Tribe Logo

Is your load testing fit for purpose?

Author Image
By Luke Hipkiss 03/12/2020

Against the backdrop of this year’s turbulent retail landscape, we’ve seen the emergence of unprecedented opportunities for savvy online brands. Many retailers formerly focused predominantly on the high street have ventured into web sales, while many more have redoubled their efforts to satisfy online customer demand in an attempt to counteract the savage contraction in in-store sales.

In some ways, the wholesale swing towards internet shopping has merely fast-tracked an ideological shift that was already underway.

Nowhere is this contrast between bricks and clicks more starkly drawn than in the period leading up to Christmas – the so-called ‘Golden Quarter’ of the retail calendar. By the middle of November, and with online sales already up 58 percent on the same period last year, experts at IMRG were already predicting that this year’s Black Friday week could see internet takings rise to £5.8bn, outstripping 2019’s online figures by 35 to 40 percent and leaving in-store receipts looking weak.

One size doesn’t fit all

While it’s welcome news for beleaguered brands struggling to tempt customers back to city centres, it’s a correspondingly big ask for underprepared web platforms and understaffed logistics operations.

Most retailers have been mindful of this and have responded by cutting prices earlier to avoid a sudden surge in ordering: IMRG found almost a third of the retailers they looked at were running offers throughout November. Nevertheless, spending will undoubtedly spike in the final weekend of the month – it’s thought that shoppers may spend as much as £7.5bn over just a few days.

Some sites have already had their mettle tested – the online scrum that followed the release of a limited number of Sony PlayStation 5 consoles really pushed the tolerances of major retailers, including John Lewis and GAME.

Counting customers

So, here’s the – virtual – rub. While no-one can entirely predict the volume of online demand, it’s essential that retailers can confidently accommodate an expected increase in traffic so they can continue to deliver a fast and error-free user experience (UX) to their customers.

Knowing how many active users there are on your site any given time doesn’t give you enough information for effective planning, though – mainly because you have no idea what they’re doing. Imagine 10,000 people visit your homepage briefly and then leave.

As far as Google Analytics is concerned, this will count as ’10,000 concurrents’ for the next half-an-hour. On another day, you might welcome 10,000 visitors to your site but this time, half of them embark on searches, while 1,000 place orders.
Two identical concurrent user figures but two dramatically different days’ trading.

So, what’s a better metric?

Once you know how big your store is at peak time – that’s to say how much you’re transacting, how many orders you’re processing and how many searches customers are making per hour – you have a sound basis for calculating capacity.

The best metric of store size is to compare it to your last peak, measured – crucially – by the same traffic profile mix. It’s the only way you’ll know if you can handle, say, 15 percent more orders or more transactions without negatively impacting the user experience (UX).

The only way to thoroughly understand your site’s true capacity is by load testing. Not via a system focused on concurrent users but one that creates a realistic mix of users based on actual traffic patterns.

By performing this kind of in-depth load-testing analysis, you’ll gain important insights about how different behaviour impacts your web provision and, importantly, will have the information you need to accurately assess whether your site can handle predicted traffic over the busiest periods.

At thinkTribe we use load testing protocols that are based on real user traffic data to provide a better overview of what users are doing, including the devices they’re using, the routes they’re following and their ‘drop-off’ ratios. A load test that replicates traffic to a 98 percent accuracy level is the single most effective way of ensuring shoppers will enjoy unhindered progress through your site.

    We just need a few details first




    Subscribe me to your mailing list