Enhanced CPC is a new AdWords bidding option (for search as well as for display network) I was anxious to test. It seems that this option uses the conversion history to increase/decrease max CPC bid so we will get more conversions and a higher profit. If the system estimates that an auction has a high conversion probability, the CPC will be increased, otherwise will be decreased. Max CPC can’t be raised with more than 30%. So with Enhanced CPC, it can be paid more than the max CPC you specify.

Google does not specify if there is a necessary minimum number of conversions for this feature to work well. I already activated for few campaigns (with different conversions history – small or big number of conversions). I will let you know when I get relevant results .

What happens in online? When people start to search specifically for Christmas offers or services? When we should be present on search engines? We get an answer, with relative data from Google Insights for Search.

Google Insights for Search is a free tool offered by Google who shows the searches volume (relative to the maximum point of searches) in a specifically geographical area on a given period of time.

For example, is we have a hostel in Romanian mountains and we want to see if there is seasonality related to cazare searches in the last 3 years in Romania, we will choose in Time Range: 2009, 2008, 2007 and in filter area cazare as below; we can see that there is a searches peak in April – September.

For cazare munte the searches volume decreased in 2009, and seasonality is more fragmented. We can notice though, peaks in Easter, summer and autumn seasons.

If we do the same for Christmas, we see that the users start to search for cazare craciun at beginning of September, earlier (with few days though) every year.

It’s interesting than in 2009, there is a peak in December 20th-26th.

Results changes if we add another term search pensiune craciun:

It seems that people started to search more specific, and we can see that pensiune craciun and cazare craciun searches volume is greater in 2009 than in 2008 or 2007. Very important, the highest peak for these searches in 2009 was recorded in the week October 4th-10th!!!

So, September is coming. If you work in tourism, maybe you should start preparing for Christmas advertising .

Dimensions tab can be displayed by clicking the arrow next to the campaign tabs.

As dimensions, we can choose from:

We can display the below columns, the most important being the clicks, conversions and conversion rate.

Interesting to see a report that proves statistically that the greatest number of conversions is recorded on Friday, isn’t it?

Of course, to find out if it is Friday, you have to test it for your own product and campaign.

The test has run on two campaigns, both targeted on Google search. First I wanted to end the test on August 12th for both campaigns, but seen how the results are, I considered that I should delay the end date.

Test objective: to see how max CPC increasing modified the number of conversions
Experiment increased max CPC with 25% for both campaigns

For the non-branded keywords campaign these are the data I got:

More, the control conversion rate was aprox 4 times greater than the experiment conversion rate. Explanation I found is that the ads position increased. This brought a lot of traffic, but not necessary clients. Data being statistically relevant, I stopped the experiment on August 13th.

For branded keywords campaign, data are not statistically relevant, that’s why the test is still running.

Initial post version. Test Explanation

I noticed few days ago, a new tab in AdWords Campaign Settings.

It is called Experiment, it is in Beta, and it allows us to … what do you think? We can test a lot of things, from adding keywords to setting max CPC for an ad group. Depends a lot of what you would like to get how you will use it. More details about the steps you have to take for doing an experiment you can find on Google AdWords Help.

In this account (where I noticed for the first time the experiment section), I noticed that the conversion rate decreased while I was increasing max CPC. I decided to test and see what happens with conversion rate if I use a max CPC smaller.

I defined the experiment, as start date, end date and traffic – how much is controlled, how much experimental.

After I setup the experiment, an icon appeared in front of all the ad groups.

After I did the changes, I started the experiment.

I will let you know what I got and how the results look like .

Challenged is too much, as the solution is already offered by Google here.

To make the story short, you have to create a new empty page that should contain the below Java code:

Visit that page. And create a customized filter as below:

If you need more help, leave a comment at this post or contact me by email.

Update:

The nice part was when I realized that I implemented Google Analytics new code . So, what I wrote above wasn’t right. Especially thanks to the fact that the function was deprecated. Anyway, thanks to Madalin Dinu, I did it. For Google Analytics asynchronous code, the function is _setCustomVar and it can be called as below:

Function could be called in the body of the file, too.

Just became Google AdWords Certified Partner on PFA.

Wondering for how many days .

I would like to see what bidding is more efficient from budget and performance indicators point of view. So, I ran a test: two campaigns with the same landing page, same ads – as message and formats, targeted on the same placements (to be sure that the ads placements on pages are almost similar). The campaigns ran in the same period of time with the same budget.

Test results (July 18th 2010)

Notes:
CPM bidding campaign started first. CPC bidding campaign started two hours later (the number of impressions at least started to be displayed in the account 2 hours later than at the CPM bidding campaign).

Image ads have been not approved in time for the test (they are still in Pending Review ), so both campaigns ran only with text ads. I did not optimize the costs, as it was a test campaign and I wanted to see how system works. I set up max CPC of 4RON for CPC bidding campaign and max CPM of 8RON for CPM bidding campaign.

Campaign budget was 60RON/day for each of them. Period of the test: 2 days, enough to see how system works.

The test results were surprising:

As you can notice, CPM bidding campaign had better performance indicators than CPC bidding campaign. The equivalent CPC was with 12% smaller than the CPC paid in the CPC bidding campaign. CTR was 2 times greater. In the same time, the number of impressions was much smaller (aprox 2 times).

At the first sight, it seems that a CPM bidding campaign on placements will bring the wanted traffic with a smaller budget and better performance indicators than a CPC bidding campaign (important for account history who influences the real CPC/CPM paid).

A possible explanation would be the text ads expanding on slots.

Please don’t take the result of this test as absolute truth. The period of the campaign was short, the vertical will not be the same all the time. It’s just a question I raise regarding of how system works.

And, I really think that if you are interested in advertising on content network, you should not neglect CPM bidding campaigns.

PS. Content network is created especially for awareness. Is it possible that CPM campaigns have priority vs CPC campaigns??? The way how campaigns started and their performance suggest this. This would be an interesting question for Google .

Original version of the post – Test explanation

Let’s start with a question .

I want to advertise on the Google content network. What type of bidding shall I choose? CPC or CPM?

At the first sight the answer is easy: depending on the campaign’s objective. If the objective is awareness, visibility, we should choose the CPM bidding, as we will pay for impressions. If the objective is traffic, with CPC bidding we will pay for clicks. In the essence, it should be the same, as the content network is used especially for branding and awareness (in Romania still does not bring very good results for conversions).

Theoretical, the system should optimize the budget in such a way, so we will get what we want: impressions for CPM bidding and clicks for CPC bidding.

I would like to see what bidding is more efficient from budget and performance indicators point of view. So, let’s run a test. Two campaigns with the same landing page, same ads – as message and formats, targeted on the same placements (to be sure that the ads placements on pages are almost similar). The campaigns will run in the same period of time with the same budget. Haven’t decided how many days (2-3 until I run out of money ).

Targeting for both campaigns: content network only on selected placements; country: Romania; languages: Romanian and English, all devices; ads scheduling: 24 hours/day, any day of the week

Max CPC bid for CPC campaign: 4RON
Max CPM bid for CPM campaign: 8RON

Values are huge, as I want to be sure that we get as many impressions/clicks we want and to be sure that the ads have visibility (initially). I will not optimize the values of CPC/CPM as I want to see how system works.

Ads used: text and image built with Display Ad Builder (336×280, 728×90, 120×600, 300×250, 160×600). You can see the messages below. I hope that the image ads will be approved before test ending . The landing page of the campaign: the Romanian version of this post AdWords – Licitare CPC sau CPM? (test)

The placements have been chosen with Placement Tool. Of course, if there is no traffic I will add more placements.

Test results will be published as update at this post.

Results expected: low CTR as the ad is for experts mainly. CPM campaign will have better CTR due to text ads expanding on slots.

We cannot judge the relations from the social networks with the same criteria as offline relationships. Of course, the 3000 friends I have on Facebook will not help me to move my furniture in my new house, and it will not be there when I am crying.

Have you ever had the curiosity to see how many relationships from Facebook become real? 3000 friends on Facebook; you are lucky if 10 become virtual friends. Maybe you will meet 1-2 in offline. And then, what good to use social networks??? I asked myself few months ago, when I did for the first time above statistic. Then I decided to take a pause from online socialization . Anyway I returned to communicate with the few friends I had. What good to have 3000 friends on Facebook? To find out what happens on markets I do not have access to. To encourage people abroad when they miss their country. To filter those 3000 and find those 10 I would never find in another way.

One week ago someone was giving me advices about how can I use Facebook: cool stuff – applications, games to advertise online trainings and services. My answer was direct and short: “I could do it, but I do not want to use personal accounts to build my brand”. Using social networks accounts for advertising means to say what you have to when you have to. To hide yourself making yourself known. To talk and answer but not direct, not sincere and not profound. This will be the capital guilt.

I believe in personal brand, but not in the one built based on social network advertising. I believe in people who fails and rise and those who will never be applaud on social networks.

No, my dear, I will not start building my brand.

Even before we start to build the keyword list, to ask “what keywords search users to find the product/service I promote” we should be careful at duplicates.

Did you know that…

… caps have no importance? If we write a keywords with capital letters and one with small letters, these two keywords are identical so, they will canibalize???

… there are symbols unrecognized by system? You will get error if you will try to add ! @ % ^ * () = {} ; ~ ` <> ? \ | in keywords.

… there are symbols ignored by system? – . are ignored.
So e-commerce is identical with e commerce, and Str. is identical with Str
Both versions will compete if appear in the same search campaign.

… there are recognized symbols by system? & and accents are recognized. So Romania is different than România. Depending on the match type, we can target search queries with accents or not. If we choose Romania in broad match, most probably the ads will be displayed for România as well. But if Romania is in phrase or exact phrase, this thing will not happen.

Coming next: Keywords in AdWords – part 2 – strategy for building keyword list.

R: There are more than reasons for these values:

1. One user could click on an ad more than once in the same browser’s session. This could be a normal user behavior, if he opens the site in a new tab and comes back to the search results page and clicks the ad again. I can’t say it’s normal, but it could happen.

If we are worried that these clicks are invalid, we could run an Invalid Clicks Report for the account or for the campaign.
How can we do this? In Reporting section choose Reports –> Create new report –> Account or Campaign
In Performance Statistics area choose invalid clicks.

After you run the report you could see something like:

2. Data we see in the account are not updated in real time. Clics update at every hours, impression at every few hours, that’s why it’s possible that the data you see in a given moment it’s incorrect. To avoid this errors, I recommend you to not take into consideration the current day for reporting. .