8 Rules to Avoid a Manic Cyber Monday

Every year, Cyber Monday becomes bigger and bigger, with last year’s sales overshadowing even that of Black Friday by $1.56 billion. With so much potential for Cyber Monday, risking website, application, and API availability and performance is simply not an option anymore.

Anyone with an interest in revenue, brand exposure or even service-level agreements (SLAs) on Cyber Monday should have a vested interest in how your company is preparing to handle the influx of visitors to your website. However, we know that trying to predict all possible failure points can seem like a daunting task. That’s why we’ve crafted an important list of dos and don’ts for optimal performance on Cyber Monday.

4 Must Dos

1. Functional Testing: Validate essential functionality like login processes, filter functionalities, sort capabilities, form fields, drop-downs, and cart functionalities. Ensuring that users can browse your site and interact with features in a predictable manner is key for success.
2. Browser and Mobile Testing: Identify your customers most commonly used browsers and devices, then run tests for usability and performance across the various combinations of devices and browsers so you know that the digital experience you provide is incredible regardless of how each visitor gets there.
3. Load Testing: Maybe one of the most important tasks for any business looking to introduce a flash sale on Cyber Monday — load testing. Understanding what the breaking point is for your website, web application or API that supports user transitions is key in ensuring 24x7 availability on the biggest sale day of the year.
4. API Transaction Monitoring: One of the most overlooked, yet imperative, tactics is to ensure that you set up API transaction monitors in your synthetic monitoring tool. These monitors work to ensure that the APIs that support key functionalities of your website and user experience are available, fast, and returning the right information. Oftentimes, API calls are done in sequence and pass on contextual information from one call to the next. Make sure your tool has the capacity to monitor the entirety of the transaction, including any information that may need to be continuously called from one interaction to the next.

4 Don’ts

1. Just Monitor Website Availability: Nowadays, it’s not necessarily enough just to know that your website is up. If your website is slow, it might as well be down. On Cyber Monday, shoppers aren't going to have the patience to wait for your site when they can find what they need on a competitor’s site that runs 10 times faster. Be sure to monitor user journeys. Set up synthetic monitors that can mirror an entire user transaction to validate that your site is performing as intended and functioning as designed.
2. Let SLAs Slip: SLAs, whether you’re providing the service or are a consumer of the service, are always at risk in "stress" scenarios where there’s a larger-than-average interaction rate for your digital assets, be it a website, web app, mobile app, or API. Managing SLAs by tracking performance deviations for the agreed upon threshold can make all the difference in remediating performance problems, and it doesn’t have to be hard. Use a synthetic monitoring tool to manage SLAs by setting benchmarks and automating alerts when something strays from the performance level you need.
3. Send Alerts to Everyone or Just One Person: Alerts are the first line of defense for operations teams and businesses as a whole. Neglecting to configure the alerts can almost render a monitoring tool useless. Sending big blasts of alerts often causes a mess — either everyone is trying to solve the problem at the same time without coordination, or the bystander effect kicks in and individuals think other people will respond and fix the problem. However, it’s also problematic when alerts are too isolated. Depending on one individual to remediate issues might not always make sense. Make sure your tool supports alert routing based on factors like error code, step level failure, etc., so that when something goes wrong the right team is notified rather than everyone or just one person.
4. Forget a Remediation Plan: It’s inevitable that something will go wrong, but the only thing that matters is how you fix that problem. Mean time to resolution (MTTR) is a KPI that most ops teams are measured against, and from a business perspective, it should be something that's taken into consideration because it essentially means how long your customers are waiting for you to solve the problem. Having a solid plan of action, including alert routing, root cause best practices, responsibilities, availability of team members, etc., can drastically decrease MTTR, making it faster for your customers to access your website or application.

Don’t let performance slip when it matters most. Putting these tactics into practice now can help keep your customers happy and accelerate revenue on Cyber Monday and throughout the busy holiday shopping season.

Saoirse Hinksmon is a digital content marketing manager at SmartBear, an information technology company that delivers tools for application performance monitoring, software development, software testing and API management.