Fail to prepare and prepare to fail: the cost of downtime during a big event

Fail to prepare and prepare to fail: the cost of downtime during a big event

World Cup, Superbowl, Grand National. Three of the world’s largest sporting events and all have fallen victim to unforeseen downtime within the last 12 months.

Just this February, William Hill’s sports betting app crashed in Nevada on Super Bowl Sunday. Last year’s World Cup Final took down Bet365 and Betano. And numerous major betting platforms broke down due to traffic spikes during the 2022 Grand National, making headline news.

So, why do even the biggest online sportsbooks struggle with large-scale annual events?

In this blog post we’ll explore why big events are challenging for betting platforms, the financial and reputational impact of unforeseen downtime, and measures to reduce disruption.

What causes downtime?

For all web-based platforms, from banking apps to ecommerce stores, hardware issues and server outages are an unfortunate inevitability. But for applications that receive traffic spikes around annual events, like streaming services and betting platforms, preventing platform outages can be even more challenging.

And a lot of that has to do with traffic spikes. The opening weekend of this year’s Indian Premier League saw streaming platform JioCinema experience skyrocketing simultaneous live streams resulting in reports of glitching, buffering, and quality issues. In short, volume exceeded server capacity.

Big sporting events like these have started to put immense pressure on streaming platforms. And now that online sports betting has grown in popularity, betting platforms are starting to face similar traffic spikes around game time.

Betting platforms rely heavily on databases to facilitate bets, calculate odds, and process financial transactions. So, when big events come around, the strains on databases mount.

“When people are going through and placing bets, every transaction ends up in a database, so the more load, the more work… and when the database is at capacity, then you’re going to see outages,” comments Nate Stewart, Chief Product Officer at Cockroach Labs.

Even if betting platforms are load tested (placed under simulated demand), it can be difficult to accurately predict the exceptionally large volume of traffic triggered by big sporting events. Potentially hundreds of thousands of individual users may flock to a betting platform, surpassing server capacity and causing system overload and server outages.  

The Grand National is a prime example of an event that continually causes platform downtime. The 2022 Grand National temporarily brought down Bet365, William Hill, and SkyBet after unprecedented traffic spikes exceeded server capacity, leaving users experiencing failed page loads, error messages, and platform outages.

The Super Bowl is another repeat offender. William Hill’s mobile sportsbook suffered a multi-day outage in Nevada during the second quarter of this year’s game. The previous year server outages and volume backlogs hit numerous online sportsbooks including BetMGM, DraftKings, FanDuel, BetRivers, and Barstool.

Fail to prepare and prepare to fail: the cost of downtime during a big event

The impact of downtime

Whether due to server outages, hardware failure, or traffic spikes, downtime causes two significant problems: loss of revenue and reputational damage.

Online sportbooks are some of the worst hit. William Hill’s outages during this year’s Super Bowl may have contributed to the reported decrease in Nevada’s Super Bowl handle which saw dollars wagered on the game down by $26.6 million (14.8%) compared to 2022. But (and perhaps more importantly) the disruption left users of the app angry and frustrated.

The reputational damage caused by downtime can be just as bad as, if not worse than, revenue loss. In 2022, the D.C. Lottery received almost $500,000 in compensation from the city’s official sports betting platform for lost revenue and reputational damage incurred after a technical mishap kept the app offline during the Super Bowl.

According to D.C Lottery’s Director, Frank Suarez, the payment compensated for lost bets, promotional spend, and included a sum of $428,000 to be used to “overcome the negative sentiment” incurred.

What’s the solution?

All applications have to contend with traffic spikes. But for betting platforms, maintaining 24/7 availability during peak concurrent usership is even more critical.

At the infrastructure level, three things often happen during peak traffic times: CPU becomes overloaded, servers run out of storage space, and/or the network interface fails to support sufficient network traffic. In short, server capacity is exceeded.

Optimizing the caching and indexing on a database increases application efficiency but this can only achieve so much. To cope with larger traffic spikes, it’s the scalability of a database server that matters most.

Databases can be scaled vertically (by increasing the processing power of a single server or server cluster) or horizontally (by sharing the load between additional storage nodes). The best scaling technique will depend on the type of database in operation. For instance, it’s easier to scale horizontally on non-relational databases than it is on relational databases since within relational databases it’s difficult to separate out related data across nodes.

Finally, it’s worth looking at your infrastructure stack. Implementing a multi-region architecture provides fault-tolerance for multi-region applications and, by keeping data close to end users, latencies can be kept to a minimum.

When demand is extremely unpredictable, a hybrid infrastructure approach, with a base layer of bare-metal and a top layer of cloud, can be particularly beneficial. A platform can use bare metal for its predictable infrastructure usage and mobilize cloud auto scaling during demand spikes.

Prior planning prevents poor performance

With sports betting platforms surging in popularity, traffic spikes are an ever-more-present challenge that even the most established online sportsbooks struggle to cope with. Even though demand surges are unavoidable around big events, the detrimental impact of failing to accommodate traffic spikes can be huge.

To avoid the financial and reputational hit that downtime causes, online sportsbooks and betting platforms must plan ahead. Database optimization is the bare minimum but to cope with larger traffic spikes, being able to rapidly increase server capacity is what’s needed.

Adopting a hybrid infrastructure approach combining the reliability and predictable billing of dedicated servers with the instant auto scaling offered by cloud offers the best of both worlds. A stable infrastructure foundation for the ‘normal’ days, with a scalability buffer for when that match, game, or race causes its annual frenzy.


Related articles