Comments on: OK, But How Do I Make A Word Bold?

By: Stanley Wong

Stanley Wong — Fri, 12 May 2006 00:16:12 +0000

You should test your ad copy. Sign up for something like BetterPPC.com, it is software that guides you through building a multivariate test for your ads.

By: HD

HD — Fri, 12 May 2006 00:00:12 +0000

It might be that the competition for the kinds of keywords your're using is so keen and search traffic also so high on those terms that Google Ad Words has the minimum bids pretty well dialed in for what it takes to get your ads served AT ALL, forget about ranking. If so, then you'll not experience what I did, which was a campaign that launched with $.05 minimum bids on some of the search terms (and zero competition for them on Ad Words) and saw them ratcheted upwards to $1.50 over the course of the following four weeks, despite a CTR of 1 - 1.5% on the terms, which according to the Ad Words folks was reasonable (low CTR is analyzed as 'not-relevant' and so you pay a premium for having a 'non-relevant' ad continue to get served).

The combination of inexplicably upward ratcheting minimum bids, coupled with fraudulent clicks, led me to bail from the Ad Words program.

Here's hoping you have a more pleasant experience, and I suspect you might, given that you're not doing something completely off the wall.

By: Rick

Rick — Wed, 10 May 2006 23:07:20 +0000

I always like to 'test' out potential keywords by thinking of every combo or related search term I can, then do like 24-48 hours worth of a test. I'll put a totally unrelated ad (something that would be unclickable) and then put a decent dollar amount so it'll run high. I'm sure the Adwords gods don't this, but it works for a short test and what it would take to be #1.

I also use Google suggest to find potential keywords and possible misspellings. Misspellings can be like $.03 or $.05 per click.

That's my $.02.

By: Rob

Rob — Wed, 10 May 2006 22:12:23 +0000

Yes, most high-end news/analysis sites out there are looking in the $80 region for a rectangle/mpu and slightly less for a Leaderboard sized creative.

By: George

George — Wed, 10 May 2006 17:23:21 +0000

So you really think a little more control on ad placement (which you already have to some extent in adsense/msn adcenter) deserves a $30 CPM for semi-targetted, non-intent traffic?...

Auto Sites are a whole different story because the US big 3's are slowly shifting a large amount of their marketing money online but there is not enough auto releated inventory available thats the reason the CPM is skyhigh and sites like edmunds are sold out for a year. Also remember most of the users visiting those sites are in the buying mode (almost 70% of new car buyers research online before deciding).

By: John Battelle

John Battelle — Wed, 10 May 2006 16:57:28 +0000

George – A couple of things. First, $30 is the high end of what FM ads sell for, and for that you’re getting a much larger an unit, on a specific site with specific guaranteed placement, in a specifc timeframe, with total control. And, you are part of an ongoing conversation with that site, which is not exactly how AdSense works. It’s the equivalent of a sponsorship, of sorts. In any case, $30 is far lower than the CPMs on most other professional sites like the NYT, even Yahoo (in areas like autos and travel).

By: George

George — Wed, 10 May 2006 16:36:27 +0000

John, When you play with adwords and the different keyword bids you can see how unrealistic your CPM bids ($30) in FM really is...Remember search traffic is way qualified than blog traffic.

At this moment you may sell the spots to some clueless fortune 100 company, but i doubt you can sustain this price in the long term.

By: King Troll China Boi

King Troll China Boi — Tue, 09 May 2006 18:16:58 +0000

I was wondering how you looked and I did an image search. You are a dork dude. I did think you were some oild long haired fat guy at first though, so glad to see you are not.

F#$k you to Johnny , F@#k you too.

http://www.theregister.co.uk/2005/10/06/battelle.jpg

By: rick gregory

rick gregory — Tue, 09 May 2006 18:08:03 +0000

John,

There are a couple of other things I'd suggest.

First, give all of your ads tracking paramters in the URLs! Now!

From a very quick look, it seems that each of the URLs you have out there simply points to a page, but this means you can't easily differentiate between the traffic from one word vs another. That, in turn, makes it hard to calculate ROI, etc.

This doesn't have to be complex - make up a parameter, then iterate it and append it to the end of the URL. For example, put ?src= at the end of each URL and just have the part be an incrementing number, i.e. ?src=0001. Associate these tracking parameters with the words in a spreadsheet.

Second, you can have multiple pieces of creative associated with one keyword. Google will rotate these evenly until it determines that one is doing better (has higher CTR) at ehcih point it will start to favor that one. This is great for testing messaging, etc. For these, give each keyword/creative combo unique tracking parameters.

Finally, think about campaign and adgroup structure. You can move words between campaigns, but if you have enough words, you may want to track spend and ROI for one set vs another.

By: Kate

Kate — Tue, 09 May 2006 17:40:16 +0000

John, be careful though. In some keywords, bolding has the opposite effect. If every title is bolded, people will go for the unbolded ones. Think Monster.com and the bolded listings there. It’s more money, but for some people the bolding deters them.