Three must-know load testing tips to prepare for Black Friday & Cyber Monday

Yippee ki-yay, performance testers

Time with family, sledding, snowball fights, hot chocolate, great movies and ONLINE SHOPPING.

These are all traditions we hold dear, but online shopping is very important to many people around the world, and many of those people will simply not be patient if your website or app takes too long to load.

When somebody wants to order a football for their child, jewelry for their significant other or any of the zillion other things people exchange during the holidays, your website better be fast enough to display it and get people checked out quickly.

Start Load Testing Now!

If you haven’t already been continuously load testing your website or applications (tisk, tisk), then now you know it’s time to start.

Here are three bits of wisdom to help you get the most from your load tests in crunch time leading up to Black Friday and Cyber Monday.

Plan the user scenario(s)

The best place to start is by figuring out the user scenarios that you want to test. For quick reference, a user scenario — which can also be referred to as a “load script” — is a program that describes every action Virtual Users in a load test should perform.

Let’s say you’re running an e-commerce website where users are required to login using a username and password. That’s an important part of your flow, and you need to include those HTTP requests in your user scenario.

You’ll want to make these requests dynamic so the Virtual Users aren’t using the same username and password in every iteration. You can do that by using data stores. Learn more about that in this helpful article.

So, how do you do get the most actionable data from your load testing?

Our most popular option is the Load Impact User Scenario Recorder, which is conveniently located on the Chrome Web Store.

After you install the extension, simply click the extension, select “Start Recording,” and then go through the user scenario you want to test on your website.

When you’re done, simply click the extension icon and select “Stop.” You’ll then be taken to another tab, where your Load Impact account will open and you’ll save the new scenario.

It’s super simple, and that’s why it’s been so popular for our users over the years. Check out this quick tutorial if you want to see it in action.

Determine the number of concurrent users in your load test

Let’s get to the fun part of determining how many concurrent Virtual Users to configure in your load test. We recommend starting with this formula:

Hourly Sessions x Average Session Duration (in seconds) / 3,600

Google Analytics has made it super easy to find the data you need to fill this formula. Here it is in seven easy steps.

  1. Login to your Google Analytics account
  2. Click the “Reporting” tab across the top
  3. Select “Audience” in the sidebar menu
  4. Click “Overview”
  5. Set the time period you want to base your data on in the top-right corner
  6. Make sure “Hourly” is selected in the top-right of the line graph
  7. And the data you need is right in front of you!

If you want a little more insight into the reasoning behind this magic formula, check out this blog post, aptly named: Determining Concurrent Users in Your Load Tests.

Establish benchmarks with load testing

From the mathematical formula and insight from your analytics, you’ll easily figure out how many concurrent users your website serves now during-and-off business hours.

And if you’ve been around since this time last year, then you might already have data from last Black Friday and Cyber Monday — and that could be worth its weight in gold.

Run a few tests with varying durations, and you’ll get the behind-the-scenes view of your website’s performance and establish important benchmarks. It’s also very likely that you’ll discover little optimizations to make immediately, too.

Once you’ve done that, when you run your larger load tests to understand your site’s scalability, you can more easily correlate how much additional traffic effects load times.

For instance, maybe 10x the normal traffic level increases load times 5x, but 15x the normal traffic crashes your entire site.

When explaining the need of optimizations or more infrastructure, it’s always good to have this data and these comparisons available to help other stakeholders, perhaps marketing or the C-suite, understand the site’s overall performance.

And for fun, here are some other potential bottlenecks you could find:

  • The need to compress images on your website
  • Adding more caching to your middleware layer
  • Optimizing APIs from external resources
  • Adjusting your infrastructure subscription (Could need more server space, or you could need less. We’ve seen both!)

There you go. Just as the headline promised, those are three tips that will help you be a better load tester this holiday season, and if you need more help, check out our Knowledge Base or reach out and ask us a question.

Look at the beauty in this image. A nice sunset, some trees and the implication of speed. Now, who’s ready to sign up and start using super-sweet load testing software?