Hi all,
Currently I'm using the rules of thumb in the getting started guide for pausing banners.
@1200 impressions and 0 clicks, pause the banner
@5000 impressions and not at the desired ctr, pause the banner
Because I work a day job, I programmed a script to disable banners for me following these rules.
My question is, should I test each banner at multiple different times? As in, can the ctr fluctuate significantly at vary times of the day?
Another question, is there a flaw in this approach that I'm not seeing?
Don't forget every banner will convert differently, i have had banners with an awfull ctr who were outperforming normal ctr's in conversions.
Btw, sell me the script haha.
What traffic source?
Adult traffic
If those impressions are coming in very quickly, the data becomes less certain. Generally I'd recommend giving a banner at least half a day to prove itself. It's worth not setting your initial campaign budgets too high for just this reason - impressions trickling in over the day give a more reliable overall picture.
I see, would you recommend frequency capping ads at 1 or 2 for testing, opposed to 4-5?
It really depends on the traffic source and type but generally a frequency cap of 3 is a good place to start. For adult I imagine people spam open a lot of pages = lots of repeat impressions and a higher frequency cap may be useful. My conclusion: test it yourself - duplicate a campaign or banner with all else being equal but the frequency cap and see what effect it has.
Is this just a rule you use when launching new campaigns?
I get why pausing a banner if it doesn't draw a click in the first 1200 impressions would be beneficial, but if you applied it to all banners at all times, you'd likely end up with zero banners (or a lot of chat creatives!).
I think any pausing mechanism has to be more sophisticated than measuring CTR alone. Otherwise we'd all be running Facebook popup windows.
Awesome advice, it would explain why absolutely non of my banners last more than a few hours!