Google Analytics Numbers have Mysteriosuly Jumped
2007-07-21 by Yohay
Google Analytics‘ Average Time on Site has mysteriously jumped to much longer times. This big leap in numbers on past days is strange. Am I supposed to feel better now, believing that my readers have actually spent double time on my sites?
I remember seeing an average time on site of 2:50 minutes on a specific day, and suddenly I see 5:23. On another day, I remember seeing about 3 minutes, and now I see over 7 minutes.
This week, Google Analytics abandoned their old interface. Only the new well styled interface can be viewed. I still miss the “by hour” analysis that isn’t seen in the new interface. Maybe there’s a connection between both events.
Anyway, I haven’t seen a report on it elsewhere on the web (using Google’s search).
Did anyone else run into this strange phenomenon?
Update, May 3rd 2009: Also in my forex blog, numbers have moved strangely.
Check out about a Forex Demo Account.


זה פשוט – לקוראים האומללים שלך כנראה לוקח הרבה יותר זמן לקרוא באנגלית מאשר בעברית… (:
חוץ מזה, פונקציית “לפי שעה” לא הלכה לשום מקום, היא עדיין שם. תסתכל טוב בכל הלשוניות.
Islay says that reading in English makes my readers stay longer, becuase it’s hard. She also claims that the “by hour” metric is still there. Where?
I have the opposite issue over the past week – my numbers dropped drastically in terms of visitors. Very strange…
Stephanie, that is very strange indeed… Did it happen to you with the same data? For example, when I looked at Monday’s data on Wednesday, it showed around a 2 minute average visit. When I looked at the same data (Monday) on Friday, it showed a much higher number.
Anyway, you’ll get a visit from me
We’ve also experienced the same problem this week (our time on site numbers nearly doubled); I found your blog post by googling for answers. Drop me a line if you find something out? I’ll come back and comment if I do…
Deanna, thanks for dropping by.
I’ve sent a support question, and I’m officially supposed to be answered in one business day.
I’ll keep you up to date…
Great, thanks!
I just found a couple more people having this issue because GA installed an upgrade this weekend, as well as the abandonment of the old interface… no word yet if the new numbers are correct or not.
I’ll keep checking in on what I find out!
Hi. I also noticed that my avg time on site recently increased dramatically, from around 3 minutes to about 7.5 minutes. One thing that makes this especially strange is that the change has affected historical data going back to October of last year. It seems that GA is interpreting the same raw data much differently now. Very curious to hear an official explanation from them about what’s up.
Amos, that’s exactly the same experience I’m having. Yup, a totally different interpretation.
I still didn’t get answer from them…
Same issue … I remember TOS in about the 4 minute range, now it’s about 8 minutes … happened within the last three or so weeks. If I look at Visitors / Visitor Loyalty / Length of Visit it doesn’t make sense since only 12% of visitors are at the site more than 600 seconds (10 minutes) … unless some are there for a VERY long time!
Greg, I’m happy to hear that yet another person noticed this strange data. It sure doesn’t make any sense.
In the meantime, the guys at GA got back to me, asking fro a screenshot. Luckily, I had screenshot from July 11th (prior to their update). I took a fresh screenshot of the same data, and it clearly shows the leap in average time on site. I sent it to them, and I’m waiting for explanations…
That’s so fortunate that you happened to have that old screenshot, Yohay! And surprising that Google’s coverage of this important question comes down to one user sending in screenshots.
Amos, it sure is fortunate. I made the screenshot for another user of the site. Still nothing on their official blog.
Me again… I’ve got some historical data showing my jump in numbers in csv files. (I use them to make excel docs for the staff.)
Deanna, nice seeing again! If you want to (and it’s not too confidential) you can send this data to Google Analytics to “support my case”.
In the meantime, I’ve received another answer from them. This time they claim that it make take 24 hours for data to arrive. I explained my case again, focusing on more historical data (more than a week before that initial screenshot) that has changed. I hope that I was more clear this time.
When I get a more serious answer from them, I’ll post it.
Google Analytics: Bounce Visits are excluded for Avg. Time On Site…
Google Analytics stopped counting bounce visits when calculating the average time on site. The new numbers began showing up on July 20th.
I finally received a satisfying answer from the Google Analytics support team, following this post. This explain…
This problem still exists. I don’t believe it has to do with the July update because Time on Site metrics I recorded for multiple clients last month have all DECREASED by about 38% a month later when I look again. This is happening for Time On Site historical data too, all have decreased for ALL my clients. This is a very annoying problem!
Bryan, thanks for sharing your info. I haven’t looked into these metrics lately, and I’m bothered by your report that the numbers have changed again.
I understood that the excluding the bounce visits caused the shift in numbers, but Google has apparently made another change. I’ll check it out…
I just came across this. I guess I am not going nuts.
http://analytics.blogspot.com/2007/09/reverting-back-to-original-average-time.html
Thanks for checking it out and updating me. I’ll update the relevant post.
It’s weird that they announced this revert. Last time it was undeclared.