Archive for December, 2005

When you have a successful campaign running, you want it to make as much money as possible per day. How are you going to achieve that? You may be surprised to learn that it’s not by buying as many clicks as possible. Let’s look at some example figures.

Example 1: You make $17.83 per commission. Clicks cost you $0.10, and you ad appears in average position 3. You make one sale per 150 clicks, and you get an average of 15 clicks per day. You make an average of $2.83 per sale after advertising costs. You will make an average of $0.283 per day after advertising costs as it takes 10 days to make a sale on average.

Example 2: You make $17.83 per commission. Clicks cost you $0.09, and your ad appears in average position 5. You make one sale per 130 clicks, and you get an average of 12 clicks per day. You will make an average of $6.13 per sale after advertising costs. You will make an average of $0.566 per day after advertising costs as it takes 11.67 days to make a sale on average.

Example 3: You make $17.83 per commission. Clicks cost you $0.03, and you ad appears in average position 13. You make one sale per 120 clicks, and you get an average of one click per day. You will make an average of $14.23 per sale after advertising costs. You will make an average of $0.119 per day after advertising costs as it takes 120 days to make a sale on average.

(Why does the number of sales per click differ between the examples? Simply because someone who clicks through on an ad in a lower position is more likely to convert than someone who clicks through on a higher spot. They’ve looked at more ads and decided not to click through on those ads, so they have gone through a greater process of selection to get to your ad. Of course, the click-through rate will probably be lower.)

Example 2 shows practically double the income per day after just one change in advertising price point, compared to example 1! You really need to try each price point, and see what income you make over a period of time. It’s unfortunate, but that’s the way the statistics work. I suggest you wait until you have made at least five sales or 250 clicks on a campaign before making changes to improve the accuracy of the results.

You will probably find that there is a specific Max CPC that maximises the net commission per day. This is what you’re looking for. A graph showing net commission per day as a function of maximum cost per click will probably have one peak, with the value dropping to either side of the peak.

You must keep diligent records of your changes to make this work. Record the average, or mean, commission; the mean cost per click (use the campaign cost divided by the number of clicks, and not Google’s rounded figure); the maximum CPC; the mean number of clicks per day; and the mean number of clicks before an action occurs (sale or lead). Also record the start and end dates of the period, and the number of impressions over the period; you’ll need these for comparison purposes. You can get some of this information easily through Google’s reports, but not all of it. Not all of these numbers are needed for the optimisation calculation to be performed, but they are all useful pieces of information.

To help you, use a spreadsheet package. Use Excel if you have it; otherwise get OpenOffice Calc from the free office package at OpenOffice.org, which is worth far more than it costs! I use OpenOffice Calc. I suggest you set it up with these headings, and where I have put formulae in square brackets, you can enter these directly into OpenOffice Calc, and probably also Excel, onto the second row. The cells can then be copied throughout the column, and the cell references in the formulae will have the row numbers automatically updated:

• A: Start Date
• B: End Date
• C: Number of Days [=B2-A2+1]
• D: Gross Commission over Period
• E: Number of Sales
• F: Average Commission [=D2/E2]
• G: Total Impressions
• H: Average Impressions/Day [=G2/C2]
• I: Total Clicks
• J: Average Clicks/Day [=I2/C2]
• K: Average Clicks/Sale [=I2/E2]
• L: Maximum Cost per Click
• M: Campaign cost over period
• N: Average Cost/Click [=M2/I2]
• O: Net Commission over Period [=D2-M2]
• P: Return on Investment [=(D2/M2)-1]
• Q: Net Commission per Day [=O2/C2]

I suggest you highlight the cells that are calculated on-the-fly by the spreadsheet, to remind you that you do not enter the numbers in these cells.

Note that although return on investment is included as column P, this is not the figure you would usually wish to maximise. You want to maximise the net commission per day, in column Q. Usually, maximising return on investment will give you a low net commission per day, simply because you would achieve the highest return on investment by having the lowest possible cost per click, which means you see few clicks!

The spreadsheet with the above headings fits very nicely on my 1280 x 1024 standard-size fonts screen. It is available from the download tools page. To get access for free, sign up to my affiliate marketing mailing list by using the sign-up form at the top of the page.

I hope you found this article helpful. Check back often for more hints and tips!

David Thomas, The Affiliate Marketer

I like to know certain information about the people who buy the products that I promote. The following are handy to know:

1. What keyword did they use to visit my site?
2. Did they come from the search network or the content network?
3. When did they visit my site? (Useful to know, so that you know how long after they clicked on your link before they actually made the purchase.)
4. What was their IP address? (You may then be able to find out exactly what they did on your site, by rummaging through your logs.)

Using Commission Junction’s Shopper ID tracking, here’s how it’s done!

1. Design your page using your normal page design techniques, ensuring that the SID in each link is set to TRACKMEHERE.
2. When you have saved your page with a .html extension, change the extension to .php.
3. Add the following at the top of the file, before anything else:
   <?php
  $source = $_SERVER['REMOTE_ADDR'] . ' ';
  $source .= gmdate('mdHis') . ' ';
  if (isset($_GET['keyword'])) {
    $source .= $_GET['keyword'];
  }
  $source = trim($source);
  $source = strtr($source, ' ', '+');
?>
4. Wherever TRACKMEHERE appears in your document, replace it with the following:
   <?php echo $source; ?>
5. Add the following query parameter to your destination URLs in AdWords:
   ?keyword={ifcontent:CN+}{KeyWord}
   A sample URL might be:
   http://myurl.com/index.php?
keyword={ifcontent:CN+}{KeyWord}

That’s it. You’re done.

When somebody causes an action that appears in your CJ logs, such as makes a purchase, the SID will be set to the following:

1. The IP address of the user.
2. Exactly when the page was served. This consists of five pairs of digits, with the first pair of numbers being the month, the second pair the date, the third pair the hour (24-hour clock), the fourth pair the minute, and the last pair the second. I chose not to include the year, as it’s very unlikely that cookies on CJ are ever set to run for longer than 365 days. Note: As your server may be located in a different time zone than your own, the time shown is given in UTC, or GMT. If you prefer to use local time, remove the “gm” in front of “gmdate” in the PHP code at the start of the web page.
3. The letters ‘CN’ if the click came from the content network.
4. The search keyword used when the ad appeared. (This is probably only meaningful for the search network. I’m not sure what it displays when somebody clicks through from the content network.)

You can test it by going to your URL and entering a query parameter straight away, and looking at the CJ affiliate links generated. They should all contain something like the following:

sid=123.234.123.234+1217123840+Test+Keyword

(The above would mean customer IP address of 123.234.123.234; date and time of 17th Dec, 12:38:40; not the content network (no CN after the date and time); and keyword of Test Keyword caused the ad impression.)

With this information, you can start building up a picture of which keywords convert to sales; how long it takes for the average customer to buy; what customers do on your site; and whether it’s worth paying for content network clicks.

David Thomas, The Affiliate Marketer