Amazing PPC Tactics
Good afternoon from the Big Apple. Ready for an amazing session? Amazing PPC Tactics is a safe bet, wouldn’t you say? Roll call:
Moderator: Matt Van Wagner, President, Find Me Faster
Addie Conner, Director of Search Marketing, Course Advisor Inc.
Brad Geddes, Founder, bg Theory, LLC
Dan Soha, CEO, Five Mill, Inc.
David Szetela, CEO, Clix Marketing
Brad Geddes’s up first and he’s going to look at location targeting, specifically for national companies. He asks the audience who advertises across the country. A majority. Do you write ads tailored to the location? Just a few hands go up. This presentation is for those with their hands up for the first question but not the second.
Have you heard of Coca Cola Light? You can’t get it in the U.S., where we call it Diet Coke. Meanwhile in Japan they sell diet beer, and no one would buy that in the U.S. There’s a city, state and country level of cultural differences.
“Convenient locations near you” doesn’t speak to people in rural Pennsylvania but speaks well to NJ consumers. A Geographic Report in Google AdWords helps you see this. See the metro areas, the ad spend and the conversions.
Google Analytics Map Overlay helps you drill down into goals per location. Google Trends is a good place to look for info on local targeting. An area that searches most for a product often has a higher conversion rate for that product. When you have a limited ad budget, target just the areas where you have high response rates and create campaigns based on those geographies.
About 1 in 3 people live in top-20 metros. If you have a limited budget and want to reach a broad audience, increase your budgets in the areas where you have high ROI. Or, find promising locations with low conversion rates and re-tailor your ads for those markets.
Dan Soha steps up.
Did you know:
- With “broad-phrase” match, the order of the words does affect your results.
- Headlines are usually 25 characters but 28 character headlines can be created with dynamic keyword insertion.
- Google always gives you one chance. When you’re advertising with a domain that Google has not seen before, your ad will always get a chance. Google will assume your ads perform at the CTR of the average ad on the page. Even if your ad is malformed, no matter what your landing page, you’ll get a chance even if you use a new account, new user, new credit card, etc.
- The Quality Index is at the Adgroup level. You’ll benefit from highly targeted adgroups. Keywords should be similar to each other as well as the ad.
- There is no broad match. Advanced match is not the same as broad match. The more keywords the better.
Divide and conquer:
- Analyze keywords’ CTR performance in an adgroup. Note the ones with high CTR relative to position and ones with low CTR relative to position.
- Pause all low CTR keywords.
- Duplicate adgroup.
- In the new adgroup, delete the active keywords and unpause the remainder.
- The result is a higher CTR adgroup. If done properly the Quality Index of this campaign will increase.
- Hopefully you can find the common denominator between these keywords and write/find an ad that is more appropriate. Analyze the search landscape on these keywords and you’ll probably find commonalities.
Yahoo! Dynamic Keyword Insertion
Why would you want to use DKI? DKI will increase Quality Index because Yahoo! gives extra credit to DKI. With two exact same ads where one has DKI and the other doesn’t, the one with DKI will have a higher Quality Index. He gives us several examples of this at work.
And next is David Szetela. He’ll be talking about winning at PPC content advertising. He asks who in the audience ever advertised on Google’s content network, if they’re still advertising on Google’s content network, and if they’re making good money advertising on Google’s content network. Hands went up high for the first question and gradually decreased along with the questions.
While Google’s content network has not traditionally performed that well, he says content advertising doesn’t actually suck.
Why care about contextual advertising?
- Growing faster than search click inventory.
- Clicks are cheaper.
- Faster/easier to create successful campaigns.
Google’s ad network just doubled in sized and over the last few years, most advertisers have seen their CPC for the program go up. While the cost of paid search advertising is going up, it’s staying steady for contextual advertising.
Why content advertisers lose money:
- Ads appear on irrelevant pages and get bad clicks (low conversion rates).
- Ads don’t distract attention from site content.
- Search and content should never exist in the same campaign.
Contextual is not search:
- Readers are not searching for you.
- More like banner or print advertsing.
- The first job of your ad is to distract.
Three AdWords content flavors:
- Placement targeted
- Enhanced (combination of first two)
- A keyword-targeted ad group’s keywords should describe the kinds of pages where you want your ads to appear.
- Keyword list = the words that appear most frequently on such pages.
- Keyword list need not — and frequently should not — include names of your products/services.
- No more than 2 to 10 keywords per ad group
- Match types are irrelevant (except negative)
- Individual keyword bids are irrelevant
- Negative keywords are almost the same
Google Placement Performance Report shows performance (clicks, conversions) by site where ads are served. Use the site exclusion tool to remove your ad from sites where you’re not getting quality leads.
Addie Conner is going to talk about, well, Matt says, we shall see… Her presentation is going to be way over my head, isn’t it?
ANOVA: Analysis of Variance. It’s a statistical technique for comparing means for multiple (usually 3) independent populations. It’s used to answer the question: are there any main effects or interactions between three independent variables?
A three-factor ANOVA (analysis of the variation) has seven significance tests: 3 mean effects, 3 two-way interaction effects, and 1 three-way interaction effect. To make it simple, we restrict each factor to have only 2 possible designs. In this case, we have 2 x 2 x 2 = 8 possible “appearance” of the ads. We then start to collect the data of their conversion rate.
Take your data and create a graph, where the control the Title is controlled so you can see whether body text and image interact. If there are two parallel lines, there’s no interaction between the two factors. If the two lines aren’t parallel, there is likely to be interaction.
Reasons to split by match type:
- Uniform bid either understates or overstates value leading to inefficient bidding.
- Bidding on a more granular level allows you to maximize the good and minimize the bad.
Did you get all that? I see fellow liveblogger Marty Weintraub covering this session for Search Engine Roundtable as well, so there’s a chance he’s got some complementary info that might help you get a more complete picture of the presentation.