|Crazy Bounce Rate Drop - Starting October 19th|
From October 19th / 20th, my bounce rated on my main ecommerce site dropped to only being about a third of what it was previously. Unfortunately, sales have been way down - maybe off by 30%.
Normal bounce rate on my site is about 65%. Has been that way forever.
Starting October 19th, bounce rate dropped to 39%
From October 20th up until November 5th, bounce rate has averaged about 24%
So basically, bounce rate is a third of what it was, with little changes to the site.
It appears that most of the other metrics only had slight variations when comparing October 19th to November 5th with the previous period of October 3rd to October 19th. for example:
Visits: UP 4.5%
Absolute Unique Visitors: UP 3%
Pageviews: UP 7%
Average Pageviews: UP 2.25%
Time on Site: UP 5.4%
New Visits DOWN 0.74%
Direct Traffic UP 5.73%
Referring Sites UP 7.19%
Search Engines DOWN 2.72%
So New Visits and Search Engine referrals are down slightly, while other metrics are up slightly.
Anybody else seeing this? I am suspecting a GA glitch more than anything else. Otherwise, wouldn't Average pageviews and Time On site be much higher?
Perhaps a link to the official Google definition of "bounce rate" would be in order but I can't find such a page.
Anyway, a question: Are you seeing a lot of newly found "direct" traffic that you didn't have say a year ago too?
I'm seeing something similar re:bounce rate and I know for a fact 50% of my traffic on this site isn't direct as analytics states it is.
It's direct traffic. Email blasts, mail drops etc.
But a component of it also is more targetted referrals.
|Anyway, a question: Are you seeing a lot of newly found "direct" traffic that you didn't have say a year ago too? |
Yes, but my direct traffic (and referrals) has always been real low. So for this year, compared to this time last year, here have been the changes in terms of traffic patterns:
* 10.29% Direct Traffic
Previous: 6.93% (+48.35%)
* 20.10% Referring Sites
Previous: 12.54% (+60.30%)
* 69.61% Search Engines
Previous: 80.53% (-13.55%)
This seems odd for a couple of reasons:
1) I am pretty sure that our email newsletters have been caught in the junkmail filters for both hotmail and gmail accounts. I get lots of people telling us they never got our emails, so I would suspect that direct traffic SHOULD be down significantly.
2) We have been trying to work on SEO, and apparently we are going in reverse, since GA says that search engine traffic is headed south.
I can understand referrals going up, because we only started doing any facebook this year. But still, 20% seems pretty high for us.
In comparison, for fourth quarter 2009, our traffic sources looked pretty much the same they have always looked, namely:
* 6.33% Direct Traffic
* 11.19% Referring Sites
* 82.49% Search Engines
I think that GA might be messed up...
Are these metrics from raw logs or from Google stats?
I'd be inclined to believe the former and if relying on latter that a glitch is involved.
Personally don't use any Google stats...
What is bounce rate exactly anyway? Is there an industry standard that defines how long a visitor needs to stay on a website in order for a bounce to not be counted?
My main website has a 20-22% bounce rate, I have always tried to improve on this number but these days its so hard. I use clicky stats, so I dont know what measurement they use as a bounce stat.
The closest I could find on Google to an official description was from this page - [code.google.com...]
|Bounce rate is referred to as a single-page visit to your site, but is strictly defined as a single interaction request during a user session. For this reason, a bounce rate for a page is also affected by ecommerce transactions and event tracking requests. This is because these features co-exist with page tracking and, when they are triggered, they result in additional interaction requests to the Analytics servers. |
It might be described elsewhere but I haven't found it if it is. It's interesting that Anaytics is mentioned since *I believe* data is gathered from visitors via Google tool bar and not just by your page loading analytics (which you may not even have).
|It's interesting that Anaytics is mentioned... |
Considering that the page is about Analytics that isn't too surprising ;-)
In a recent video [youtube.com] Matt mentions that GA data is only from that acquired through the JS.
Regarding the definition of bounce rate in Google Analytics, I'd read a detailed definition recently, but couldn't find it again. This is the closest thing I found from Google's own mouth - it's one of those things you have to piece together with definitions in different places. Combine it with the quote from Sgt_Kickaxe:
|Bounce Rate |
The percentage of single-page visits (i.e. visits in which the person left your site from the entrance page).
The number of times your visitors have been to your site (unique sessions initiated by all your visitors). If a user is inactive on your site for 30 minutes or more, any future activity will be attributed to a new session. Users who leave your site and return within 30 minutes will be counted as part of the original session.
This field identifies the number of single-page visits to your site over the selected dimension.
ok got it...so when the user gets to your page and doesnt click anything or take any action, just closes the window or hits the back browser button.
|Are these metrics from raw logs or from Google stats? |
These are the stats from Google Analytics. I have access to the raw logs, but the only thing to analyze them is the parallels control panel, and that doesn't have bounce rate.
Takes a bit of work with raw logs, but one can find that bounce rate. That's the one I go with because that is my EXACT TRAFFIC and no filter from a third party. Might want to look at that metric, if you have access to raw logs.
Yes, Google Analytics uses js. If you want to count real people, it can be argued that js tracking is actually a better way to go than site logs, even though some visitors will have js off.
Is there anything to suggest that the type of traffic you've been getting recently... ie, in say research vs commercial... is different enough to cause the changes you're seeing? With regard to traffic from Google, or from sites that found you via Google, this thread might provide some food for thought...
Google & Traffic Shaping - a hidden method to the quality madness?
Also, in the period you're looking at, keep in mind that this was an election year, and that the economic situation is uncertain at best.
Re direct traffic, a PS to the above as I'm catching up with reading after PubCon and just came across this in the update thread... [webmasterworld.com...] ...SEOPTI's Nov 11 post, msg#4229333...
|Analytics shows the preview visitor as direct traffic which is nonsense. |
Take analytics data with a big grain of salt. Quoting myself here: [webmasterworld.com...]
|What this looks like to me, is that GA is tracking each subsequent visit from this particular user as resulting from their initial Google search. This is not in line with reality and inconsistent when comparing to other RSS feed users. |
|Is there anything to suggest that the type of traffic you've been getting recently... ie, in say research vs commercial... is different enough to cause the changes you're seeing? |
Nothing that is obvious. The most popular pages for the period after the drastic drop in Bounce Rate is the same for the corresponding period before. That is the same for the entrance keywords as well. Some slight variation, but nothing major.
I wonder if it might be due to the fact that while I am using the ga.js code in my pages, it is still the (older) version that goes in the footer of the page, right before the closing body tag, as opposed to using the newer asynchonus code that goes in the head element.
On my other site, I still use the same older version of the code, and HAVEN'T seen the same drop in bounce rate.
However, I think the google analytics code might work differently on different sites.
For example, I had been using the ga urchin code up through 2008, and suddenly in late August 2008, it stopped working on my site. GA was reporting zero traffic. I then had to change to the ga.js code (in the footer) to track any traffic.
However, I talked to other web masters who were still using the ga urchin code up through 2009 and they said it was working fine on their sites. So I don't know why the urchin code would stop working on my site but continue to work on other sites.
Just going over my web analytics stats again, and one thing I just noticed is that google is saying that my SEARCH ENGINE traffic has a bounce rate of ONLY 11%
That CAN'T be right!
My DIRECT TRAFFIC, which is from an opt-in email, is about 50%, which SEEMS about right to me.
For Referring Sites, the bounce rate is listed at 48% which also seems LOW to me (I would imagine it should be around 70% or more).
so in short, I think that something is definitely affecting how google measures bounce rate from Search engine referrals.
this is because the google preview tool - its a bug in analytics, it count every request from the bot as a direct visit with a 100% bounce rate, we have the same problem, suddenly more direct traffic in analytis, but also a higher bouncrate from "direct traffic"
Welcome to the forums, swisspanama.
This thread is about a bounce rate going very LOW, and it began on October 19. Instant Previews were announced on November 9 - three weeks later. Also, I can't see how previews would make the bounce rate go lower, only higher.
|Also, I can't see how previews would make the bounce rate go lower, only higher. |
Precisely. The bounce rate for Search Engine Traffic DROPPED from around 65% to around 12% on October 19th.
Direct Traffic remained about the same.
Referral traffic dropped slightly, but I expect that is an anomaly in analytics as well.
Have you seen something in the past where all of a sudden, GA would just start returning garbage numbers for one of the main metrics for no apparent reason? (aside from the above-referenced case where GA stopped supporting the urchin code and moved the the ga.js script).
This may be more of a topic for our Analytics Forum [webmasterworld.com] - but certainly there are times with GA that the data get buggy. I think anyone should have another source of metrics in addition to the freebie, especially if they're going to make business decisions based on the data. I know Google wouldn't see things that way, but I certainly do.
Are you certain that you didn't introduce something on your home page that GA is mistaking for a second page-view/click?
|The bounce rate for Search Engine Traffic DROPPED from around 65% to around 12% on October 19th. |
Agreed, this seems almost too good to be true.
Grasping at some straws, there has been discussion in the Traffic Shaping thread that fits some of what you're describing, but nothing to the degree you cite above....
Google & Traffic Shaping - a hidden method to the quality madness?
The OP in the traffic shaping discussion notes a 20% increase in conversions with no increase in traffic, but with a different set of referrals. Another poster cited 10% decrease in bounce rates. Have your conversions been affected?
Does anything in that thread suggest similarities to what's happening with your traffic?
|Does anything in that thread suggest similarities to what's happening with your traffic? |
Well, unfortunately the discussion in that thread is a bit above my head.
What I do know is that we had no significant shift in conversions either way, no significant shift in the geographic regions (we are an ecommerce site with customers primarily from USA, then Canada, then UK. Then a scattering of European countries and Australia. No significant traffic shifts.
We did minimize the number of links in the left hand navigation / category tree, not by eliminating page, but by making some sections that were one click away from the home page into sub categories, so now they are two clicks away. In all we reduced the left hand category tree from about 45 down to about 30
I am trying to get AWSTATS set up, but I am not really having an easy task of it. I will see if it is something my host can help out with so that I can compare it to GA.