Originally posted by Dominique
You better test a larger sample of visitors than Dean, because his experience can tip to the opposite tomorrow.The samples are way too small.
I wouldnt be testing anything but hits – EG if my stats showed 185 hits sent then their stats should show 185 hits sent … it wouldnt matter if it was 1 hit or 10000 hits – a hit sent should be recorded as a hit received visa versa …
Sure there may be inconsistencies and/or acceptable deviations but I ASSURE you IMHO the 20% another program stated is unacceptable for loss of tracking.
Heck if your losing 20% of tracking and by chance that’s the 20% of the traffic that’s converting your making nothing! That high percentage to me is an excuse for shaving!
So my proposal would be simple – send X number of hits from various sources and compare the numbers to establish whos accurate with regards to hits and who is not.