically answers people’s questions on any topic.
We provide users with exactly what they are looking for, instead of a list of links to web pages which might be what they looking for.
Behind the technology is a large and continuously growing knowledge base of the world’s knowledge, which is in a form that the system can understand and reason with. The Knowledge Base contains both factual and common sense knowledge, which enables the system to model the world and understand what users are asking. As the knowledge base grows the platform can answer more and more.
I invented the technology and founded the business which has subsequently been taken forward and transformed by the team of 25 fantastic people who work here.
As far as revenue is concerned, the www.trueknowledge.com website has about seven million monthly users and is growing at around 20% per month. We monetise that traffic with advertising.
Additionally we generate revenue by licensing the underlying platform to a number of third-parties via APIs to power their products. These include other websites and mobile products such as Siri (www.siri.com)
We are also working on additional products powered by our technology.
Which company do you consider to be your biggest competitor, and what sets True Knowledge apart from other search engines and Q+A platforms?
We have competitors in a number of different spaces.
In the Q&A space are many companies like Quora (www.quora.com) and Answers (wiki.answers.com) where the answers to questions are written by people and no attempt is made to answer the question automatically. This approach doesn’t typically give users an instant response to their question and has some serious challenges around scaling and keeping results fresh.
Wolfram Alpha (www.wolframalpha.com) has some similarities to us in that it answers questions automatically and uses structured knowledge to do so, but there are fundamental differences both in approach and technology.
What did you do, professionally, before establishing True Knowledge?
My occupation has always been developing and commercially exploiting innovative computer software. I have been doing this continuously since the age of thirteen (breaking only to do a computer science degree at Cambridge University). I’ve always had a passion for both business and technology and have been very fortunate in being able to combine the two without ever having to do anything else.
Was it a sequence of events or one specific “Aha” moment that lead you to want to launch the platform?
Many of the products I developed before True Knowledge involved very deep technology typically involving Artificial Intelligence (AI) but in very niche markets. For example, I developed the anagram software that Dan Brown used to create the anagrams in the Da Vinci Code book and movie (www.anagramgenius.com) and I also developed the only software that can automatically solve and explain cryptic crossword puzzles (www.crosswordmaestro.com). I’m proud of the technology behind these products and they continue to sell to this day, but they are never going to result in an industry changing business.
In the late 90’s I thus decided that I wanted to tackle a difficult AI problem with a very big market and chose the biggest problem I could find which was to improve internet search. The internet was becoming very big and internet search very important but I couldn’t believe that 10 or 20 years in the future people were still going to be finding things out by guessing keywords and browsing links.
Having made the decision though it was not a fast process to get going. I needed almost two years to get the early technology developed and patents filed and then the dotcom crash put pay to turning it into a business for a while. However, in 2006 I filled in the remaining technology and business gaps, got some initial funding and hired the company’s first staff.
Once you’d made the decision to obtain funding, how easy was it for you to raise?
I raised friends and family money and a £75k government R&D grant in 2006 at the same time as hiring the company’s first staff. This lead to an angel round in 2007 and then venture finance (from Octopus Ventures). Raising equity finance is never easy but we have been fortunate to have found enthusiastic and supportive investors each time we have needed to raise money.
Does being based in Cambridge work well for you and the team? Do you feel that you miss out on anything by not being in London?
I have jokingly referred to Cambridge as “the second best place in the second best country” for an internet business like True Knowledge. Although “second best” is usually pejorative it is actually not bad considering the size of the world. Cambridge is a fine place for a technology business with a great technology and start-up culture, the world’s best university and great people.
London has some big advantages though in terms of sheer volume of talent within commuting distance and True Knowledge has a London office as well for this reason. Also our Cambridge office is located a few minutes from the train station enabling access to London in under an hour. The United States, and in particular the San Francisco Bay Area (first best country, first best place) have significant advantages again – particularly for talent/know-how and access to finance for an internet technology business like ours.
What is the single best thing that you’ve done, and the biggest mistake that you’ve made, as CEO?
The single best thing I did was start the business in the first place. When I started, the idea of founding a business on a new search technology, arguably in competition with highly successful companies like Google and tackling a very difficult AI problem that nobody had solved for decades has (with some justification) been characterised as nuts. It would have been very easy to have listened to the sceptics and never let it get off the ground. However, the team here has proven the cynics wrong with a system that is now performing at scale and serving millions of users, being used by paying customers and with lots of other validation of our approach. We still have lots more to do but I have never regretted the path we set out on.
I have made many mistakes and possibly the biggest was not focussing early enough on showing what we can do with a highly slick product that engages users and looks perfect from a design point of view. Our early strength was always the core technology and the focus was on showing what the technology could do. We presented the technology in a fairly raw form hoping that others would see its capabilities through the imperfect presentation and use it in their own products. With hindsight, bringing product and design people into the team earlier could have brought some benefit. We’re in the process of fixing this now.
What would you expect to be the greatest challenge to your business over the course of 2011?
Possibly the biggest will be growing the business to massive size, whilst dealing cleanly with all the issues that result, in a way that doesn’t slow us down.
You’re in the process of building a great team there: is there anything in particular that you look for in new team members?
We are a business that is very ambitious: we are trying to build something enormous and change the way world interacts with the internet. What we are doing is also extremely difficult. We are thus looking for highly talented and intelligent people who are excited by this challenge and unafraid to be involved in something big.
Finally, before University you briefly worked on a project helping computers to recognise and understand human speech – is this something that we can expect True Knowledge to be able (one day) to do?
Yes – but not (at least in the short term) with our technology. Speech recognition technology turns speech into text. Our technology responds to text so getting the user experience of talking to a computer and having it understand and respond can be readily done right now by combining our technology with state-of-the-art speech recognition from third parties.
Thanks William – best of luck for an exciting year ahead.