Optimizing Landers – How many views on lower traffic targets?
I’m assuming this will vary a bit from person to person – i’m trying to optimize my LP for a potentially good campaign but the targets aren’t exactly high traffic, getting about 600-800 views/day split between 2LP’s and Directlink, so waiting like 3-4days to collect "significant" (say ~800views/LP) data to compare LP results and make further tweaks to the page to split-test. Given that your not gathering data very quickly would you make calls if you see significant differences early on eg~250views, or would you kill the direct link (which is currently at breakeven without optimization) to drive more traffic to the LP’s for quicker data collection?OR not be impatient and wait 4 days at a time to do a testing round?
love some advice here
A significant difference is a significant difference. Once I get a signal, I go with it.
WITH THAT SAID . . .
The less data you have, the bigger the difference you need between two options to know if it is truly statistically different.
So . . . . if it is the difference between 15 and 17% CTR, you may want to collect more data. On the other hand, if the difference is 4% vs 28%, there’s little benefit in collecting more data.
One thing to remember, it is the number of actionsthat is important in determining when you have enough data, not the percentages. In other words, 18 clicks is statistically better than 3 clicks regardless if those clicks came from 100 or 1000000 views. But, 3 clicks is never statistically better than 2 clicks, regardless of the number of views.
The Article Published IN 08-04-2011 08:02 AM