In a previous post, I detailed how to analyze sales with visitor keyword searches, specifically how you can use PHP to assign a unique id to a visitor, store that id with their Google Analytic information, and pass that to the affiliate network. The reason for going to all that trouble is so that you can analyze the keyword and campaign info that sent the visitor to your site for visitors that actually bought something from your affiliate vendor.

Well, I now have a really practical example of how this analysis saved me money.

One of the companies that I am an affiliate for sells collectible dolls. They have lots of different types: newborn dolls, dolls that breathe, dolls that have heartbeats, anatomically correct dolls, etc.

I have a Google Adwords campaign for each of these different categories. Without my PHP magic, I could tell how many impressions and clicks each campaign got me, but I had no way to tell which of those visitors actually bought a doll because I was sending them off to the other website.

Using my PHP technique, I was able to run a report that showed me all the keywords just for those people that purchased a doll. And I found out something interesting.

Even though I was getting lots of impressions and clicks for anatomically correct dolls, not a single person bought one. That’s out of about 200 visitors who came to my site because they searched for anatomically correct dolls and clicked on my ad.

I was showing them a page with anatomically correct dolls, exactly word-for-word what they had searched for. So why wasn’t anyone buying an anatomically correct doll?

To find out, I created a survey (the topic of another post I have yet to do). I directed all the people who clicked on that ad to my survey first, then on to my website. I asked them a couple of questions like “what were you hoping to find when you searched for anatomically correct dolls?” etc.

And I found something interesting.

All the people who filled out my survey were looking for anatomically correct dolls that could be played with. One person gave me more helpful information: she’s opening a day care and her child development background prompted her to search out these types of dolls.

Armed with that information, the lack of sales made perfect sense. My site sells collectible dolls which can cost up to $199.00. These are dolls that are made to be put on a shelf and admired, not played with. These people were looking for dolls as toys.

So I disabled that particular Google Adwords campaign. Without the ability to relate keywords to sales, I never would have discovered this and I would have continued to waste money on a Google campaign that wasn’t resulting in any sales.

Posted by susb8383 under Affiliate Marketing & Google Adwords | No Comments »

One of the latest big changes by Google is to display a message next to a keyword that says something like “Bid is below first page bid estimate of $0.20.”

Now, logically you would think that you’d see this message for every keyword that causes your ad to show on > page 1, and this amount is what you need to move it to page 1.

But this isn’t necessarily the case. This message also depends on the quality score for the keyword.

For example, I got this message for a keyword that showed my average position as 3.1. Now, clearly 3.1 is on the first page of ads. So why was Google telling me this?

On the other hand, I have a keyword that shows an average position of 12.1—clearly not on the first page, and yet I do not see this message.

So what happens if there is only 1 page of ads total? Logically you’d think you would never see this message, but that isn’t the case.

I had the good fortune to bid on a keyword that had absolutely no ads associated with it. That doesn’t happen very often, but it pointed out something interesting. I assumed that since I’m the only advertiser, I could bid as low as I wanted and still get the #1 position. Wrong!

I bid the lowest amount possible—$.01. Google showed me the dreaded bid is below first page message. And I found that if I went to Google and typed in that keyword, my ad wasn’t showing at all.

So in other words, even though there were no other advertisers, Google would rather not show my ad than show it and only charge me .01.

Interesting…

Posted by susb8383 under Google Adwords | No Comments »

The Google Adwords keyword tool is a free tool that gives you keyword suggestions.

I’ve always assumed that the suggestions are a full list taken from things people searched for in Google.

But I just discovered that it’s only a partial list.

Here’s an example. I put these two keyword phrases into the tool:
jasmine
waltraud


The results I got back showed me 150 results. The first two were the words I typed: jasmine and waltraud.

The other 148 were all related to jasmine: jasmine tea, jasmine plant, paul jasmine, etc. There were no other results related to waltraud.

I reasoned that maybe it only shows you 150 results.

So then I changed my keywords to:
jasmine doll
waltraud


Now I got back 86 results. There were 31 results relating to waltraud including:
waltraud klasnic
suzuki waltraud
waltraud reiner
waltraud hanl dolls


Now I refined the search even more. I typed in waltraud hanl. I got back 6 results including:
waltraud hanl dolls
waltraud hanl doll


It’s interesting to note here that “waltraud hanl doll” was not a result I got back when I searched for waltraud and jasmine doll.

Why not? From my first search it’s clear that Google can return 150 results. So when I got 86 results for jamine doll and waltraud, why didn’t it include “waltraud hanl doll” and give me more than 86?

I’m going to assume that the more general your search, the more likely that Google will show you results that got more impressions.

So the moral is:
1. Don’t assume the keyword tool is showing you everything that was searched for.
2. Do multiple searches when you use the tool—the more detailed, the better.

Posted by susb8383 under Google Adwords | No Comments »