About The Client: AO
AO.com is a multi-award winning eCommerce retailer that offers in excess of 4,000 products across its platform. The company originated in 2000 and has rapidly grown to become a household name.
“It made absolute sense to us to deploy the TrafficDefender Virtual Waiting Room to ensure that we could continue to serve more customers simultaneously than we ever had, even if the event turned out to be much bigger than we expected”Adam Warne, Group IT Director
In 2014 AO.com embarked on a Black Friday campaign to maximise sales revenues prior to Christmas. The marketing investments they made performed exceptionally well – too well.
The surge in website visitors caused by the promotional campaign pushed their IT systems and infrastructure to breaking point. Although the AO.com website maintained operability during the sale, without any degradation of load speed, the experience highlighted a vulnerability that their team understood needed to be addressed.
In 2015, Group IT Director, Adam Warne, and Head of Software Development, Carl Phillips, commissioned TrafficDefender to deploy its Virtual Waiting Room solution in preparation for that year’s peak event.
Following the experience of 2014, AO.com started preparing for Black Friday and Cyber Monday much earlier in the year. Adam and Carl’s strategy was to ensure all aspects of their web platform were scalable so they could cope with unprecedented levels of visitors.
The strategy required a methodical and iterative approach to measuring the maximum capacity of the platform, identifying what the limiting factor of that capacity was. Once the limiting factor had been identified AO.com could either look to squeeze more out of the platform through optimisation or increase throughput by scaling up or out.
Although AO.com have a strong team of software and infrastructure engineers, they realised they required specialist support. TrafficDefender were initially brought in to provide performance testing on their web platform. The objective was to help AO.com load test their platform to capacity and identify constraints and limitations.
The platform was re-engineered to enable it to auto-scale in line with increases in website traffic. However, with peak events such as Black Friday, Cyber Monday and Christmas sales, the sudden and unpredictable nature of the traffic spike can catch the best websites off guard. This is because most websites cannot auto-scale fast enough to cope with the increased load.
Adam Warne, Group IT Director, commented: “TrafficDefender told us about their waiting system, which comprises visitor and bot control capabilities, and whilst we absolutely wanted to avoid having people queue to buy from us, we recognised that we could not rely on the accuracy of our sales forecasting to predict the scale of the traffic spike. It, therefore, made absolute sense to us to deploy the Virtual Waiting Room to ensure that we could continue to serve more customers simultaneously than we ever had, even if the event turned out to be much bigger than we expected”.
In addition to Visitor Control, AO.com utilised the power of Netacea by implementing its Bot Detection capability. As with most eCommerce companies, AO.com did not fully understand the composition of their website traffic or what the traffic was doing on their web platform. Initially, it was activated in ‘observation mode’ to help AO.com understand and segment their website traffic between humans, good bots, bad bots and grey bots. The grey bots were given further analysis to classify them into either good or bad pools. Once the composition of the web traffic was fully categorised Bot Detection could be activated in ’protection mode’.
Bot Detection’s ‘protection mode’ blocks all known bad bots from the AO.com web platform, which enables them to either reduce their hosting costs as a result of reducing their overall traffic volumes or increase conversion rates and revenues by replacing bad bots with more human visitors
- Business and customer base growing faster than IT systems can support.
- Huge spikes in traffic pushed infrastructure to its limit.
- Automated bot traffic consuming resource and posing a security risk.
For AO.com, Black Friday and Cyber Monday 2015 was an unprecedented success, generating revenues above and beyond previous sales events. The eCommerce platform stayed open for the duration of the weekend and customers enjoyed fast load speeds and positive user experiences.
TrafficDefender’s bot control functionality has also helped them block content and price scrapers and, guard against brute force attacks and potential malicious competitors. By reducing bot access to their website platform they have also increased the capacity of their system, enabling them to handle even more human visitors looking to purchase their products.
Whilst the strategy around managing peak events remains the same for AO.com, they are always looking for ways to improve the performance of their platform and cope with even bigger traffic volumes and spikes.
- Can serve more customers than ever had before
- Business can push more flash sales events with confidence
- Malicious bots blocked from accessing the website
About The TrafficDefender Solution
VIRTUAL WAITING ROOM
When the maximum number of visitors is reached, TrafficDefender delivers a waiting page to all additional visitors without any access to your servers ensuring protection for your servers and quality of service to your site visitors.
By putting all additional visitors into a fair, first-in-first-out waiting room we ensure that all visitors get the best possible experience and your site remains available 24/7 no matter how busy it gets.
TrafficeDefencer’s Virtual Waiting Room is a scalable, cloud-based service that sits in front of your website and controls access for all visitors. To learn more about our solution, sign up for a Virtual Waiting Room Trial to demo the product and test it on your live site or mobile app.
Our partner Netacea provides advanced insight and visibility into your web traffic, allowing you to create powerful actions based on deep machine learning insight, using APIs.