Executive Director, Multi Channel at House of Fraser
16 September 2009 10:23am
Hello. Please can someone help us to determine what the best length for our cookie expiry is? I have read that 30 days is the best practice but i dont understand why that would be better than say 30 days, 90 days or 365 days.
If a customer only makes a purchase every 60 days then my understanding of analytics means that a 30 day cookie will not give us the ability to understand the relationship between the original purchase and the second browsing session.
Any thoughts on this?
European Web Analytics Specialist at Canon Europe
22 September 2009 11:25am
The limiter to cookie length is the speed at which they are deleted or blocked - long length will lead to skewed data.
HOw about considering how you can track your regular sustomers using a form of cookie and email matching - pretty standard practice for ecommerce...that way you segment between your established and aquisitional traffic...what do you think?
Head of team at Aedgency
22 September 2009 16:28pm
Talking about site analytics, the longer, the better. Provided that it has no prohibitive additional cost to store the data, the more information you have on customer loyalty, the better.
Regarding online advertising, it's the contrary. The longer the cookie window, the more you will pay, if you are on CPA deals such as affiliation or blind networks. It also depends whether you're considering post view activity (as a lot of blind networks default to) or only post click. Definitely cookie window for postview should be very limited.
Yet, if have a unique tracking solution for all your traffic sources, it's always usefull to understand the customer journey, especially when it comes to search patterns (see http://www.rimmkaufman.com/rkgblog/2008/11/05/cookie-windows/).
Director at Yard Associates
23 September 2009 16:53pm
Just to add to the comments already provided, the cookie length for analytics normally refers to a period of inactivity, after which a user is counted as having expired from a marketing point of view. For tracking repeat sales, the cookie length is normally set to some distant date, which is extended on every subsequent visit.
For example, I visited the WebTrends website today, and was tracked using their analytics tool. The cookie expiry date currently set in my browser is 19th September 2019. So, whilst my visits will no longer be attributed to the marketing method I used to enter after 30 days, my transactional history will continue until 2019.
...unless of course I delete my cookies, or buy a new computer in the meantime!
CEO at Econsultancy
29 September 2009 14:41pm
You're correct that 30 days has often been the default, at least for marketing/advertising tracking online. However, I found out recently that Google Analytics' default is set to 60 days.
It's the same with the length of a 'session' or visit which, by most web server standards, is usually set to 30 minutes.
But what is right depends on your business, the way you wish to understand the data, or, just possibly, what you want to make the data say!
- If you want to make it look like you're getting more visits to your site just reduce the timeout value on your webserver and/or web analytics software. Bingo ! More visits. But probably also annoyed customers whose session data evaporated when they came back to your site having just gone to make a cup of tea.
- If you want to pay your affiliates more, or less, then just adjust the cookie value up or down.
- If you want to make various forms of online marketing look more, or less, effective, then just adjust the cookie expiry length up or down. Some might say that because (paid) search has often received the 'last click' (coupon sites now doing well here) then it is in Google's interest to set the expiry length default in GA as long as possible because it will probably make paid search look more effective?
And, of course, with all the above, if you change your expiry length you'll have data / stats that don't quite match up for a while until the changes have worked through and you have new benchmarks to work with.
I guess a lot depends on how you measure value, how you choose to reward others (affiliates etc.) for value creation, and your product/customer base e.g. if you're selling CDs then the consideration/purchase window is *much less* than if you're selling mortgages so your cookie expiry length would, presumably, be much shorter too.
Crudely speaking I guess your cookie expiry length could be as long as the consideration/purchase window for, say, 80% of your customers?
Free market research on digital marketing
Daily Pulse: award winning newsletter
It takes 30 seconds to register