Bounce Rate 86% - Is this possible?

NO it isn't

[ergophobe]

So I have seen some strange things with Bounce Rate in Google Analytics. One site which has hovered at about 16-18% for a couple of years, shot way up (over 50%) and has slowly come back down. Now 32%. Could these numbers be accurate. Must be, right? GA would never lie.

So how about this. I have another site which I just added to GA. It has an 86% bounce rate. I KNOW that's not right. How do I know? Because there is exactly ONE PAGE on that "site". So I KNOW the bounce rate is 100% right? There is not a link on that page. It's just a domain placeholder.

So unless 14% of the visitors are sitting there staring at the site right now. Staring. Then refreshing. Then staring. Doesn't seem real likely now does it?

It makes me think I could really get the bounce rate down if I added a second page ;-)

[g1smd]

If someone refreshes the page, then leaves, they aren't counted as a bounce.

[incrediblehelp]

Sure it is possible. Someone types in something and gets to a page that is not relevant ot the query then the bounce rates is probably 100%.

[ergophobe]

incrediblehelp - I think you misunderstood. 100% is the expected rate.

g1smd - I thought of that. But really, there's pretty much nothing there. Why would 15% of visitors refresh the page? I guess that's like asking why many of the top searches are for domain names. It's just so. So you're probably right.

[Receptional Andy]

I don't know how easy it is to get the info from GA, but (using other systems at least) I would just look at some of the visitors who didn't bounce and see what the actual page views were. To be frank, it wouldn't be the first impossible data I've seen GA provide, though.

[MrWrite]

Bounce rate is supposed to be a bad thing as people come to your site and leave without going to any other pages. But perhaps you provided exactly the information they were looking for on your website and left happy. My bounce rate is pretty high (about 65%) and I like to think that people drop in, get what they want, and leave happy, but I could be wrong...

[ergophobe]

MrWrite - understood. The curious thing is how a "site" with only one page can have any bounce rate below 100%

Andy, I'm not sure GA has that data, but I'll have to check around.

I think g1smd must have the answer though. Maybe they think if they reload the page, the navigation that failed to load the first time will suddenly appear. It's kind of an official-sounding single-word domain, so they may think something went wrong and if they reload they'll see the rest of the site. Ha!

[Lexur]

More ideas:
- 14% has some kind of failure or bad setup or privacy tool to prevent its browsing path to be recorded
- 14% are the real readers of your page and spend/waste there more time than GA takes as session and are no counted as bouncers
- 14% close the browser or tab without next page to see and the GA thinks that's not bounce

[ergophobe]

BTW, it might seem like idle curiosity as I have no interest in the stats on this site. What it seems useful for is figuring out what types of error and bias there might be in bounce rates on sites I might care about. So thanks for the ideas everyone.

- 14% are the real readers of your page and spend/waste there more time than GA takes as session and are no counted as bouncers

Six months or so when I last read up on bounce rate, GA counts as a bounce any visitor who doesn't visit a second page. They have no way of knowing how long a single-page visitor is on that page. To figure that out, GA would need it's javascript to communicate with the server (i.e. an AJAX request or some such where data gets sent without a page load). That does not happen (you can test that simply enough by looking at a page that has some plain text and GA on it and watching to see if there's any traffic through your connection).

#3 - closing the browser tab would fall under the same category as #2, wouldn't it? GA doesn't know and doesn't care if the user closes the browser entirely, closes the tab or just goes to a page for which google has no tracking (at this point, that would be a text file on the hard drive I suppose!).

#1 - I don't see that either. GA doesn't use the browser path to calculate Bounce. It's simply a measure of whether or not the GA javascript for that site gets requested again or not by that user.

[cgrantski]

Are there any links to this page? What about arriving at the page, backing out, hitting the same link again to come back? If it's a domain placeholder, it could be search-indexed by now, right?

[ergophobe]

That is possible. It actually is indexed and gets a tiny amount of traffic (half dozen SE referrals per day) so it is, I suppose, possible that someone is actually coming in from multiple searches, so there are in fact two page views of that single page.

I guess the thing to do is check IPs and requests and see how it looks.

So those are reasonable scenarios
- people reload (thinking perhaps the navigation portion failed to load)
- people find it via multiple searches or otherwise forget they've already clicked through.

That might explain it. And given the small numbers - a couple hundred visits, there's room for a lot of error. So the 14% non-bouncers surprised me, but I suppose it is possible, despite what I was thinking.

It will be interesting to see how it evolves as the numbers grow.

[ergophobe]

Just a look through the logs and saw something new that explains this.

There is more than just one page on the site! I just didn't know it because it wasn't "on" the site.

So what do you think of this URL from the GA report:
search?q=cache:PO4xfadsfasdfasdf:example.tld/+keyword1+keyword2+Keyword3&hl=en&ct=clnk&cd=4&gl=ca

[explorador]

MrWrite - understood. The curious thing is how a "site" with only one page can have any bounce rate below 100%

Depends on hows the page you put there.

• People might hit refresh thinking something went wrong.
  
• Depending on your domain previous content, old visitors might stay wondering what happened... And perhaps hitting refresh, or back on the browser only to get in again.
  
• Perhaps your domain name is interesting... Some webmasters won't even read the page, but would stay at a site trying to figure out its traffic. I've been there with parked pages (those with links) moving the mouse to find out where the links go.
  
• You mentioned no links, so, then whats on your page?
  

There is more than just one page on the site! I just didn't know it because it wasn't "on" the site.

So what do you think of this URL from the GA report:
search?q=cache:PO4xfadsfasdfasdf:example.tld/+keyword1+keyword2+Keyword3&hl=en&ct=clnk&cd=4&gl=ca

Sounds like people trying to visit your site via cache and thus, calls to your server for images and files that are not there.

[ergophobe]

It's just a one-page article regurgitated from other sources. Literally, we're talking 10 unique visitors per day. A project I'll get to probably never.

It's a single word domain, so you might expect it to be official. I thought it might be the next big class-action lawsuit. Never materialized, but I figured that whoever got mesothelioma.com probably bought quite a few domains that didn't pan out either and for $8 bucks....

I suspect people are looking at the cache to see what happened to the rest of the site because they expect either link farms or serious sites on one-word domains. Also, they could be domainers looking for info on the site.