Bright Ideas: Predicting the News Can Be Profitable

Hubdub’s founder, Nigel Eccles, says his news prediction website appeals to news addicts and amateur punters alike. This week it launches its UK “news predictions” site, having spent the last nine months building up a US version that now attracts 200,000 users a month. More…

Popularity: 5% [?]

We’re delighted to be able to offer our UK Hubdubbers their very own UK Edition of Hubdub. Now Hubdubbers based in the UK can now see news stories that other UK users are predicting on, simply by selecting the UK Edition at the top right of the home page.

Interestingly, the US site’s most popular stories today are all around the Presidential Inauguration and Obama’s cabinent appointments, while the UK site’s top listings are X-Factor, Strictly Come Dancing and Everton v Middlesborough. What does this say about the UK, I wonder? 

Here is the press release that we issued today in the UK:

Edinburgh, 10 November 2008: News junkies in the UK can now get an additional fix, by predicting the outcome of news stories – engaging in news stories instead of just reading them.

Users of Hubdub, which launches in the UK today, use a pot of virtual money to place bets on the outcome of any of over 2000 running news stories, covering the likes of politics, sport, business and entertainment.

Since launching earlier this year in the US, Edinburgh-based Hubdub has garnered a substantial and loyal following, with over 200,000 unique visitors per month, 800,000 predictions to date and over 2 million Hubdub dollars staked every day. Some users are spending several hours a day on the site, with many of them confessing to a definite Hubdub “addiction” on the site’s forum.

The fact that it’s Hubdub dollars at stake is apparently irrelevant – Hubdub users are so competitive that the most important thing is that they prove they were right to the rest of the vibrant community. How much is won or lost depends on not only how the news story turns out, but how others have predicted the outcome. Groups of friends can set up their own, dedicated leaderboard to show who has predicted most accurately.

Globally, prediction markets, the mechanism behind Hubdub, are becoming increasingly popular. In the US, the Presidential election sent the numbers of users soaring. They have proven to be more accurate than polls, and because they respond in near real-time to developing stories, more agile too. Indeed, Hubdub successfully predicted the final voting results of the US election where the majority of polls and pundits failed.

Lesley Eccles, one of the founders of Hubdub, commented: “It’s great to be launching in the UK, despite the recession. People are going to be looking for free entertainment and will turn to the web to find it. Hubdub is free, exciting and keeps you informed. In addition, news sites are looking to engage and interact further with their readers and Hubdub gives them a fantastic means to do so.”

Having secured partnerships with leading media news sites including Reuters and The Independent, Hubdub is taking prediction markets more mainstream and making them easier for casual browsers to appreciate. Sitting beside high quality news content, Hubdub provides news sites with additional ways to engage their readers, keeping them online longer and returning on a regular basis.

Current UK Markets include:
• Where will Russell Brand next get a full-time job?
• When will the next breach of confidential data occur?
• Who will win Strictly Come Dancing?
• All the games, goals and scorers in the English and Scottish football season

Popularity: 5% [?]

Over the past couple of weeks I have been speaking to lots of members of the Hubdub community. One thing that has come into the conversation a number of times is what is the vision for Hubdub? Where do we see ourselves in 5 years time? While history loves stories about the flash of inspiration or the single event that sparked an idea, in general I believe the process of innovation is much more messy than that. Hubdub was no different. Part of the idea goes back to my university days when I developed a fascination for game theory and gambling (pure and applied!). Then in 2000 I joined a start-up called flutter.com. Flutter’s original vision was to enable people to make person-to-person bets on anything that interested them. There were a number of problems with that vision, which I blogged about here, but we quickly changed course and were ultimately successfully acquired by Betfair. For those of you who aren’t familiar with Betfair, it is a betting exchange, like an eBay for sports betting.

While Betfair has been highly successful, there were two parts of their service that I was always surprised that they never developed in 7 years since I left. The first part was that they have one of the most active forums on the internet with people discussing everything from sports to celebrities to everything in between. However the forums have always been kept at arms length from the main betting exchange. The second part is that behind all these markets is an amazing set of data that clearly shows how news events like the US elections have evolved. Over the past 7 years I have used Betfair prices to follow news stories like that (and also who is next to get kicked out of Big Brother!). Unfortunately, unless you can read digital odds and you can navigate their interface you won’t be able to get the same benefit.

The second last element behind the Hubdub concept was an exceptionally bad stock portfolio I had from 2005 to 2007. I had spent weeks and months of my free time working out PE ratios, momentum, free cash flow, debt ratios etc. I even bought a couple of Domino pizzas to check if their product was up to scratch (the pizzas were good but I passed on the stock). However, despite all my work my portfolio starting dropping from day one. Owning Sportingbet shares when the US passed the Unlawful Internet Gaming Act didn’t help much either. However the thing that I discovered was (a) how frequently I checked my portfolio (er, every other minute when things were quiet) and (b) that all the information I wanted to know was boiled into the price. Only if was there a significant price movement did I dig into see if there was a significant news event. This made me realize that if I could place bets on future events and news stories I was interested in then I could let the system ping me when something significant had happened (however I would probably still want to check my portfolio every 15 minutes!)

The last thing that pushed me over the edge to starting Hubdub was working in the news industry I had found that newspapers on-line had been really successful at reaching out to millions of readers. However while they could reach millions of readers those readers weren’t hanging around for very long. For example the Guardian is one of the biggest UK on-line newspaper sites with around 20 million monthly unique visitors. However each of those visitors only views about 11 pages on average per month. That made me realize that if we could build something which combined the reach of news with the engagement and excitement of a betting exchange then that would be really something. Thus Hubdub the idea was born on July 23rd, 2007, exactly one year today.

*A lot* has happened in that past year with the help of a fantastic team who have taken an idea and made it their own. In many ways Hubdub looks nothing like the mock-ups we threw together last summer. Some things we tried never really panned out the way we expected (did anyone ever vote a news story up or down?!) but the great thing is that the idea core to the concept, that people would find the idea of trading predictions around news stories and future events highly addictive. Even more exciting is that many of our original ideas we still haven’t had the chance to try out.

All that is just the pre-amble to let me put off trying to describe what my vision is for Hubdub in 5 years time. The Hubdub concept has evolved over the past 12 months as it has gone from one person owning it to thousands of people owning it. That makes it tough to lay out a vision so here is my best shot:

In 5 years time, Hubdub is the market platform for any and all future events of public interest; where millions of users actively trade tens of thousands of stories and future events. Streaming out of those public markets is data on:
- Forecast probabilities of the likelihood of future events
- Peoples’ individual views of the future, which they share and discuss with friends
- Peoples’ reputations for accurately analyzing and forecasting future events
- A filter of news stories that have a significant impact on peoples’ outlook on the future (that is, only newsworthy stories)
Hubdub would be the exchange for this activity with both predictions flowing in and data flowing out via partner sites, distributed functionality and third party apps. Quoting Hubdub forecasts in news stories about future events will be as common as quoting stock prices in financial stories is today or (in the UK) quoting betting odds for political elections.

Tim O’Reilly recently wrote about the success of web 2.0 applications but asked question “But what good is collective intelligence if it doesn’t make us smarter?”. I truly believe that achieving the above goal would go a long way in answering that question. I would love to know your thoughts.

Popularity: 9% [?]