When to dump a targeted site?

How many impressions is statistically sig?

[ardent]

I've played around with Adwords often, never with a lot of success. I'm now starting to target sites.

At what point should I drop a site from targeting? For example, there have been several thousand impressions on one site without a click - is this enough impressions to can the site?

[arinick]

I hate to give this response, but it really does depend on the individual conditions. That said, here are some questions to get at more detail as to what I mean:

Is your goal branding or direct traffic? Sometimes impressions can be just as valuable if you are trying to increase brand/product awareness. If your goal is traffic-- it is certainly possible that site match, or image ads for that matter, are not for you. But, some industries certainly enjoy high CTR and excellent conversion with this inventory. So, the bottom line is to test and see.

Do you measure your ROI? If you have a high-ticket item, for example, it may take many impressions/clicks before you see a return. If you know your ROI, there are two real variables to control-- choice of sites and CPM. I generally recommend, therefore, keeping the number of sites in an adgroup to a minimum so that you can see what is working and what is not, and control things at the most micro level possible.

Lastly, and I think most importantly-- visit every site on which your ads are appearing. Remember, if the site isn't following the Adsense principles for maximizing CTR, then you aren't going to get a good CTR. Plain and simple. Many of the sites available bury ads so deep that they will never be clicked-- bottoms of pages, below footers, etc.

Hope this helps somewhat.

[ardent]

Even though you started with "depends on the situation", that was a very helpful response - thanks.

I want to drive traffic (not branding).

I do watch ROI. The ROI on zero clicks is 0, but when do you know you've let the campaign run long enough to be meaningful?

[arinick]

It is difficult to set a hard and fast rule for when to abandon any unproven tactic--- but, here's a way to look at it that may help:

What is your ROI elsewhere? If you know that, then perhaps the best strategy is to stop at some predetermined multiple of the amount typically spent per transaction that seems right for your business conditions (ability to take risk, etc). If you track how many transactions it would take to bring you back to a positive ROI, or your typical ROI, then you will have a better sense of what to expect and when to pull the plug.

I have seen this multiple range from 2x to 50x, so it really does depend on what you are looking to achieve.

[edited by: arinick at 6:12 pm (utc) on Dec. 15, 2005]

[DamonHD]

Hi Ardent,

For me "long enough" is either one week, to allow for the full gamut of changing traffic patterns, or when your testing budget has gone (though I try to make sure that the initial budget will last more than a week).

Rgds

Damon