A common question asked by marketers today – Why is there a big difference between Adwords conversion revenue and Google Analytics Paid Search conversion revenue?

Below are a few reasons why there will be a difference.

Google Adwords:
Google Adwords has a cookie expiration of 30 days.
Google Adwords matches conversions to clicks.
Google Adwords campaign attribution cannot be overwritten by other traffic sources.

Google Analytics:
Google Analytics, by default, has a campaign cookie expiration of 6 months.
Google Analytics matches conversions to visits.
Google Analytics campaign attribution is overwritten by the most recent campaign visit (by default), excl. direct traffic.

Scenario:
A visitor clicks on a Google Paid Search text ad on 19 July and does not make a purchase. The same visitor, using the same PC and browser returns to the website 2 days later by typing the website address in their browser and makes a purchase.

Outcome:
Google Adwords: Records the conversion at click-date, 19 July.
Google Analytics: Records the conversion in the visit which resulted in the transaction, 21 July.
* Even if Google Adwords is linked to Google Analytics, Google Analytics will assign the conversion to the date of the transaction (the visit in which the purchase occurred), not the original date the ad was clicked.

Scenario:
A visitor clicks on a Google Paid Search text ad on 15 June and does not make a purchase. The same visitor, using the same PC and browser returns to the website 35 days later by typing the website address in their browser and makes a purchase.

Outcome:
Google Adwords: Does not record the conversion, as the conversion occurred after the cookie expired.
Google Analytics: Records the conversion in the visit which resulted in the transaction, 19 July. (35 days after 15 June)

Notes: Adwords has a cookie expiration of 30 days. Google Analytics, by default, has a campaign cookie expiration of 6 months. This can be amended, using the _setCampaignCookieTimeout() method. More information relating to changing the length of the cookie here: http://code.google.com/apis/analytics/docs/gaJS/gaJSApiCampaignTracking.html#_gat.GA_Tracker_._setCampaignCookieTimeout

Scenario:
A visitor clicks on a Google Paid Search text ad on 19 July and does not make a purchase. The same visitor, using the same PC and browser returns to the website 2 days later, after clicking on an email, which is tagged with _utm tags.
(Utm Tags: http://www.google.com/support/analytics/bin/answer.py?answer=55518)

Outcome:
Google Adwords: Records the transaction on 19 July (date of click), as the conversion occurred during the 30 day cookie period.
Google Analytics: Records the transaction, but it is not attributed to Adwords. By default, the most recent campaign (email campaign) receives credit for the conversion, although this can be amended.

Understanding the differences between Google Adwords and Google Analytics will help you understand why each platform is reporting different revenue.

Here is a simple Google Analytics checklist to help you to ensure you are using all features.
Please let me know what features might be missing, or any comments you may have. Thank you

1. List all URLs / Sub-domains that require tracking.
2. List who requires access to profiles and their access permission. (Standard/Administrator)
3. Look at the website – Ask yourself:
   1. What is the purpose of this website?
   2. Identify Micro conversions / goals. (Track as much as possible)
   3. Which metrics are going to be important? Which metrics will help me identify where I can improve my website?
4. Decide the naming conversion for campaign tagging. (Non Adwords campaigns)
1. Use standard naming conventions.
5. Setup:
1. Ensure an unfiltered raw profile exists.
2. Setup asynchronous version of Google Analytics code to be included on all pages.
• Ensure the correct code is used for sub domains (SetDomainName, SetAllowLinker etc.)
3. Setup Site Search. [include category parameter, if possible]
4. Setup Goals.
• Give the goals a value.
• Setup a Goal funnel.
5. Setup E-commerce tracking for purchased items.
6. Do reports need to be scheduled? Setup report schedules.
7. Do events need to be tracked? Track email links / Video buttons.
8. Setup events/Virtual Page views if applicable. E.g Ajax sites
9. Implement filters on profiles. (Except on the raw profile) e.g. exclude company IP.
10. Link Adwords to Analytics.
• Enable Auto tagging.
11. Setup custom alerts to be notified if abnormal activity occurs. [Intelligence]
12. Implement Custom variables. (If applicable) e.g. member status.
13. Advanced Segments: Once you have enough click-stream data, apply Advanced Segments to help understand your data better.
6. Test:
   1. Purchase more than 1 item to ensure E-commerce code renders correct.
   • Ensure the transaction appears within Google Analytics. (You may have to wait 24 hours)
   2. Check that the new Asynchronous code is on all pages. (Keep a close eye on the profile ID within the code)

Yesterday, I started reading Web Analytics 2.0 written by Avinash. The content in this book is outstanding. It has already started opening by eyes and is beginning to mould the way I think about web analytics.

One thing that I found quite funny in the early stages of the book (reason for this blog post) was the reference to Hippos and Customers. I thought I had to blog about this! and upload a picture of a hippo
I guess it was funny as I have experienced it often … Hippo – the Highest Paid Persons Opinion…counts the most, but it could be ruining your site/conversion rate completely!

Well, I’m really looking forward to finishing your book, Avinash. Love it!

Happy Holidays everyone! Big challenges in 2011!

Too often, I find that bounce rate is incorrectly understood. Comments like, “Look, the bounce rate is so high, it must be a bad website”, or even worse “The traffic you are sending to this website is not targeted”. Yes, both is possible, but I always recommend digging deeper with the information that is available to you. Take a closer look.

One of the reasons for this post is that I came across a lead generation website the other day, which uses AJAX to “thank” the user when a successful lead is submitted. This means that the page does not reload and the URL remains the same. Therefore, visitors that are completing the form (the ultimate goal!) are contributing to a higher bounce rate – this is obviously not ideal and would affect your bounce rate metric. Any interpretations you have made on the bounce rate are therefore wrong? Yes.

To fix this, you can use Google Analytics “fake page views” or event tracking or custom variables to capture the lead submission/conv. All credit to http://www.twitter.com/avinash Credit to Avinash for explaining the solution in one sentence. He is good!

I confess, I have a lot to learn with Bounce Rate and gaining a deeper understanding on this metric. My current rule of thumb, is to always ask “Why / Where” are visitors bouncing and always check your web analytics setup to ensure it’s gathering the correct data.

I work quite a bit with Google Analytics and have recently come across two “keywords” that are generated by Google Analytics, that are not actually keywords.
{ysmkey} is a keyword that is created by Google Analytics for a Yahoo Paid campaign, where it was unable to get the actual keyword that was clicked on that attributed the site visit.

The Other keyword that I came across is “boomuserlist”. keywords that start with boomuserlist are visits that belong to a Remarketing campaign within Google Adwords. As keywords do not need to be specified when creating a Remarketing campaign in Google Adwords, Google Analytics attributes all visits/revenue to “boomuserlist”, which I think has to do with the list/audience to which the visitors belong to, as defined in Google Adwords.

It was a Sunday Afternoon and I was bored, after a fantastic weekend. So I decided to create www.GoogleAnalyticsLessons.com. I did not want to create something big and I had around half an hour spare time, so I quickly developed it, registered the domain name and added a few “lessons”. Check it out: www.GoogleAnalyticsLessons.com .. I know – This is probably one of the biggest ‘geek’ things I have ever done

I had a thought the other day – I believe it is possible to see a higher bounce rate and (at the same time), a higher e-Commerce conversion rate…even with lots of visit data.

It may not make sense – I higher bounce rate so more people are leaving my site immediately, so how can my conversion rate is possible to be higher.. Well, I believe it’s possible.

Anyone agree or disagree or have any thoughts?

Every day I read about Analytics for at least 30 minutes, and although I love Google Analytics, I still think that its biggest downfall is campaign attribution which other paid enterprise analytics solutions offer.

To only see the first or last click that lead up to a conversion is not ideal, whereas with other paid solutions, it is possible to capture both clicks and even the clicks in between which might help you allocate funds to your campaigns differently and understand the conversion process better.

I really hope that Google Analytics add this as a feature.

Do you add Google Analytics to your site, just to tick it of the checklist? “It’s a cool product, everyone uses it, hey its free, I might as well add it to my site.” Cool, but are you using it properly?… These 5 things below certainly don’t prove it’s power(I’ll write about that sometime), but for now, here are some common mistakes…even amongst large successful eRetailers.

Last Click Attribution. Beware. Use Nooverride to switch attribution from last campaign back to first campaign.

By Default Google Analytics attributes all conversions to the last click. So if you running an Adwords campaign with converion tracking (a Google product!) as well as Analytics (also a Google product!) and numbers differ, it may not only be because of clicks vs. visits, but also because, by default, Google Analytics attributes
the conversion to the last click…so if a banner on another marketing channel was clicked with UTM tags, or an email newsletter link, with UTM tags was clicked and a conversion is made after the user clicked on an adwords ad, Analytics will attribute the conversion to your the most recent click. It is however possible to allow Google Analytics not to attribute the conversion to the last click, simply by add nooverride=true to the landing page URL.

Tracking Subdomains – such a common mistake.
While reviewing many Analytics accounts, I find that the most common mistake is tracking sub-domains. if your website has subdomains like search.domainname.com or chat.domainname.com or checkout.domainname.com (I have seen this), then ensure Google Analytics is aware that these subdomains are part of your normal website flow and should be tracked within the same profile.

Why track your own traffic. Use Filters
Why do so many analytics profiles not include a filter to exclude their own traffic? Surely you do not want to measure your own visits? I recommended keeping a raw profile and creating another profile with an added filter. Filtering by IP would probably work well, especially if all users in your office share the same static IP address.

SiteSearch
This is so easy to setup and usually (98% of the time) requires absolutely no web development changes. Tracking Site Search can be realy useful. Why? Well, if you running any campaign, you can segment your data by that campaign and see what users are searching for when they land on your site. Is what they are searching for aligned with your campaign? Perhaps you running a PPC campaign and many users are specifically searching for the same keyword that you do not have in your campaign. If it is a relevant keyword, then why not add the keyword as an exact match keyword and deep link to the most relevant page? Refine your campaign in to increase user intent? Optimise your SEO?

TimeZone
Please tell me the timezone in your Analytics profile is correct. I’ve come across this (more than once) that the timezone of an Analytics profile is incorrect…If you looking at data that shows hours of the day that receives most visits or sales, then you looking at incorrect data. Your time-zone can be seen under Settings.

There’s a lot more I would like to cover; Goals, Funnels, Segmentation, Custom Varibles (ever heard of them?), Annotations. I’ll write about this in another post.

Cheers,
Lloyd

Well today I put event tracking on a site using the new Google Analytics code.
It’s done like this:

onClick="_gaq.push(['_trackEvent', 'name', 1]);"