My friend Jane asked whether there are any best practices for error pages. Today, I’ll examine that issue, looking at the various error types and the best ways to deal with them. The number-one rule?
Never blame the customer.
User-generated errors are only errors because the system didn’t expect the user to do what she did. For instance, an error occurs when a user doesn’t complete a form or types letters into a U.S. Zip Code field. While these are technically errors, they can be minimized by carefully explaining what the expected user input is supposed to be. For example:
- Enter your e-mail address (e.g., email@example.com):
Enter your phone number (e.g., 555-555-5555):
Logic errors happen when a known series of events results in an unexpected outcome. Coding errors occur due to poor programming or a user scenario not being thought of during a design’s user experience strategy phase. Either way, the user isn’t getting the anticipated results. For instance, I used to use a site that cleared the shopping cart once you logged in. If you logged in midsession, the system would empty your (possibly full) shopping cart. Another site had a short session timeout and would log you out (and clear your cart) after 20 minutes. The average user time spent on the site, however, was longer than that because it was a research tool as well as a store.
These errors can be hard to find, as they don’t trigger any error flags. The system just isn’t operating in ways that benefit users. Many of these errors can be caught by looking for abnormal or unexpected trends in your analytics package. Customer service might have insight into these errors, as it’s most likely to hear from customers with complaints or requests for help.
The two most common server errors are the 404 (not found) and the 500 (internal server error). There are actually many more error codes defined by the W3C, an organization that promotes standards. It’s possible to shield the user from most of these messages. Since the most common one is the 404, let’s see how we can avoid it. The 404 message generally occurs for two reasons.
First, the URL the user is trying to access was misspelled or simply doesn’t exist. In this case, your system should analyze the URL line that caused the error. Most onsite search engines feature a spelling corrector. It’s quite possible something was misspelled. If the search engine is able to come up with results, show that page instead. On the other hand, if your system has a list of common misspellings, you could use that to reconstruct the URL the user intended to visit, though it’s unlikely the user typed in a long URL herself.
A second common reason for a 404 is when the page no longer exists. Most people don’t type in long URLs themselves; instead, they click on links from search engines or another site page. If the URL is outdated, the user will generally get a 404 error.
In this case, you have better options than returning a 404. The most transparent option for users is to have the system perform a redirect on the server side to the page’s new location. This way all your old links remain intact and users see the content they were expecting without knowing anything is wrong. A second option is for the server to return a 301 error, which tells the user the page has permanently moved. This error page comes with the new page address, and the Web browser can redirect the user to the new page automatically. But if you have enough information to generate this error message, you should handle the redirect on your side. You won’t have to worry about whether the browser handles this error properly, and users will have a cleaner experience.
When your pages are supposed to have a lot of product or content and they don’t, users wonder if there’s a problem. For example, when users navigate down your product categories to a product listing page and find only one product in the category, they have a bad user experience. Your online store feels a little empty to them. Similar situations occur when users click on an editorial page only to find no editorial, just links to other pages. These pages aren’t technically errors, but they do make your company look bad.
As a rule of thumb, every product category should have several products, not one or two, and every page that’s supposed to have content on it actually should. If you find your product categories consistently have only a couple of products on them, your categorization probably goes a level too deep. Let’s say you’re a paper company and have a category called “Card Stock,” with five products in that category. You don’t need subcategories specifying the paper’s weight. If you did, each subcategory would probably have only one product in it, and that’s a bad user experience. Let “Card Stock” be your categorization’s lowest level, as it’s the lowest level category that would have enough products in it to justify its existence.
Another no-content problem occurs when a user searches and the result is a page with no products, no links into the site, and no other guidance for the user. I’ve covered what to do with an empty search results page before. Take a look at that column for best practices.
Solicit User Feedback
Users are more than happy to help here. Each site page should have an e-mail link like “Spot an error on this page? Tell us about it!” in the footer that goes directly to whoever is in charge of that site area. This link’s form should automatically list page in question and allow users to give feedback. Unless you operate a Wiki, this is the closest you’ll get to users contributing to make your pages better.
To Err Is Human. To Let Users Know About It Is Foolish
Errors happen. Some are user-generated errors; some are system-generated. Most can be hidden from users, however. Carefully plan your use cases to account for all types of user behavior. Use your analytics package to spot unanticipated errors and identify trends that go against your expectations. When users make a mistake, correct them as quietly and quickly as possible so they don’t waste a lot of effort. Finally, design your site so that every page has adequate content, because an empty category or one with only one product in it looks like a mistake to the user — or at least a poor selection.
Questions, thoughts, comments? Let me know.
Until next time…
While ad fraud has become part of every marketer’s vocabulary, attribution fraud—the practice of gaming outdated attribution models to justify self-serving means—has ... read more
When you’re just starting out as a business owner it’s easy to become wrapped up in the seemingly endless number of metrics ... read more
Something I’m asked frequently at conferences and from marketers is what metrics they should be striving for from their social media marketing. ... read more