London.AI is a vibrant community of AI practitioners hailing from large technology companies, startups and academic institutions. Alex Flamant and I have run London.AI since 2015 to accelerate the science and application of AI to solve impactful real world problems.
While London is in our name, many of the businesses, entrepreneurs and Universities in our community are based across the UK, the Nordics, France, Germany, the US and Israel. This diversity puts us in a great position to sample trends in the AI space that help us understand its health.
Last night, we held our 11th London.AI event to kick off 2018, in collaboration with the European Xoogler community and supported by Google Cloud. The jam-packed room at Google London prompted us to analyse our attendee data over the last 12 months to see how we’re doing!
London.AI is >1,812 people strong in 2017 alone!
Over the last 12 months, we have run 5 events (London.AI #7 through #11). On average, we collected 478 registrations per event over 4 weeks leading up to the day (min=312, max=630). In total, we have received 1,812 unique requests to attend our 5 London.AI events! These individuals represent 54 different academic institutions, 301 large technology companies, 519 startups and 139 investment firms. You can see the trends in representation of these groups in the graph below.
>1/3 of our constituency are startups
Our events regularly feature early-stage technology companies that are developing and/or applying machine learning technology to solve real-world problems. Examples included Improbable (large scale simulations), SwiftKey (predictive mobile keyboard, since acquired by Microsoft), Tractable (auto insurance), and Magic Pony Technology (video processing, since acquired by Twitter).
As such, it’s not surprising that Startups account for 38% of our attendee base, on average. These individuals commonly identify as software engineers, machine learning researchers/engineers, data scientists, designers and CEO/CTO/VP Engineering leaders. Next, we see 22% from Corporates (mostly large technology companies), 18% from Academia, 12% are investors and 10% are a mysterious Other :-) The graph below shows how this breaks down per meetup.
As the future engines of our technology industry, we place a particularly strong emphasis on welcoming as many students to London.AI. We see consistent attendance by students from institutions such as University College London (specifically their CSML program) and Imperial College London. Further to our goal of being a hub for the AI space in the UK, we’re happy to see students from Harvard, Columbia and Tel Aviv attend too when they happen to be in town.
Referrals from friends drives up to 75% of interest for London.AI
Over the last year, we’ve experimented with various online and offline distribution channels to reach people whom we think have an interest in London.AI. By far the most successful acquisition channel is in fact word of mouth from someone who attended or heard of our events before. While we’ve got some new ideas for how to evolve London.AI this year, we think this metric lends support to our current product-market fit :-)
Room to grow! 2/3 of attendees are experiencing their first London.AI
On average, we see 1/3 of our registrations from repeat London.AI attendees and 2/3 from people attending for the first time. We feel this strikes a good balance between catering to a core group of loyal fans and opening our doors to make brand new connections. It also tells us that the community of interested people runs far, far deeper! As such, we’ll make a bigger push this year to grow our reach. The graph below shows you how this metric has evolved.
How to reduce attendance churn by >50% and support a great cause
Our very first London.AI event was free-to-attend, allowing anyone to register and show up. We weren’t sure how many attendees would actually turn up and inevitably, quite a few people churned. This wasn’t ideal for figuring out our logistics and making sure everyone who wanted to attend could do so. Ever since London.AI #2, we put in place a strategy that hit two birds with one stone:
- Anyone interested in coming along to London.AI is asked to complete a short registration form on our website.
- We send tickets to 150–200 registered individuals who are demonstrably involved in AI research, engineering, product and company building.
- Each ticket costs £5 to purchase, with 100% of proceeds going directly to Code Club, a nationwide network of volunteers and educators who run free coding clubs for students aged 9–13.
Two things happened as a result: 1) Our attendance churn rate dropped by >50% almost immediately, and 2) each event raises enough to support two Code Club classes for an entire semester!