Engineering resolutions for 2019: Speed it up
2018 has certainly been a year. As the days grow shorter, we can’t help but look forward to a brand new 2019 and all the amazing tech trends in store for us. In this article, Debbie Levitt explains the most important resolution any web developer needs to make for 2019: improving the speed and performance of their sites and apps.
How much longer can you put off fixing or improving speed issues on your website, in your app, or the system you offer customers? It’s official: no more procrastinating.
In the past, customers separated a good digital product that they liked from server issues or back-end technical choices that made the site or system slow. They would notice the speed issues but decide that was secondary to their perception of whether or not the product fit their needs, offered delightful innovation, or allowed them to accomplish necessary tasks.
Attention spans have gone from short to bizarrely short. Standards and expectations are constantly on the rise. Most companies have at least one competitor and customers have more choices. Slow internet modems and overnight downloads are further and further back in your customers’ memories.
The user perceives the moments between the visuals loading and the page being usable.
Nearly 1 out of 3 users left the site before the home page was done loading; that’s how little patience we all have. It’s a wake-up call for companies and teams deprioritizing or continuing to put off fixing speed and performance.
Google’s Lighthouse gives us another view of the issues on this page, and let’s re-order them to help us make sense of what the user is waiting for:
- First Contentful Paint – how long does it take the user to see the first text or images on this page. This is fast and passes Google’s standards.
- First Meaningful Paint – how long does it take for the primary content of the page to be visible. Still doing well so far.
- Speed Index – how quickly the contents of a page are visibly populated. Google calculates that it takes nearly 9 seconds for the page to appear fully populated. As 2018 becomes 2019, this feels like an eternity.
- First CPU Idle – this is how long it takes for the main thread to be quiet enough to handle user input. Notice the seconds between when the customer will believe the page is ready to use and when it might be usable.
- Time to Interactive – how long it takes for the page to be fully interactive without delays or, “I clicked and nothing happened.” 12 seconds by Google’s calculation.
If you didn’t like a 4-second wait, you’ll really hate a 20-second wait.
I was once building a realistic, clickable, interactive UX prototyping for a bank’s app. There was a screen where the app hits the server, leaving the user waiting to see their latest banking transactions. I built a 4-second wait into the prototype but received a note back telling me that 4 seconds was ridiculous; change the prototype to show a 2-second wait before revealing transactions.
Later in the project, QA included UX in testing some of the builds to see what bugs or issues might be found, being so familiar with the interface and functionality. The first time I went to load the transactions, it took over 20 seconds for them to appear. That was on internal company Wi-Fi. One can only imagine the speed a customer might experience if on 3G data.
If 4 seconds seemed like an eternity to expect customers to wait, then 20 seconds are off-the-charts unacceptable.
Everything your customer experiences is part of the “User Experience”
More and more, site or app performance is perceived by the customer as part of the user experience. Instead of excusing a slow system, they will be more likely to find slow response times frustrating and a deal-breaker.
And you know this. I’m polling audiences of mostly engineers at the conference presentations I’ve been doing at Agile and DevOps events. Attendees are asked to rate certain software development issues on a scale from 0, representing “customer doesn’t care,” to 10, representing “furious customer might leave us.” This graph shows the average answers across hundreds of attendees at three different events.
The first four deeply affect the user’s experience, their perception of your software, and their perception of your company. The last one about timing and budget is bad for DevOps but mostly doesn’t affect the customer. Perhaps customers will have to wait a little longer for that next release.
Engineers, product managers, and everybody on the cross-functional team are keenly aware that problems that cause users frustration, confusion, disappointment, distraction, or other negative emotions can inspire customers to abandon. They might abandon the process, purchase, product, or entire company.
Anything perceived as a poor user experience is enough to make customers consider leaving. This reminds us of its importance and ability to be make-or-break for our teams and companies. We need to work smarter and get qualified experts in every step of planning, prioritization, and design, and many of the development steps.
2019: The year you finally fix it
You hate slow sites, slow search results, and slow checkouts. These things had positive qualities to begin with! We’ve been making excuses too long and hoping that customers will give us a pass.
As people become less patient, speed and performance will truly kill the customer experience. Fixing server issues, refactoring, and improving database issues should be at the top of everybody’s 2019 list. It’s time to get your DevOps and your User Experience out of the intensive care unit.