Why building Web Summit is as much maths as marketing
Published 17/09/2015 | 02:30
'Wait, let me get this straight: You want me to switch working in a particle accelerator and come work for a conference company in Ireland. I'm a physicist not an event manager, that makes no sense." That physicist now works for us.
I started Web Summit five years ago. It came from my frustration as a startup founder attending conferences. I felt there was a different way to bring people together. That difference would use lots of software we'd build and then lots more graph data we'd analyse. It's an approach that's helped propel Web Summit from a tiny conference to the largest startup-cum-tech gathering on the planet. Web Summit has become a crossroads for the world's largest companies and the world's newest companies.
We've now over 100 people full time in Dublin, including some incredible physicists and other rare creatures who've left their jobs in universities, traditional software companies and research institutes around the world to join us. Including that physicist.
But it takes more than just instinct and energy. It takes algorithms.
It might sound odd, but Web Summit is run as a giant algorithmically optimised dating festival.
In our view, as conferences scale, they become giant omnishambles. What use is a searchable or alphabetical attendee list when there are 3,000 other attendees, never mind 30,000? But it turns out you can go a long way to solving that problem, or hack conferences, if you just think about them in terms of momentary social graphs.
Many years ago, LinkedIn, Facebook and Twitter cut through the problem of prioritising from thousands of possible connections, friends or followers to suggest to you by using 'recommender' systems. LinkedIn suggests similar profiles. Facebook suggests friends. Twitter suggests followers.
They do this by analysing the network data inherent to their social graphs and personalising recommendations to you. Ultimately what they do online, we do offline. The maths they use is neither proprietary, nor overly complex.
Many people look at what we do and wonder how could a bunch of outsiders from Ireland grow conferences around the world faster than anyone else. The answer is pretty straightforward: a field in maths known as graph theory.
Let's be very clear: Web Summit is not some sort of magic elixir for attendees. Attendance does not guarantee entrance into the pantheon of startup or business gods. The chance of creating the next Uber and raising a funding round in a Dublin pub, is probably similar to winning the lottery.
To complicate matters further, over the years I've watched visionary entrepreneurs, who've already built billion dollar businesses, fail the second time round, despite an incredible team and a mesmerising roster of investors and advisors. On the other hand, I've watched twenty-somethings who haven't a sense for design or code in their bodies build billion-dollar startups, first time round.
So where does Web Summit come in?
Let's start with Y Combinator. Y Combinator is incredible in my view. It's produced a staggering number of unicorns over its short history. Being part of Y Combinator increases your chance of winning the startup lottery by exposing you to incredible people and ideas. But it doesn't guarantee you'll win. After all, Y Combinator has produced more failed companies than unicorns, along with many middling successes.
Outcomes at Y Combinator are not binary, they are distributed. In other words, outcomes are neither universally good, nor are they universally bad for every startup in every Y Combinator class.
If you were to look at the shape of the distribution curve for success versus failure for Y Combinator startups versus startups in general, you could assume that being part of Y Combinator is on aggregate positively correlated with winning the startup lottery. In other words, it almost certainly increases your chances.
Web Summit is a different animal entirely to Y Combinator. We attempt to compress exposure to relevant ideas and relevant people into 3 days. Internally we think of Web Summit as an accelerated accelerator, a collision of 30,000 people from almost every country and industry on the planet.
We accelerate the probability of meeting people relevant to you and your business by applying graph and network analysis to engineer serendipity at the scale of 30,000 people over three days. In other words, we use maths to try and recommend who you should meet out the 30,000 people in the room.
Think of it this way: that pub crawl you find yourself on in Dublin isn't a random collection of execs, entrepreneurs and investors, it's machine curated.
But, and this is important, similar to any product, experience is distributed at Web Summit, similar to Y Combinator.
Does every attendee achieve incredible outcomes by attending Web Summit?
No. The probability of incredible outcomes for attendees at Web Summit is distributed. Yes, it's true that startups have emerged from Web Summit with tens of millions in funding because they met a single investor in a pub. But it's also true that some startups have met tens of investors in many pubs, and walked away with nothing.
Not every attendee will have a life-changing experience at Web Summit. Some will, some won’t. The probability however that you will have an above average experience at Web Summit as compared to other events is, I believe, higher. And it’s that marginally better experience on aggregate that has driven our growth.
Your challenge as a Fortune 500 CEO, early stage startup, seasoned investor or the many other types of attendees that come to Web Summit is deciding whether we’re building something that on average delivers value. We’d argue that statistically you’re increasing the probability of success.