| 5:20 am on Jan 16, 2008 (gmt 0)|
google also recommend using broad match for all your terms. ;-)
You should test it for yourself and see how you do. Maybe remove a lot of your long tail from your existing campaign and measure the results.
| 11:04 pm on Jan 16, 2008 (gmt 0)|
|Does this basically mean there is no point in having an ad group with 1000 long tail keywords in it? Just stick in the Top 30 keywords and be done with it? |
Assuming that you are talking about campaigns targeted to the content network, then my vote goes right along with the Inside AdWords blog team.
Except that I might even say 20 or fewer very tightly targeted keywords, all on the theme as the type of page on which you'd like your ad to appear. ;)
|[...] You should test it for yourself and see how you do. [...] |
Always a very worthy idea.
| 1:11 am on Jan 17, 2008 (gmt 0)|
Well my ad groups are always "tightly themed" but they can still have tons of keywords. For example if I had an ad group called "work at home" I could easily have 500-1000 keyword phrases in that ad group, all containing the base phrase "work at home". The theme of the ad group would obviously be "work at home" so basically just at the top 20 or so keywords and that's it?
I guess what I'm trying to do is just confirm how the system works. Is 15-30 keywords all that's necessary for the system to understand what kind of pages you want your ad shown on, or is that just the general recommendation because they were trying to reinforce the fact that you should have tightly themed ad groups, and they assume most people aren't able to have a tightly themed ad group with 500 keywords in it (maybe true).
| 8:27 pm on Feb 10, 2008 (gmt 0)|
I too wish we could get a definitive answer to that.
| 5:46 am on Feb 11, 2008 (gmt 0)|
My understanding is that the content network categorizes your keywords. IE the system reviews all your keywords and then assigns it a category to display ads for when it is run.
If you run too many ads you run the risk of it not hitting the right category. I would assume you just need enough keywords to make sure Google clearly knows what you mean and where to place you.
It would actually be useful if we could see the category a content network ad is run in though.
So if you want to show up on sites that are categorized as relating to "red apples", you should probably stick with like "red apples, buying red apples, eating red apples." Probably avoid things like "red and golden delicious apples", "red apple cider" etc.
| 11:07 am on Feb 11, 2008 (gmt 0)|
But what about "i love eating red apples in the morning"?
Let's say there is a website somewhere with that phrase and an Adsense block. Would adding a phrase like that help my ad get on that page? That's the main question here. And I'm afraid only the Google employees would be able to answer it.
| 5:15 pm on Feb 11, 2008 (gmt 0)|
I am testing three adgroups in content. One is based on the google recommendations to mix and match keywords (flowers fits the same theme as florist, etc.). One is based on just themed keywords (red widget, fast widget, buy widget, etc.) and one is a combination of my old approach (just themed keywords) and Google's approach (combined keywords). From an impression standpoint, I'm getting the most impressions from the mixed adgroup with the second most impressions coming from my traditional adgroup (themed keywords).
I'm still analyzing my placement reports to see how differently Google distributes the ads but its not by category...it goes deeper than that...to individual sites. In other words, my three test adgroups are not necessarily showing ads on the same domains and URLs even though the theme and ads are identical and the only difference is the slight variations for keywords that I used in each group.
| 9:45 pm on Feb 11, 2008 (gmt 0)|
justshelley, I would interested to hear how your tests turn out. For myself, some content campaigns have had great success with many keywords while others have had great success with just a few keywords. I'm not coming up with anything definitive at this point, but leaning towards a combination of 'what's easiest' plus 'what google recommends'.
| 2:54 am on Feb 12, 2008 (gmt 0)|
Are you running these content ad groups simultaneously and from the same account? If so, and this is just a thought, ads from two ad groups will not display at the same time. In other words, two or even all of the ad groups could theoretically display an ad on that site or URL, but won't because only one of your ads can display at a time. Google may simply choose an ad with a higher CTR or some other factor of their algo. IMO you should test these and run only one at a time to get a true test.
Do you have the same number of keywords in each of your ad groups? or are you running 500-1000 keywords in one and 15-30 in another? Doesn't seem it would be as accurate if your only measuring impressions.
Also, are you looking for impressions, CTR, and/or conversions? In my situation, I look for conversions so that would be the metric to measure not the number of impressions. Obviously it would depend on your situation not mine. If you want more impressions and more clicks (more cost) then you may be on the right track, but if your looking for a better conversion rate or most profitable then excluding sites and/or categories to get the right mix may be the way to go.
I'm curious to hear about your findings as well. Please post results.
| 2:58 am on Feb 12, 2008 (gmt 0)|
Do you run and monitor your content campaign separately from a Google/network campaign?
| 8:38 pm on Feb 13, 2008 (gmt 0)|
I thought about the competition factor but I've had some of the test groups take off and over-power the old groups within a couple of days so I don't think that's an issue.
I'm not using the same number of keywords in each group and I'm using a lot less than when I originally started working with content a while back. ;) I have between 30 and 50 keywords in my test groups.
Right now, I'm looking at impressions and placements. Conversions are important but so is figuring out how to get my ads placed on the types of content sites that I want them to be seen on and that's what I'm working on right now.
But related to conversions...I have other test concepts running too. Instead of doing adgroups for tips (widget company tips, widget business tips), savings (widget company savings, widget business savings), strategies (widget company strategies, widget business strategies). I set up new adgroups for company (widget company tips, widget company savings, widget company strategies), business, etc. and those adgroups are converting MUCH better than my old adgroups (tips, savings, strategies).
| 9:08 pm on Feb 13, 2008 (gmt 0)|
|Conversions are important but so is figuring out how to get my ads placed on the types of content sites that I want them to be seen on and that's what I'm working on right now. |
Have you tried to use the content site targeting feature? You can actually pick the sites where your ads display and they won't display elsewhere. Of course you may have to research which sites you want to include (Google rep could even help with this), but your doing that now with your test campaigns. Ultimately you could run both campaings side-by-side to continue your testing.
| 6:06 pm on Feb 15, 2008 (gmt 0)|
I was actually in the middle of a test where I was taking converting sites from my placement reports from content and setting them up in site tarteging...but I lost that client before I could finish the test. Haven't had the opportunity to set it up again. I never got to really dig into the data...conversions were great but it seemed like the cost was higher in site targeting. I'd really like to see if anyone else has tested that same way because I'd like to know what kind of results they got.
I had several test scenarios running. We actually combed our campaigns for converting sites and set up identical site targeted adgroups using only converting sites. At first we actually excluded the converting sites from the regular content network so we could control them in site targeting but I didn't like how that was working so I was in the process of putting them back in both places when I lost the client.
I also want to try something I read...take my high conversion sites from content and set up individual site targeted adgroups running just one high converting site in each adgroup and then testing ads just targeted at the audience for that site. Anyone tried that?
| 6:20 pm on Feb 15, 2008 (gmt 0)|
justshelley, keep in mind that site targeting is often more expensive per click than the content network. Often times your effective CPM has to be higher than the combined eCPM of various advertisers you are replacing. If a site-targeted ad is replacing 4 content ads, you won't show unless your eCPM is higher than all those 4 combined.
| 1:38 pm on Feb 16, 2008 (gmt 0)|
Someone told me that about the higher cost of site targeting but I've been trying to find information on it online. Do you know of any links that talk about the higher cost of site targeting?
| 7:09 am on Feb 18, 2008 (gmt 0)|
Googles recomendations suit Google.
Your goals and their goals are very different, always remember that.
| 4:08 pm on Feb 18, 2008 (gmt 0)|
They may have changed the wording on this:
The key paragraph is the following:
"Keyword-targeted text ads often appear with other ads in one ad unit. If a placement-targeted ad's Ad Rank is higher than that of all the competing keyword-targeted ads, it will be the only ad to appear on that ad unit."
I've had to bid a lot higher to have ads show on placement targeted campaigns. My ads show on those same sites for much lower bids with keyword content targeting campaigns.
| 2:38 pm on Feb 20, 2008 (gmt 0)|
The big question is...on traditional content marketing, how do you make SURE your ads continue showing on that one site that you want to target? I haven't figured that out yet...if anyone has any tips for making sure that your ads show up on "specific" sites in the traditional content targeting, please let me know how you did it!
| 3:54 pm on Feb 20, 2008 (gmt 0)|
Google's likely answer: that's what placement targeting is for. Unfortunately, that costs you more money.
I'm fairly sure there is no way to ensure you appear on a specific site using the traditional content network.
I just wish there was an option in placement targeting for you to select if you want to replace traditional content ads or appear along with them in the same ad unit.
| 8:18 pm on Feb 20, 2008 (gmt 0)|
I tried moving all my converting placements over to placement targeting and then excluding those same placements from content to keep them from overlapping. Even though my conversions were good in placement targeting...they were never as high as they were when I left those same placements alone in content.
I guess one strategy for taking advantage of the lower costs per click and conversion in regular content and for controlling which sites the ads are shown on would be to just keep excluding every site that pops up for a specific adgroup except for the site or sites that I want my ads to be seen on.
| 8:31 pm on Feb 20, 2008 (gmt 0)|
I would imagine that would work, but be a management nightmare. You would have to constantly exclude sites.