Central to this is Ticketek’s new system of record, MongoDB. I got the chance to speak to Ticketek CTO, Matt Cudworth, at MongoDB World in New York this week about why relational databases couldn’t cut it anymore and how it is using the NoSQL alternative to support its data-driven activities. Cudworth said:
It’s a really interesting story about how we ended up with MongoDB. But we are a data driven business and it was really about freeing up the data, getting the data moving. That’s the key thing that we’ve done over the last couple of years.
Our transformation started about three or four years ago, and it had all the things that you’d normally see - the removal of the legacy environments, the uplift of the technology, move to the cloud. But I think particularly in the last couple of years we’ve benefited from that MongoDB environment, it’s really freed up the data and now we’re moving at pace.
The biggest part of our business is now real-time data streaming. As part of that architecture we are moving data to our partners in real-time. We are delivering to the venues, the hirers, it’s a real ecosystem of data. I think we make a strong commitment to banish batch processing from the industry.
The decision to move to MongoDB came down to one particular scenario, which Ticketek internally calls a ‘hot show’ - when a major event, such as a concert, or the recent Commonwealth Games, goes on sale. Those ‘hot shows’ see more demand than supply, Ticketek knows they are coming, and most of the action happens in the first 30 to 90 seconds of tickets going live. A database needs to be able to handle that intense load and move it through the system, with data moving at scale and at pace.
Cudworth held a beauty pageant of databases to support this scenario, and MongoDB came out as the clear winner for out of the box functionality and scalability. He said:
When we looked for a new system of record, we looked at Oracle, Microsoft SQL, MySQL, MongoDB, Couchbase. It was a lineup, anything was on the table. It’s not that we were particularly aligned to NoSQL versus SQL, we came from an SQL-centric sort of world.
We said that we have to have something that we are comfortable that can handle this workload. That’s one of the key moments of time in our business. And MongoDB was the only one that could handle the load that we put it under, out of the box.
You could argue that maybe Oracle could do it, but it would need customisation, they’d need to modify it to achieve that. So out of the box, MongoDB was the only one that came close to handling it.
However, MongoDB and a NoSQL environment has also driven simplification for Ticketek, compared to its previous SQL infrastructure. Cudworth said:
In the modernisation we wanted simplification. If you think about the SQL environment, all those stored procedures, all that business logic that’s embedded down in my key data stores, that’s the piece I was trying to get rid of.
I was trying to free up not only the data, but the business logic around it. That’s going to help us scale and keep things simple going forward. Mongo has effectively become our core system of record. It’s the order’s database. It’s kind of a perfect use case if you think about it, an order is a document.
Creating a platform
Cudworth also explained that Ticketek is using this data-driven architecture to create a platform that is being exposed to external partners for their own use and to build upon, with the aim of creating a better customer experience. He added that the ticketing business has probably changed more in the past couple of years than it has in the last 30 and Ticketek is focused on building an environment exposes its platform for scale and use beyond its four walls. Cudworth said:
Rather than thinking about it as an application, think about it as a platform. What we’ve done is in essence is open up the platform. It wasn’t just opening up technically, it’s opening up our mindset to build a whole ecosystem around the platform.
The key tag line that we use is ‘innovation through integration’. Because of that deep knitting of the venues and hirers, we’re more like a marketplace than we are and e-commerce company. We are selling other people’s inventory, we are selling our own inventory, it’s back and forth. There’s 20,000 events a year.
Cudworth also isn’t keen on the term ‘digital transformation’ as he think it sounds like an inward looking term, a project that has a start point and an endpoint. He prefers to use the term ‘digital acceleration’. He explained:
It’s no longer a project, it’s no longer a destination, it’s just speeding up every part of our business. We’re opening up the platform. We took those internal APIs and put them in front of our customers, told them we will give them every part of our platform, real time data streams. And we are starting to see all of these new customer experiences pop up. No longer it’s just the one website, there’s a whole set of new applications. We are now feeding their sites.”
Cudworth said that Ticketek could probably charge for this access, but it doesn’t need to, because it’s more about building this ecosystem. He said:
We’ve pretty much doubled the size of the business in the last two to three years, a lot of it has been organic growth. We are expanding up into Asia, that’s our stated goal. We bought a ticketing company in Malaysia, that’s running our platform now as well. We’ve got a joint venture in China. We want to extend our successful model up into Asia.
Key to this expansion has been Atlas, MongoDB’s database-as-a-service, which runs on Google Cloud, AWS and Microsoft Azure. Atlas allows Ticketek to be able to port its platform globally with ease and has also introduced an enhanced layer of security into the system. He advised other customers thinking about going to MongoDB to start with Atlas, not to delay it. Cudworth said:
We are an Atlas customer with MongoDB, and as part of that we were starting to deploy our platform tactfully. We went into the UK with a family entertainment event. We took the opportunity to deploy our platform to sell some of the tickets. We took a lot of those learnings back and said, how would we shrink the footprint of our platform and enable it to be far more agile to move into these regions?
One of the best things we did was move to Atlas. We were a successful MongoDB client, and it was fine, we had a cluster. What we wanted out of it was that portability. Do I really want to be managing it up in Asia and the UK? But what we gained out of moving to Atlas was quite surprising.
It actually washed a nice layer of governance through the business. We got encryption at rest, we didn’t ask for it, we just got it. That’s a bonus. Those are the things, if we had just started on that, it would have been better. The costs aren’t that different. The big cluster we were managing on MongoDB versus Atlas, the ROI is actually pretty good. We have mandated that.