A reader asks: "If I want to improve my click through rate of underperforming keywords in a certain adgroup by peeling those keywords and sticking them into a new adgroup… should I delete or pause the low CTR keywords in the old adgroup?  How do I do it so my quality score isn’t affected?"

I write back: "What are the quality scores of the underperforming keywords right now?"

Reader responds that one keyword is great, others are OK or Poor.

I procrastinate for four days by playing with Google Desktop for Mac and rereading the amazon reviews for AdWords For Dummies (hey, there’s two new ones, and I don’t even know these people!) before sitting down to respond.

My eventual response:

You want your quality score to be affected, if it isn’t great. One of the reasons to peel and stick is to improve keyword quality by increasing relevance between keyword, ad and landing page. I would probably delete the offending keyword, and recreate it in a new ad group, either in the same or a different campaign.

If Google tells you to delete it, then it’s probably a good idea not to tick off Mama Google - just do what she says, until she turns her attention elsewhere. Then sneak the keyword into a better ad group. (That’s actually what Mama Google wants, anyway.)

How do you know what Mama Google wants? She now tells you, with a little icon right next to your quality score in the keywords tab. Mouse over it to see what Google thinks:

Click the "Details and recommendations >>" link to bring up, amazingly enough, details and recommendations:

Read the rest of this entry »

I didn’t feel like doing any real work today, so I created an AdWords For Dummies checklist. it consists of 36 Things You Should Be Doing But Probably Aren’t (we in the consulting profession like to refer to them as "best practices" and we often charge a lot of money to rub your nose in them).

But as I said, I’m a little too tired to do any nose-rubbing today, so you get to have them for free.

The goal of this checklist, as with most free checklists, is to make you feel really stupid and helpless, so you’ll call me to fix everything for you. Well, forget it. I’m tired, remember?

And do you know why I’m tired? Because I spent 8 months writing the damn Dummies book, that’s why.

So if you get inspired, motivated, or otherwise turned on by the checklist, do yourself a favor and buy AdWords For Dummies. It’s not tired. It’s perky, in fact. Ready to party. It’s wearing a green velvet smoking jacket, a jaunty blue fedora, and a faux-ivory tipped faux-ebony cane.

Here’s the AdWords For Dummies Official Checklist, handsomely converted to PDF for your checking-off pleasure.

Update: The Checklist, all updated and shiny for late 2008, is now included as a bonus with the Look Over My Shoulder (LOMS) AdWords Success Videos.

A lot of campaigns I see combine Google search, partner search, and the content network. That’s almost always a bad idea. Here’s why:

1. Different networks require different ads

Ads on the search networks appear on search results pages. Your prospects have actively typed a keyword or phrase into a search engine. They are looking for something, and are doing so intentionally.

By contrast, ads on the content network appear next to web content, like blog posts or newpaper articles. If you want someone to click your ad, you have to interrupt their train of thought, rather than just continue it.

Rarely will the same ad perform optimally in both situations.

2. The content network ads change slower

When you create a new ad or edit an old one, Google assumes that ad is fine and starts showing it within minutes - on its own networks. If the ad doesn’t conform to Google’s editorial policy, it will be disabled later.

On the search partners and content network, however, Google checks the ads before serving them outside of GoogleWorld. That could take up to a week.

So if you mix the content, search partners and Google search networks, the results of split tests will be unclear. You won’t be able to tell when the new or edited ads started showing on which network.

A question from a subscriber in the speaking/info-products business:

"How do I manage multiple websites (multiple products) and link them together in the most strategic - and profitable/productive - way?"

In the old days (circa Spring 2006), having lots of mini-sites was a good strategy. Register the domain names with different registrars, host them on different servers, link them all to each other, and keep churning them out to create your own happy little web-ring.

Trouble is, mini-sites tend to get the cold shoulder from Google these days. Too much salesy-ness, not enough content, and virtually no "authority."

The Google Slap of summer 2006, along with subsequent Slaps that continue to this day, penalized non-relevant and non-authoritative sites by increasing their minimum bids to up to 100 times the bids of the best sites. Instead of a dime a click, folks were paying ten bucks.

You can keep your mini-sites, but don’t use them for Google landing pages. Use them for two other functions:

1. Testing URLs

If you have product-specific URLs, they might convert better than your main site. For example, www.free-lead-generation-course.com is more descriptive and benefit-laden than askhowie.com). You can use the mini-site URL as long as you forward the page to your "authority site."

2. Experimenting with conversion from traffic sources other than search

You may want to try things that Google doesn’t reward or allow, like pop-ups, flash sites, and other things. Remember, if you send offline traffic to your site, you can do whatever you like on it.

But if you intend to drive search engine traffic (paid and/or organic), then spend the time to turn your main site into an authoritative source of information on your topic. Remember that inbound links are the things that make Google go "ahhh," and create a site that other Web sites will link to as a service to their readers.

Minimum Cost Per Click (CPC) is one of the crucial AdWords metrics. How much you are required to pay for a single click on a keyword can determine if you can afford to advertise on that keyword. If you can get your average CPC lower than your competitors, you have a small advantage that you can easily turn into a market-dominating one.

Four ways to lower your average CPC:

1. Find "off the beaten track" keywords with less competition
2. Use quotes and brackets to include phrase and exact match keywords (see Chapter 5 of AdWords For Dummies for details)
3. Bid lower to accept an inferior position and/or fewer impressions
4. Improve the Quality Score (QS) of your keywords

The fourth option - lowering the QS - will change the CPC of an individual keyword and often improves its position. And the QS typically has the greatest impact on CPC - a Great keyword may require only a nickel, whereas the same keyword with a Poor score may demand five bucks to ante up.

But Google doesn’t recalibrate the new QS instantly. So you improve your ad and your landing page, and yet your keywords muddle along with exactly the same CPC as before for up to a couple of days. That’s painfully slow feedback.

The Presto Change-o Solution

Delete and re-enter the keyword, and Google will instantly recalculate the minimum CPC.

If you keep track of the date when you delete and re-enter the keyword, you can compare the performance of the "new" keyword and the "old" one (I know, they’re exactly the same, but don’t tell Google) to see what difference an improved quality score can make.

Google likes its advertisers to show their better-performing ads more often. If you never venture into the Campaign Settings area and change it, Google will "sunset"  your low CTR ads by default.

In AdWords For Dummies I offer several compelling reasons to move the radio button from "Optimize" to "Rotate" (to show the ads more evenly, so you can more quickly and accurately determine the results of your split tests).

In a perfect world, you could choose "Rotate" and know that each ad was getting equal air time - over a period of 7 days, for example, Ad 1 and Ad 2 would both get roughly 6500 impressions each. As long as the impressions are about equal, the split tests are valid and the ultimate measure of an ad’s effectiveness - profit per 1000 impressions - is valid as well.

Quality Score: a wrench in the works

Trouble is, Google’s quality score metric gets in the way here. Different ads with different quality scores get shown in different positions - ads with poor quality may get shunted to page 2 or 3, or not get shown at all. So your scientific split test to find out which ad is favored by prospects, or which one makes you the most money, is now seriously skewed if Google determines the two ads are of different quality.

To make matters worse, Google doesn’t tell you the quality of an individual ad. Quality scores are attached to keywords only, and the quality of a keyword is the same as the quality of that keyword in relation to the worst ad. In other words, if the keyword "inflatable cordless phone" triggered one ad with the headline "Inflatable cordless phone" and another with the headline "Give your dog viagra", then the quality score for that keyword would be "Poor" because of the poor fit with the second ad.

The Tony the Tiger Solution

If all your quality scores are "Great," that means all your ads in that ad group are "Great." Like blue eyes, baldness, and susceptibility to poison ivy, "Great" is a recessive trait. They all have to be great for your keyword to be great.

And it’s pretty easy to make your ads highly relevant to your keywords. Use the keyword phrase a couple of times in the ad, preferably the headline, maybe even in the URL.

But of course, if your landing pages aren’t also highly relevant and part of high quality Web sites, you still won’t know if your ads are of the same quailty. So like much of AdWords, and advertising, and business, and life, you need to feed, weed, and water the entire system. If your ads aren’t getting an equal number of impressions despite chosing "Rotate," then your first task is to get the quality scores of all your keywords to Great.

Here’s how:

1. Peel and stick to create tightly focused keyword lists
2. Match keyword lists to highly relevant ads
3. Match ad groups to highly relevant landing pages