With more than $20 billion dollars in total consumer spending, Mother’s Day is the third largest retail event of the year, right after the winter holidays and back-to-school. While more and more people are shopping online for Mother’s Day, many of these shoppers don’t have a lot of confidence in the sites they visit: 64% worry that transactions will fail, and 83% say they’ll take action against a site that performs badly.
Today, let’s unpack a Harris Interactive poll of people’s expectations around Mother’s Day shopping and look at it alongside SOASTA’s recent findings about the impact of just 1-second page delays on Mother’s Day sales.
Who shops online for Mother’s Day?
Last year, SOASTA commissioned a survey from Harris Interactive, asking more than 2,000 people about their online shopping habits around Mother’s Day.
One of the stats that’s been cited a lot since we released that research is this one: 29% of Americans plan to shop online for Mother’s Day. But what I find much more interesting is the breakdown by age. According to our survey respondents, 39% of people aged 35-44 and 57% of those aged 18-34 say they’ll shop online.
What do Mother’s Day shoppers worry about most?
In the same survey, we found that two out of three shoppers worry their transaction will fail. Security, slow load times, and site crashes were among people’s top performance concerns. While it’s no surprise that security is at the top of the list, site performance worries popped up in a number of guises, from concerns about slow page load times to fears that the entire site will crash.
What’s also interesting is the fact that 20% of shoppers worry that they won’t be able to tell if their transaction successfully completed. This user mistrust speaks to so many ecommerce usability problems that we’ve all experienced at one time or another:
- Pages that hang or freeze during checkout
- Error messages during checkout
- Non-existent or sloppy confirmation pages
- Lack of confirmation email from retailer
How do people react to negative online experiences?
According to our survey, 83% of Mother’s Day shoppers will take action after having a poor user experience. Common next steps: going to a competitor’s site (49%), using social media to tell friends about their bad experience (18%), and vowing never to return to the offending site (33%).
If you’re a retailer, these numbers might make you cringe. No one likes to think about the negative impact of a poor user experience on customer retention and word of mouth. But there are tangible, here-and-now bottom-line effects, too, which you should be at least as concerned about…
How much do slow load times hurt retailers?
If you’ve been in the web performance space for any length of time, chances are you’ve encountered a few case studies that show the correlation between slow load times and reduced conversions.
If you’re newer to this space, then this information may be new to you, but you probably intuitively feel that yes, it makes sense. What may surprise you, however, is just how dramatic the impact of seemingly small changes in load time can be.
To illustrate this, I worked with our data science team here at SOASTA to analyze data for four customers who use our mPulse solution to monitor their web performance and business metrics. These customers are all leading online retailers who do a booming business around Mother’s Day. We wanted to understand how their sites performed in the four weeks leading up to and including Mother’s Day, so we aggregated the data for every user session — totalling more than 81 million unique sessions — between April 27 and May 19, 2014. Then we used our Data Science Workbench — which is now embedded in mPulse — to do some fun slicing and dicing.
Finding 1: The sweet spot for load times was between 1.9 and 4.2 seconds
This histogram showing the distribution of converted and unconverted shoppers is a thing of beauty. As you can see, the blue line, which represents conversion rate, more or less follows the traffic distribution. The orange bar segments represent the converted traffic.
Retail sites aim for conversion rates of at least 2%, so we wanted to find the performance sweet spot, AKA load times that yielded conversion rates of 2% or higher. We found that the sweet spot for this set of sites was between 1.9 seconds and 4.2 seconds. The highest conversion rate — 2.4% — correlated with the average load time of 2.5 seconds.
Finding 2: A 1-second difference in load time yielded up to 20% more conversions
Even within the sweet spot, small changes in load time added up to significant conversion gains and losses. Pages that took 3.5 seconds to load – a mere 1 second slower than the 2.5-second peak performance – had a conversion rate of 2.2%. This may not sound like a big difference from the 2.4% conversion rate enjoyed by pages that loaded in 2.5 seconds, but it actually represents an 8.3% decrease in conversions. That’s extremely significant.
This drop in conversions gets even more dramatic with each incremental 1-second page slowdown. Pages that took 4.5 seconds to load experienced 13.6% fewer conversions than pages that took 3.5 seconds to load. Pages that took 5.5 seconds to load experienced 15.8% fewer conversions than pages that took 4.5 seconds. And so on.
Another way of expressing conversion losses: Pages that took 4.5 seconds to load experienced 20.8% fewer conversions than pages that were just 2 seconds faster. Crazy!
I whipped up this graphic to better illustrate just how dramatic these seemingly small changes in conversion rate stack up side by side:
If you’re more of a glass-is-half-full kind of person and want to put a positive spin on all these numbers, then think of it this way: faster pages enjoyed significant conversion gains. Even differences of 1 and 2 seconds within the performance sweet spot made a huge difference. Pages that loaded in 2.5 seconds experienced 9% more conversions than pages that took 3.5 seconds to load, and 26% more conversions than pages that took 4.4 seconds to load.
At 8.5 seconds, conversion rate plateaued at 1% and declined gradually from there. (This is commensurate with other research I’ve done, where we determined that the “performance poverty line” for most retailers is around the 8-second mark.)
Finding 3: While OS X users comprised only 16.9% of the total user sessions, they made up 34.9% of converted shoppers
The graph on the left shows the breakdown of total user sessions by OS. The one on the right shows the breakdown of conversions by OS. Bright blue is Windows 7 and bright green is Mac OS X.
The percentage of Windows user sessions (44%) and converted users (42.3%) is more or less commensurate. What’s really interesting is the fact that, while OS X users comprised only 16.9% of the total user sessions, they made up 34.9% of converted shoppers. Possible interpretation? Mac users convert roughly twice as often as Windows users.
While site owners should, of course, aim to provide the best possible user experience for all users, they should definitely ensure that they continue to optimize for Mac users. They should also drill down into their page data for Windows users to ensure that that they’re not inadvertently delivering a subpar user experience to that cohort of shoppers.
Shoppers notice and care about performance issues more than you think, and they’re extremely sensitive to even minor problems. Making your pages just one second faster or slower can have a measurable impact on your bottom line. On Mother’s Day — and every day — site owners should aim to deliver the best possible user experience. This means having full realtime visibility into the performance of your entire site and ensuring that all pages are secure, available, and fast across all platforms.
About the Author
Tammy has spent the past two decades obsessed with the many factors that go into creating the best possible user experience. As senior researcher and evangelist at SOASTA, she explores the intersection between web performance, UX, and business metrics. Tammy is a frequent speaker at events including IRCE, Shop.org Summit, Velocity, and Smashing Conference. She is the author of 'Time Is Money: The Business Value of Web Performance' (O'Reilly, 2016).