I recently got involved in an interesting conversation on the Web Analtics Forum on the subject of large dicrepancies between traffic sources reported by Google Analytics (GA) and Coremetrics (CM).
As expected total sessions were very similar with both tools but I was surprised to discover large differences in direct, referral and search traffic:
Total Sessions
GA: 6,853,945
CM: 6,657,176
A/B: 103%
Direct Traffic
GA: 3.131,652
CM: 5,425,392
A/B: 58%
Search Engines
GA: 2,289,280
CM: 959,735
A/B: 239%
Referrals
GA: 1,357,464
CM: 79,774
A/B: 1,702% (yes 17x)
Thanks to nethab1 I discovered that the answer is simple. The mathematical geniuses among you may notice that if you add up all traffic sources for both tools the total number of visits is roughly the same (+/- 0.8%). The discrepancy arises because Google Analytics attributes some visits to Search and Referral that Coremetrics is treating as "Direct". I don't know how Coremetrics operates, but Google Analytics only attributes a visit to "Direct" if that visitor has never come from a non-direct source before.
In other words, if a visitor first visited from a Search engine but later made 9 "Direct" visits, all 10 visits would be attributed to the Search engine. Same with Referral visits. Google Analytics uses a source cookie for search as well as for marketing campaigns and keeps the duration value to the default 6 months (not sure whether this can be changed). It appears that tools like Coremetrics always attribute traffic source as the per the last visit.
More information on how Google Analytics attributes traffic source take a look at Avinash Kaushik's comments here.
Wednesday, August 12, 2009
SEOmoz | Case Study: How Much Do Rankings Matter?
An interesting case study from SEOMoz looking at the relationship between keyword ranking and keyword traffic.
"Over the 7 months in this case study, Google visitors used over 250,000 unique phrases to reach the client's site. Over 80% of those phrases didn't contain any variation of the primary keyword at all. So, while rankings obviously still matter on a keyword-by-keyword basis, being #1 for your top keyword (or even your top few keywords) is no longer good enough – if that's all you're measuring, then you're missing the big picture."
Full case study here
"Over the 7 months in this case study, Google visitors used over 250,000 unique phrases to reach the client's site. Over 80% of those phrases didn't contain any variation of the primary keyword at all. So, while rankings obviously still matter on a keyword-by-keyword basis, being #1 for your top keyword (or even your top few keywords) is no longer good enough – if that's all you're measuring, then you're missing the big picture."
Full case study here
Subscribe to:
Posts (Atom)
