How to Avoid Performance Anxiety with a Cloud Approach

By Jay Smith, Lagrange Systems


Do you want to implement a social media campaign around a new product offering but are worried that a spike in traffic will decrease the performance of your website? You are not alone. 


The increase in demand that comes with a successful campaign can pose a risk to your website. If you are not prepared, too many people show up and the site can't handle it. This over abundance of traffic puts a burden on servers and negatively affects page load times, which are directly correlated to conversion rates.  


For example, a 1-second delay in site response time can lead to a seven percent reduction in conversions, and two out of five users will abandon a website if it takes more than 3 seconds to load. For ecommerce sites like Amazon or Wal-Mart that sell commodity products, if one site is too slow, people will move on to another store to make their purchase. 


However, even specialty stores that offer more unique products can experience a negative impact from a poorly performing website under high demand. These sites are more likely to fail outright because customers often impatiently refresh the site, further compounding the problem, converting a wonderful revenue opportunity into an inadvertent denial of service attack. 


And, not all activity is equal. The checkout process is much more draining on site performance than page views and can slow down dramatically or even cause the site to fail. Checkout performance is a major factor in shopping cart abandonment-slow shopping carts often lead to lost opportunity and wasted marketing dollars.


The truth is that these traffic spikes represent tremendous revenue opportunities, but only if the application deployment is capable of handling the traffic. Too often, ecommerce sites are deployed on static infrastructure that cannot keep up with high demand while still being costly during low periods. In addition, poorly written code and database limitations can impact performance and exacerbate the need to scale out your site.  


The key to solving this problem is synchronizing performance with demand. With the cloud, you can provision additional servers during times of high demand and use fewer servers during periods of low demand. However, server scale-outs often take minutes to complete-a near eternity during a traffic spike. This means that your website infrastructure needs to anticipate the need for more servers to ensure that the new servers are actually available when you need them and not five minutes later.  


Over half of today's ecommerce companies are expected to move to the cloud in 2015 and they want their sites to perform well no matter how many people show up. 


For most, the time period between Black Friday and Christmas is when they make the majority of their revenue for the entire year-a time when they can't afford to have people abandon their site due to slow page load times or slow check-outs. Website performance needs to remain consistently fast regardless of demand to ensure that visitors convert to sales. Software solutions can enable cloud-based applications to dynamically scale out and in so that performance is synchronized with demand. This means that whatever your new peak is, your application will scale out to handle it and after the event it will scale back in to keep overall costs down. This allows top line revenue to increase while protecting profit margins.


Optimization tools can also be implemented to support a better user experience specifically designed to improve performance for both mobile and desktop users. This means that your site is fast for everyone.


It is not a matter of if, but when your site will hit a new level of peak demand. The solution is to make sure that when that wave happens you can get out in front of it. Recognize the traffic spike as it begins and proactively scale out the site before performance is impacted. 


Jay Smith is Founder and CEO of Lagrange Systems (