Relevant math for PPC

[terriwells]

I have an oddball question. I was reading another thread on PPC here and saw a comment from EGOL about how complicated the math can get for PPC campaigns. As it happens, I've been working my way through math at the Khan Academy for the past two or three months; I'm going through several areas at the moment (now I know why the UK folks call it "maths," plural). But the one that specifically came to mind was statistics.

So here's my question, and apologies if it's dumb or obvious: do you use statistics a lot for PPC campaigns? I'm thinking specifiically to understand which of your ads are more relevant or generating more income/profit (not just clicks, mind you). I'm thinking that concepts like the variance within a group of numbers, the standard deviation (which I haven't quite gotten to yet), and maybe even certain kinds of unusual charts (I've learned about more than I knew existed!) might help. Just curious.

[nooby]

Personally, I'm more inclined to do regression analysis to id optical cpc for keywords. While there may be some long-tail keywords that are more profitable than more competitive keywords, the frequency may well be too low for you to ignore the latter. Now, if you could find a competitive keyword that was so highly priced and the clicks non-purchasing that you were losing money,you have a nice benefit (assuming landing page and ads aren't the problem). But if a keyword is so competitive that you only break even, you still want to be in the marketplace to keep the competition from capturing more marketshare.

with enough info regression analysis can help you find ideal price.

[EGOL]

Terri,

Get a copy of Advanced Google Adwords by Brad Geddes.

Start plugging through it and you will find out what type of math is needed. It is mostly simple +-x/ ... but done in Excel.

It mostly has to do with these variables...

cost of an item
income from a sale
average number of clicks to produce a sale
average price of clicks
average profit or loss per sale

This is done by keyword to eliminate unproductive keywords.

You also use it to optimize the level of your bidding with the goal of producing a maximum profit per day.


After you have read the book try to attend one of these...
http://certifiedknowledge.org/adwords-seminars/

Best two days that you could ever spend on PPC learning... but don't go unless you have experimented a lot with Adwords and have read the book.

These classes are demanding. Sit in the front and turn around when the lesson reaches excel. Then you will see how few of your competitors will be really competent on the math. This has nothing to do with the quality of the teaching. Brad is excellent. However, the typical attendee thinks he is going to sit there, sip coffee, fart and go home an adwords expert.

Go here... watch some of the videos on the right..
http://certifiedknowledge.org/adwords-seminars/

If you are going to spend serious money on PPC you MUST DO THIS.

[ZenoIntl]

I don't believe it is that the math itself is complicated, it is that there are so many analyses to complete depending on how many keywords and adgroups you have that it gets to be a long process.

Excel, as mentioned, is a good way to go about it. It is pretty straightforward when you are trying to sell a product and you have a profit margin to draw break even analyses from but it gets more complicated if you are in the process of brand recognition and are not actually making a sale.

You have to decide how many of the clicks will count by drawing an arbitrary line of what constitutes a successful visit through Analytics and then deduce how much each one of those visits cost.