Ticketek’s parent TEG has embarked on a strategy to provide real-time data streams to venues, sports codes, teams and travel package bookers, opening new avenues to connect with customers.
TEG Chief Technology Officer Matt Cudworth calls it “innovation through integration”. The initiative is underpinned by an 18 month-old technology strategy to reinvigorate and open Ticketek’s “core system of record” up to external partners as well as for internal uses like data analytics.
Most of the enabling pieces are now in place: TEG has re-platformed its core system of record – previously a combination of core SQL databases and “very high speed, flat file data stores” – onto a MongoDB cluster hosted on AWS.
From that core, data is streamed in real-time to between 30 and 40 partners using AWS Kinesis and Lambda functions, or internally to a Redshift data warehouse instance for “enrichment” and analytics purposes.
TEG has previously talked up its burgeoning analytics business, which it likens to an “entertainment genome” of customer preferences that can help event partners personalise offers.
“Our technology strategy is basically to open up our platform and mindset to build out an open ecosystem that’s driven from innovation through integration … giving our customers and partners the ability to leverage our platform to reimagine the customer experience,” Cudworth said.
“Over the past couple of years we’ve built out a full API architecture, we’ve committed to real-time data streaming, and we’ve started opening that up to our venue partners like the MCG and to some of the sporting teams.
“That’s now starting to power their vision for customer experience for their particular audience.”
Like other web property owners, TEG is essentially dealing with the disaggregation of its audience.
“I think we’ve had a history where everyone came through the one website – a one-size-fits-all approach – [whereas] I think we will start to see a federated experience [develop] across a whole range of [partners] from the venues and sporting codes to people buying travel packages being able to drill into our inventory systems to find the right tickets to bundle up for an overseas visitor, for example,” he said.
“No one can convey the emotion of cricket better than Cricket Australia, same with the AFL, so we’re now supporting multiple mechanisms for interacting with the ticketing system.
“You may not even know you’re interacting with our system when you pull up your ticket. You might do it from an AFL app or venue app.
“That’s what customers demand. They want this frictionless experience wherever they are.”
Database overhaul
One of the first major projects TEG undertook was to re-platform the database cluster that acts as its system of record – that is, its ledger of orders and tickets.
Two years ago, the decision was made to shift off a combination of SQL and custom data stores into MongoDB, an open source NoSQL database platform. It began the migration six months later.
“For the last 18 months we’ve had our own built cluster of the standard community version of MongoDB hosted in Amazon,” Cudworth said.
TEG shifted its full data set into MongoDB, importing some 350 million tickets that were in the order system (Cudworth said it sells about 28 million tickets a year).
It has recently completed a slight architectural shift: rather than host MongoDB using its own AWS account, it now uses MongoDB Atlas, a database-as-a-service run by MongoDB that sits atop AWS, Azure or Google Cloud.
TEG stayed with AWS, meaning it has simply outsourced IaaS management responsibilities.
“I think for us it’s about us investing the time into the customer experience and not in the bowels of the infrastructure,” Cudworth said.
“We’re in the business of selling tickets and doing the best job we can for customers, not being world experts in how to shard the details of databases. There’s no way we’re going to be able to do better than the vendor [of the MongoDB software], as much as we could try.
“In addition, if you align to a standard you always get a benefit of shared robustness. Being different is always risky.”
Having a hosted service also makes it easier to bring acquisitions onto TEG’s systems.
“Everyone knows Ticketek from an A/NZ perspective, but we’re also delivering tickets in the US and UK and we recently bought the largest ticketing company in Malaysia,” Cudworth said.
“We’re going for a globalisation – or at least expansion – of our model.
“Atlas became an obvious choice for us to be able to deliver [services] in most regions. We can use different cloud providers, and you can’t go past [a managed service by] the vendor that makes the software. There’s no better expert in that technology.”
Ticket buyers likely interact with the MongoDB cluster multiple times for a single order.
“If you buy a ticket for your family there’s the order and then each one of those tickets is another potential row or ‘document’ in the MongoDB world,” Cudworth said.
“We’re also using MongoDB [to manage] a lot of our product information. It’s really our go-to system for all of our transactional data.”
Turn on, tune in, stream out
MongoDB is effectively a ledger of orders and tickets. As a result, most of the action on the data occurs elsewhere, either in the company’s own instance of AWS Redshift for analytics, or in systems run by event partners.
“We take the data and pump it through connector streams and [AWS] Lambda functions to enrich it,” Cudworth said.
“Effectively where we’re leading to is real-time analytics out the back that triggers a whole range of other events and things that occur behind it.”
Event partners have operational reasons to get real-time data feeds from TEG’s core systems.
“When you think of an event, from an operational and security perspective [you need] to understand the number of people in the stadium at any point in time,” Cudworth said.
“Every event now almost has real-time attendance data being sent off to operational centres and dashboards.”
The developer wishlist
There’s one other consideration for the architectural decisions made under TEG’s technology strategy: making sure that, to the greatest possible extent, developers get to work with technologies they want.
“I think there’s a power to be had when you align to what interests your developers,” Cudworth said.
“Developers are kingmakers in organisations, and being able to give them a set of technologies that excites them, is what they want to work with, and aligns to the company’s objectives, can be a very powerful thing.”
The choice of MongoDB isn’t too surprising; Cudworth notes MongoDB’s “dominance in the developer heartland”.
“I think most developers have interacted with MongoDB,” he said. “I’ve worked with it for many years.”
Selecting MongoDB for TEG’s core allowed Cudworth to tell existing and prospective developers they could “work on that technology at scale and with an unparalleled data set”.
It’s much the same philosophy with TEG’s embrace of AWS (“We’re very heavily invested in Amazon technologies,” Cudworth says).
Though TEG is using multiple AWS services, including basic IaaS, Redshift and Kinesis, it is perhaps its embrace of the serverless compute service Lambda that is paying the most dividends with developers.
“A big pivot point for us was when Amazon released .NET Core on Lambda,” Cudworth said.
“I’ve got a big investment in .NET and that instantly gave me a career path for all of my developers: this is where we’re going to go, it takes your skills to another level, it aligns to us as well because that’s what we want [as a business].
“We’re a very transactional system so serverless [computing] is very powerful in our business especially as we replicate a lot of the interactions that we’re doing.
“We stream data to 30-40 of our partners. Every event now almost has real-time attendance data being sent off to operational centres and dashboards, and serverless works perfectly for us in those environments.
“I think it’s easy to make technology decisions when you commit to certain paths and know it’s going to be right not only for the business but also your staff’s careers.”
(Courtesy IT News – www.itnews.com.au)
