The Book for SEO, PPC, & Analytics. Reviews, Comparison, & Tips at
                                                 Home       Explore the Book       Insider Newsletter       Contact Us

Insider Blog

Insider Newsletter

About analytics

About SEO

About PPC

The golden triangle

Table of contents

Book's index

The authors

Case studies


Book links

Sample website

Bonus chapter

Book updates


Press kit

Our company

Buy the book


Pay per Click: Finding the Glitch

Author: Stephanie Cota

Break it down. That's what they always say. The glitch - how do you get more of your potential customers to click on your ad?

You've got your keywords. And you've got your ads. And you've got your landing pages.

When you take x number of keywords and x number of ads, it adds up. For example, let's say you have 10 keywords and three ads. A simple multiplication produces 30. We end up having different combinations of keywords and ads that produce results. Each keyword can show any of the three ads at any given time it is searched for.

Although it adds up math wise, this isn't how it works in producing leads and sales.

For example, person number one types in teeth whitening and sees an ad that talks about free parking at the dentist office. Person number two sees an ad that talks about getting a beautiful smile through teeth whitening. And this ad has a coupon that will help him get this beautiful smile.

Two people searched for teeth whitening. Each person saw a different ad. Which of these two people will click on the ad they saw? What if neither of these people found the ad compelling enough to click on?

Here is an example of a smaller localized market. The advertising is targeted to a 30-mile radius around the dentist's office.

Our campaigns are focused on a small set of keywords, our ads are focused on a specific value proposition and our landing pages are focused on a specific offering. And let's say we want to get our campaigns up and running quickly.

Given these parameters, we ran our localized PPC campaign for six weeks. After six weeks we see there is considerable interest in one of our major product offerings.

How do we know this? We can see our selected keywords get searched on several thousand times a day. As a result of frequent keyword searches, our ads get shown frequently.

Now let's take a step back. A common factor that makes localized markets more complex is competition; particularly, those markets selling commodities.

Within a localized market, the numbers are smaller to begin with. Within a 30-mile radius, there are x number of potential customers. Let's say this potential is 1000 customers. There are several dozen other competitors vying for the same potential customers.

These customers are searching for you. Each potential customer sees the ads displayed after searching their keyword. To find out about your offering they will click on your ad. If your ad gets clicked on more, more people are finding out about your offering. If not, less people are finding out about your offering.

What does this mean? It means each one of your potential customers is finding out about another offering. Not yours.

You'll need to systematically test your ads. Yes A/B split testing is what I'm getting at. It's simple. An ad consists of a heading, two lines of description, and a display URL. Write several variations of your ad and let them run for a set time. For example, try switching two words around in your heading. For our localized market that gets several thousand searches on teeth whitening, we can run our new ads for four weeks.

Although our numbers are small, they are highly directional. Remember, with search engines, everything is exponential. For example, one ad gets a 1% click-through and another ad gets a 7% click-through. This isn't logical. Rather, it's the ad that gets the better response that produces more leads and sales. For example, an ad with a 1% response rate, may have gotten 10 clicks while an ad with a 7% response rate, got 70 clicks. Assuming a 5% close rate, our ad that produces 10 clicks will get us one new customer. Our ad that produces 70 clicks will get us four new customers. The second ad is 4 times as effective as the first ad. Drawing this out over time shows that within a few months, our customer base can increase by 4 new customers per week. Over a three month span this is 48 new customers. Compare that with twelve new customers.

Because teeth whitening is searched for frequently, ads are shown each time. As the frequency of ad displays increase, we are able to see the response rate. For example, by running six variations of ads around the benefits of teeth whitening, we can see which ad and keyword combinations grab people.

If three of those ads don't do well then we delete those and add three more ads that are similar to the three ads that are doing well. We take the headings and switch the words around. We take the benefits and present different benefits around teeth whitening.

The result - after two months of systematic testing we know what works and what doesn't work.

By split testing our ads and taking our chances on ads we didn't know would work, we've grabbed a larger number of our potential customers. They know about us because they clicked on our ad.

We didn't lose a lot of time either. We identified the glitch. How? Through isolating the items, we saw the keywords were getting searches. We saw our ads were showing up frequently. We also saw that our ads were not getting clicked on. Furthermore, we see our competitors' ads. By putting ourselves in the shoes of our potential customers, we could make inferences into the ads that were getting clicks. For example, ads that stress free parking at the dentist may not be as compelling as ads which present a coupon.

By looking at our campaign in a systematic way and by correcting early, we got the worm.


© 2007 | Privacy | Sitemap | Icon Explanation | Contact


Knowledge Base

Analytics FAQ

SEO White Paper

PPC White Paper

Case Study: MIT

Bonus Chapter

FAQ Clicktracks

PPC: Find the Glitch

FAQ: Blogs

FAQ: Adsense

Book: Web Analytics Hour a Day

Book: Actionable Web Analytics

Book: Web Analytics Demystified

Book: The Long Tail

Book: IBM on SEM

Book: Vise

Book: Batelle

Book: Goodman