If metrics for websites had a popularity contest, bounce rate would certainly be a top candidate for winning it. The bounce rate for a page is typically defined as the percentage of visitors who leave the site after visiting only that one page, without triggering any events or actions. (It can also be calculated globally by aggregating the bounce rates for all the pages on a site.)

This metric is the focus of many questions I get while teaching our course on Analytics and User Experience, likely because many organizations track it. However, optimizing for bounce rates — both of individual pages and especially site wide — is not always worth the effort.

Why is the bounce rate so popular? One explanation is that it is fairly simple to understand, and readily available to track in most analytics tools. Even those who don’t specialize in analytics data — or even know much about it at all — have at least heard of this metric and have a basic understanding of what it means.

While aiming to reduce bounce rates isn’t bad, overfocusing on this metric may lead to bad UX-design decisions.

Not All Bounces Are Bad

Consider how you navigate around the web when you’re searching for an answer to a question. If you’re like me, you go to a search engine, type in some keywords, and then proceed to open several of the resulting page links in new tabs, focusing on those results from websites you know and have used in the past. (This behavior of opening pages in different tabs is page parking; our studies with millennials show that it is particularly popular within this age group.) Next, you begin to read the articles on those trustworthy pages you’ve opened. But, because you are looking for a specific answer, once you’ve read the information on a page, chances are that you simply close the tab whether you found the answer or not. Or perhaps you leave that tab open while you navigate through your queue of tabs, but you never click to any other page on that site.

In either of these cases, your visit would be measured as a bounce! But, did you have a bad experience? Not necessarily. You read the information you needed, and just because you didn’t need any additional information from the site, your visit should not be automatically considered a failure. Each of your single-page, bounce visits moves you toward becoming a loyal user.

Unfortunately, normal bounce-rate calculations don’t take into account this type of short-yet-successful visit. Rather than looking only at how many people bounce, we need to build and measure loyalty. When sites become too preoccupied with optimizing for bounce rate, they make terrible UX decisions just to get people to click on a link. Any link, please, at any cost to the overall user experience.

Often, this line of thinking leads to artificially splitting one unit of content into multiple pieces — articles divided into multiple pages, “learn more” links hiding essential details, and product prices not displayed upfront. Adding artificial steps to a simple task degrades the user experience and may cause users to forever abandon the site after that second click. In other words, by reducing the local bounce rate, you sacrifice visitor loyalty. Without this “improvement,” those who bounced may have actually come back and converted in subsequent visits. Optimizing a page just for the sake of a lower bounce rate can end up harming the more-important, revenue-producing conversions that occur farther down the user journey.

Optimize for Returns Instead

Rather than getting caught up in the bounce-rate race, consider the reason behind looking at bounce rates in the first place: assessing whether visitors engage with the site and are interested in the organization. Building user engagement and loyalty happens over time, not in a single visit. So don’t get obsessed with getting people to click. Visitors may bounce, but what’s more important is getting them to come back.

To determine whether your site is building a loyal audience, look at the frequency and recency of users’ site visits. Tracking the percentage of users who continue to return to your site — even if it’s only to one page — on a regular basis is a much better measure of how compelling your offering is than the bounce rate. If your aim is to identify individual pages that fail to engage users and degrade their return rate, use your analytic tools to view solely those new visitors with only a single visit to your site and analyze their path through the site. Hone in on this group of users who have not returned and investigate their behavior to get a clear picture of the low-performing pages.

In A/B testing, don’t use bounce rate as an indicator of quality and usability. Instead, optimize for the conversions that happen lower in the funnel, and are much more closely linked to revenue or to your main business goals. Ask: how does each of the design variations contribute toward that deep-end goal? Don’t lose sight of these meaningful conversions by getting caught up in a shallow metric that happens to be easy to measure.

How You Should Use Bounce Rate

A high bounce rate is often an excellent signal that a page is underperforming compared to other pages of a similar type. Make sure that you have a valid reference point for the page — you can’t compare the bounce rate of a category-landing page to that of a content-focused page. Each type of page will have a different range of typical bounce rates. Look for outliers among each page type to identify those problem pages that do not retain qualified visitors either because of usability issues, or because they are falsely advertised and mislead users (due to poor referral links, overzealous marketing in ads, or the page having a high SEO ranking for less-valuable keywords).

Another reason for a high bounce rate could be that, while a lot of traffic is attracted to a page, those visitors are outside your target audience and not interested in your main products and services, so they quickly leave the site. If this is the case, look for patterns in the content topics of those pages and either edit or remove those outside your actual audience’s scope, to better focus your content strategy on your true customers and the topics that they find appealing. (How would you know if the traffic is not your target audience? Try segmenting by traffic channel or referral source to better understand the context of bounced visits, and look at the search terms used when people bounce. Also check who the bouncing users are: are they one-time visitors who bounce and never return or are they visitors who return frequently? If the former, they are likely outside your main audience. Don’t worry about returning users’ bouncing, as they clearly come back — new users’ bouncing is a stronger signal for page-design, content, or referral-link issues.)

Only when you identify these problem pages can you develop a plan of action to uncover and solve the actual issue. After a fix has been put in place, monitoring the bounce rate will show whether the issue has been resolved, and the page has returned to the acceptable range for that page type.

Note that here we are referring to the bounce rate for each individual webpage on a site, and not the site-wide bounce rate. The site-wide bounce rate lumps together performance for every page, and it’s impossible to take any specific action based on that measurement. The site-wide bounce rate is a vanity metric not worth tracking and reporting.

Conclusion

Use bounce rate as a red flag to find possible issues lurking on your site, but then switch your focus to large site goals: purchasing a product, contacting for a quote, registering for a newsletter, signing up for an account, returning to the site as a resource for information, and so on. Ask how each page supports these bigger goals and don’t push for single-visit actions. Optimize for the deeper end-goals that often take multiple visits to accomplish and keep your focus on the more important task of encouraging continued engagement and building loyalty. Of course, increasing loyalty and reducing bounce rates aren’t mutually exclusive. But strive to reduce bounce rates via usable means, such as providing relevant related content, and not by adding gimmicks.

Remember that metrics only measure what happened, and can not tell you why it happened. Sure, a high bounce rate could indicate an issue. But it can also indicate that the page is doing a wonderful job of providing the user with everything they needed to know on a single page. Beware of making design decisions aimed solely at chasing that second click — instead, focus on optimizing for the long term by providing quality content that convinces people to return in the future.