Google Panda & Bounce Rate: What Most Experts Don't Tell You

Google's Panda update has slashed traffic for millions of websites, and websites are scrambling for any lifeline that can help them recover their lost traffic and income. But Google doesn't tell us much about what the Panda update is looking for. Therefore, there's been a lot of speculation, and a lot of grasping at straws.

One of the straws being grasped is Google Analytics. It tracks how visitors find and interact with our webpages. Among the data it reports are "Time on Page" -- how long a user sticks around -- and "Bounce Rate" -- how often users back away from the page quickly.

"Aha!" cry the experts. "These are things that Panda is measuring! Improve 'Time on Page' and 'Bounce Rate', and you'll recover from Panda!"

There's a few problems with this logic which you should see immediately. Mainly, How do we know that Panda cares about Google Analytics factors? They are page-specific. Panda is a domain-wide metric. Certainly, enough spammy pages will earn a domain a bad Panda rating. But we have to be careful: metrics used by one Google product aren't necessarily a factor for all the others.

Also: NOT all bounce rates are indicative of spam or poor quality, which is what Panda is trying to penalize. Bounces may happen for a number of reasons:

• The content doesn't match the search query. (e.g.: "I asked for 'orange county', but I meant Florida not California.") Either the user didn't frame the search query very well, or the algorithm failed to understand the query and deliver relevant results.
• Ambiguous search queries. Sometimes it's not Google's fault. If I look for "avatar," I might mean James Cameron's movie, Avatar: The Last Airbender, or a graphic to represent myself on a social network.
• The person performing the searches is doing comparison shopping or some other kind of comparison. He/she needs things to compare.

In all these cases, the cause of the bounce lies in the search process, NOT in the quality of content. If Panda took those bounces as an indication of spam or poor content, it would ruin the effectiveness of the Google algorithm. Google understands this.

Here's another reason why Panda isn't taking bounces into account: many bounces cannot be measured. If someone closes a tab or window, they don't "bounce" back to Google search results, so Google Analytics can't measure it.

All these reasons also apply to the "time on page" metric. There's an additional consideration with "time on page": for most kinds of queries, people are looking for an answer, a solution, or something to buy as quickly as possible. A shorter time on page may be a "win-win" for them and for you.

Yes, Google is trying to serve up results that satisfy users and give them a good experience. But a bounce is not necessarily an indicator of user experience and satisfaction. Time on page is not necessarily an indication of user experience and satisfaction. What IS? Users' feedback.

Google collects various kinds of user feedback: the +1 button, shares on Google Plus, ratings on Youtube videos. But are any of these metrics used by Panda in particular? Yes: site blocking data.

When you bounce from a site back to Google, you see a new option added to the site you just visited: "Block all [domain name] results." This is a way for a user to specify that he/she doesn't want to see anything from that domain. EVER.

When Panda was first unrolled, site blocking data was NOT part of Panda, so it's not a central part of Panda. Panda 2.0 added this site blocking data to its algorithm. Both of those links point directly to Google's official blog announcing these changes, so we know they are so, instead of just speculating they are so.

And (back to speculation) Google delayed adding site blocking data to the Panda process, because (I am guessing) it wanted to make sure that cutthroat website owners couldn't manipulate Panda by blocking all their rivals.

I was moved to write this article after a self-proclaimed Panda expert suggested we remove multiple products or affiliate links from our webpages in order to avoid Panda's bounce rate penalty. I'm no expert, but that seems to me to be flailing. I've seen plenty of pages with multiple products listed that are still thriving under Panda, both big websites like Amazon and small-time webpages reviewing things like "dog halloween costumes."

If you want to eliminate the bounces that Panda actually cares about, be more discerning. Consider what prompts users to block a domain? Namely:

• They arrived on a webpage and looked at it for a nanosecond.
• Not only did the page not appear to match their search query, but it was so bad and spammy-looking they concluded that no page on that site would ever satisfy any search query.

Studies show that most arrivals to a webpage decide whether or not to stay in less than 3 seconds, without even scrolling down. They evaluate whether the webpage header and graphics look professional. They shun excessive, intrusive advertising. They skim the page title and introductory text to learn what the page is about and what they will get out of it. If they see from all these signals that the page has good content, even if it's not the content they're looking for, they're not liable to block the site.

If you want to tweak your page to avoid the "blocked site" Panda penalty, you need to concentrate on how to satisfy/impress the user within the first few seconds of arrival, making your "above the fold" content (everything they see on the first screen) clear, well-presented, and compelling. Prove you know your topic. Demonstrate the webpage's purpose and what someone's going to get out of it. Get to the point.

Don't worry so much about bounce rates. Worry about making a good first impression, one that puts your site in a good light even when a user thinks, "Oops, that's not what I was looking for."