“We decided early on to handle our production operations entirely in-house with a small team of crack engineers,” said Renaud Visage, CTO and co-founder of event ticketing giant Eventbrite.
Investors in Eventbrite include Sequoia Capital, Tiger Global and T Rowe Price.
‘The smart use of machine learning and real-time decision engines based on a variety of signals will make ticket scalping a non-issue in the next few years’
– RENAUD VISAGE
Earlier this year, the company announced it was to create 50 new jobs in Cork over the next three years at a new customer support centre in the city. The company also has an office in Dublin.
In cooperation with Eventbrite’s international network of customer support centres in San Francisco, Nashville and Mendoza, the new presence in Cork will serve event organisers around the world with 24/7 support.
What are the main points of your company’s IT strategy?
While we launched Eventbrite on hosted servers, we migrated to cloud-based solutions in 2009 and have never looked back. As a start-up born in the mid-2000s at the very early stages of ‘the cloud’, it just made a lot of economic sense for us to switch to cloud providers for many of our hosting and SaaS solutions.
We didn’t want to spend any time racking and upgrading servers, hiring network security specialists, procuring and configuring hardware etc. We wanted to concentrate our efforts on building the best ticketing solution in the world.
Traditional data centre solutions would have been very capital intensive and not flexible enough for our needs. Solutions like Amazon Web Services (AWS) allowed us to use as much as we needed without having to commit. We can grow and contract as needed which is very pertinent for our event ticketing business, which has a lot of variability in demand, as large, fast-selling events can go on sale at any time.
Can you give a snapshot of how extensive your IT infrastructure is?
We currently handle 2m tickets every week, and that is reflected in our infrastructure. We have a large AWS presence in the US East (Northern Virginia) region and a warm disaster recovery site in the US West Region (Oregon). In Virginia, we make use of all the ‘availability zones’ which roughly translate to data centres to buffer us from potential single data centre outages.
Our systems are designed to be highly redundant and horizontally scalable.
The ticketing industry has some unique usage patterns which cause sudden, massive spikes in traffic and transactions, so you need to be provisioned to handle these at any time. People talk about the elastic cloud and using techniques like auto-scaling, but once auto-scaling kicks in and the new servers are ready to take the traffic, the spike in traffic is typically gone. We need all of our systems to be available 24/7 to handle these spikes. We also heavily rely on content delivery networks to serve our static assets globally for improved performance.
Do you have a large in-house IT team, or do you look to strategically outsource where possible?
As a global online business, reliability, uptime, security and system performance are critical for us. We decided early on to handle our production operations entirely in-house with a small team of crack engineers. I think we have some of the best in the industry and we see this as a critical competitive advantage. I like to think that we have the best infrastructure in the events industry and can ensure smooth operations for our customers.
What metrics or measurement tools do you use to gauge how well IT is performing?
We strive to be fully data driven to run our business so we monitor and automate everything. We measure uptime, response times, peak page views and transactions per second, error rates, system failures, and a number of critical alerts at all times. We’ve built automated fail-safe mechanisms to handle unexpected hardware and networking problems, so that our systems can remain online even if parts of our infrastructure are offline.
What are some of the main responsibilities of your own role, and how much of it is spent on deep technical issues compared to the management and business side?
As the CTO of the company, my role is to ensure that our infrastructure evolves to handle the growth of the business and that we fully leverage new technologies to move fast and continually improve performance. I also need to keep a watchful eye on cost and complexity. Running our IT operations in an economically viable way is at the top of my current priority list.
How complex is the infrastructure, are you taking steps to simplify it?
Eventbrite has made a conscious choice to limit its use of the services provided by AWS. We essentially use EC2 and S3 (plus storage and some other minor services), but we do not use the heavyweight services like RDS, Kinesis, and others.
These services are thin wrappers on top of other freely available software like MySQL and Kafka. We would prefer to operate these on our own and understand how they work, how they perform, and when and why they fail. It keeps us loosely coupled with AWS in case we ever need to move to a different cloud provider in the future, or decide to build our own infrastructure.
What are the big trends and challenges in your sector, and how do you plan to use IT to address them?
The event ticketing industry has been plagued with a variety of problems that we’re aiming to solve long-term by finding the most efficient way to solve them via technology. We started our business by democratising ticketing and allowing anyone anywhere to create an online presence for their events and start selling tickets in just a few minutes. We also solved the scalability issues that large, fast-selling events still encounter way more often than they should. Our systems can handle the load that the most popular events generate.
Another focus for us in the near future is preventing bots and scalpers from buying too much of an event’s inventory. Millions of euros are lost every year to middlemen who bring zero value to the industry and move inventory to secondary markets, forgoing legitimate revenue for artists and promoters.
The smart use of machine learning and real-time decision engines based on a variety of signals will make ticket scalping a non-issue in the next few years. We aim to become the technology leader in that space.
Buy your tickets now!