This is a story about errors: It starts with a saying that’s been attributed — in error, appropriately — to the ancient Chinese:
“May you live in interesting times.”
At first glance, that looks like a blessing: Who wouldn’t want to live an interesting life? But it’s actually a curse — wishing you an existence full of turbulence and traumatic incidents.
When you think about website errors, it’s easy to see them as curses, bumps in the conversion process that cost you both in terms of sales and resources to fix them. But actually, they’re the opposite of ‘interesting times’. Look closely and you’ll discover the opportunities lurking inside them.
Be honest: Do you have the time to tag every type of error that your website visitors encounter? Error tagging is usually one of the first areas to fall by the wayside when things get busy, especially when you’re more focused on squashing bugs than identifying them.
Conversion teams say errors are being tracked but they’re usually talking about a subset of what your website visitors come across. There’s an understandable narrowing of focus — the things you consider important gain weight, while other things wither away.
Everybody at the error party
An unfortunate example of the rare pig-in-cupcake error.
The attention tends to cluster around common errors such as payment issues and technical problems like session timeouts or failing back-end components. Because these errors are so well understood, they end up forming the default position — “These are the things we know go wrong so they’re what we’ll focus on fixing.”
Of course, there’s an obvious benefit to focusing your resources on that default: You can easily reach the minimum level of performance required. But staying in that mild discomfort zone means you don’t address less obvious issues that can give you an advantage.
Think of assessing errors as like being at a noisy party. The loudest attendees can seem like they have a great deal to say but when you listen closely, they could be making the same old boring arguments. Focusing on the reserved partygoers can reveal some much more valuable points.
It’s natural for you to wonder whether putting in extra effort will deliver a large enough benefit. But getting down into the detail — talking to the quieter guests at the party — reveals lots of different types of errors that affect the user journey.
Technical issues and payment glitches are the most obnoxious guests at the error get-together. They’re more likely to cause customers to drop off, so they’re more likely to get your attention. But taken together the whispers from validation issues, problems with promotion codes and disruptive business rules can add up to a cacophony.
We all need validation
Eating the doughnut was not the error. Wearing that tie in the first place was.
Let’s take validation errors as an example. They’re boring, right? Just one of those things that everyone has to put up with as a cost of doing business online. Nope…
Even the humblest validation message hold insights into customer behavior. You need to stop seeing them as hygiene point and instead measure how often they occur for different groups, for example visitors who converted vs. those that don’t.
Our analysis shows that validation errors are a common feature of sessions that convert. So, while we don’t want them getting in the way, they can be a good sign of a healthy session. Your intent with this type of error should be to reduce how often they happen and explore the patterns they produce — they’re always telling you something.
Listening to the canary
This man has allowed a bird to land on his tongue. Serious error.
You have to identify positive behavior hidden within seemingly unimportant data. When those positive things stop happening — such as when validation messages go wrong — it’s all going to go very badly for you.
Those small errors are the canary in the coal mine, the indicator of a much bigger problem waiting to explode.
Login errors are a great example. It’s so common for a login validation to be fired that you could take an approach to modelling the conversion journey that expects to see it happen. Most of the time, you’ll end up being right.
But looking across industries, our data shows the same pattern: Login errors contribute to conversion and are commonly experienced in converting journeys. When that no longer happens, there’s a direct link to dropped conversions.
Dealing with the pigeons
This should be a picture of pigeons. We’ve placed some ducks here in error.
Do you know your error message patterns — how, why and when visitors encounter errors on your website? Most people don’t. They’re missing a trick and it’s one you can master.
Your error message patterns are — at best — causing friction. At worst, they’re on course to becoming a new source of basket abandonment.
It’s also very easy to think of business rules (such as restrictions on where you deliver or an item being out of stock) as a tedious inevitability. Everyone expects them to be a pain and lots of your visitors will be resigned to a bad experience. But the effect of your business rules on customer journeys is measurable and can directly contribute to visitors abandoning their goals.
Business rules have become like pigeons in a city. They’re so common we tend to ignore them. The streets may be splattered with their unpleasant consequences but we rarely track those issues back to the source. I’ve encountered websites with hundreds of different ways of saying, “We simply aren’t going to do what you expect of us.”
Finding creative ways to disappoint someone doesn’t make up for the disappointment you caused in the first place.
Your error in handling errors
Here’s a man whose terrible error was to sleep in the presence of a Sharpie.
We look at a wider spread of errors. For example, SessionCam considers promo codes in a category all of their own. That’s because in our analysis, they come up so frequently. You could see them simply as a variant of validation and business rule errors but they’re so common that they deserve that specific attention.
As you think about turning errors into opportunities, look for areas you’ve always wished you’d spend time on but never seem to get to.
It could be an application process that just works and ‘never’ goes wrong. Maybe it just never gets reported. Or you may discover errors lurking in obscure or infrequent journeys that don’t get addressed because the effort required seemed too high.
Errors are rich data. They’re a great proxy for customer struggle and can provide real evidence of dormant problems. But we don’t expect you to have to pick out which errors are costing you the most. Our machine-learning algorithm, the Customer Struggle score, automatically exposes error messages that are affecting visitor behavior and ranks them according to the revenue at risk.
You should know how errors are interrupting your visitors, the impact they’re having and how often they occur. Remember, not all errors are equal and spending all your resources on the ones that shout the loudest could leave you dealing with interesting times.
Images by Ryan McGuire.