Cinderella.jpg

The Cinderella Visitor

 

It’s surprising, every time I open Analytics, I learn something new about this product and I learn something new about my site … like the fact that I have to deal with Cinderella Visitors!

Here is a maths problem for you. In fact if you have pored through your Google Analytics account, I am sure even you would also have been stumped by the same problem.

Below is a screenshot of my site’s monthly traffic stats –

Monthly Stats

Now, yes … these are humble beginnings! That is why some of you should visit this site more often!! So if you look at the unique visitors (that’s the total number of visitors coming to the website), then it is 530. This is good, that means through the blog I am getting in touch with at least 530 people a month!! Holy cow!! That’s roughly 18 conversations in the day.

Out of those 530, a huge number are new visitors (511). So does that mean only 19 visitors are returning back to the blog for those 126 visits? Hmmm … that should not be that hard to find. One look at the Returning Visitors Advanced Segment should shed more light on this …

Returning Visitors

Well, this seems to the problem don’t it. Things don’t just add up here! 45 visitors are responsible for those 126 visits, which means that 485 visitors are responsible for those 511 New Visits. How can this be possible? These 485 people have come on the site ONCE ( … sigh!! aee jaanewaale ho sake to laut ke aajaa …), but the number of visits tell me otherwise.

Which should bring us to the real question of what does Google Analytics define as a visit? Here is an insightful article from Google about How Visits are Calculated in Analytics. There we have our answer –

A single visitor can open multiple visits. Those visits can occur on the same day, or over several days, weeks, or months. As soon as one visit ends, there is then an opportunity to start a new visit. There are two methods by which a visit ends:

  • Time-based expiry (including end of day):
    • After 30 minutes of inactivity by the visitor
    • At midnight

Ahhh! So if the visitor comes to the site, sticks around for more than 30 minutes without doing anything and then clicks on any other links, then its considered as ANOTHER visit. This should have been a sufficient reason IF the average visit duration was on the higher side, but since its not the case (and I write reasonable short posts too!), so the other method seems to be in action here!

At least 26 visitors have been around the site around midnight!! A quick check on my GA for hourly visits confirms my suspicions! I had 26 visitors coming to the website around midnight … and as the clock struck twelve, these Cinderella visitors fled (or at least Google Analytics made them fade away) and came back as New Visits!! By the way, in case if you are wondering why have you not heard about this term Cinderella Visitor before, then do not worry … I just coined it!

So where are the glass slippers? The glass slippers are the ubiquitous utmz cookies!!

Prasad Ajinkya

An epicurean who has read a Rand too many

3 comments
Vineet Jawa
Vineet Jawa

hahah - insightful! Had a similar case yesterday. Under engagement, GA classifies any < 1 page depth visitor as 0-10 sec bracket. So if you go to the homepage of the site and spend 2 minutes... you're still classified as only 0-10 secs... Other sore point is how GA and adwords numbers never match

Harshaja
Harshaja

Very informative thumbs up!

Prasad
Prasad

Vineet, the sore point you say aint that bad. GA measures visits v/s Adwords measures clicks. Any click within a 30 minute gap (in the same day) will be counted within the same visit. However if a user comes to the same site via multiple campaigns within the 30 minute period, then they are counted as separate visits. That should explain the discrepancy. What we do is create a dashboard by pulling data from both GA and Adwords, I had written about it some time back on how Subtotals in Excel can make such dashboards more awesome! Here is the link - http://kidakaka.com/blog/2012/12/07/subtotal-in-excel/