As we saw in part II AdSense Arbitrage: Keyword Selection, finding the right set of circumstances where AdSense Arbitrage is going to work requires a bit of work. You’ve got to get a large enough set of related keywords, get some prices, and then compare and look for the right opportunity. If you read Rae’s blog then you might remember her post on Legitimate Use of Mechanical Grunts, which is a perfect for this situation. However after you get the keywords, and prices you’re still going to have to sift through all that data. I was talking with Scott (aka Web Professor) last week and he pointed me in the direction of Standard deviation.
Ok there is going to be a little math here, and it may look really scary and look like it’s going to make your head hurt. Trust me it’s OK we can get through this really. I hate reading those long boring overly complicated search engine patents just as much as you do, so I’ll keep things simple. The big formula is coming brace yourself …
Ok I promise I won’t do that again. The formula comes from Wikipedia, but since most spammers aren’t interested in this kind of stuff we’re going to assume it’s reasonably accurate. We’re going to work with a sample of 4 bids for illustrative purposes, however I’d suggest a I wider sample range once you actually get going.
Bid 1: $4.58
Bid 2: $4.53
Bid 3: $4.10
Bid 4: $4.05
We are going to get the mean of these numbers (add them together and divide them by 1 over amount of numbers)
(1/4) * ($4.58+$4.53+$4.10+$4.05)
(1/4) * (17.26)
we end up with 4.315
Explaning the next part in words is too complicated so I’ll just show you the formula that funny thing is a square root
√[ (1/4) * [(4.58-4.315)2 + (4.53-4.315)2 + (4.10-4.315)2 + (4.05-4.315)2]]
Inside of each of the parentheses we subtracted the mean from each of the individual numbers and squared the result.
√[ (1/4) * [(0.265)2 + (0.215)2 + (-0.215)2 + (-0.265)2]]
√[ (1/4) * [(0.070225) + (0.046225) + (0.046225) + (0.070225)]]
√[ (1/4) * [0.2329]]
√[ 0.058225 ]
The end result is 0.241 (rounded to three significant digits). OK great your head hurts and 37 people how now unsubscribed from my blog because they had high school math flashbacks, but what the heck to you do with that number? Well the lower that number is the worse the situation actually is, here an example of four other bid prices for an actual keyword:
Bid 1: $16.99
Bid 2: $16.98
Bid 3: $12.00
Bid 4: $10.00
You can see the bid gap is larger by looking at it, but remember were trying do this without having to look at the numbers. I’ll save you the trouble and tell you result is 3.065. So a higher standard deviation will clue you in to a possibly favorable condition. Now I’m not going to recommend you turn your scraper on full tilt and let it make your decisions, what I am saying is get a scraper/number-cruncher to help you identify the best possible candidates. Let the machine do the grunt work, save you time and energy, and you do the thinking and decide if it’s worth pursuing or not. It’s not rocket science, but with a little programming, a little math, and a little creativity you may find a new way to build your income.
Related posts:
- Is Google Pushing Universal Results Doing a va
- Thesis Tutorial – How to Add Adsense Section Targeting Using Adse
{ 5 trackbacks }
{ 11 comments }
Brilliant stuff, Michael, but do you realize how many business models you just ruined! This stuff isn’t supposed to be out in the open
How do you think search marketing firms can charge so much for PPC consulting/management?
I like liked it. That was a comfortable read, and pans out well
Cheers.
Here’s another attempt at explaining Standard Deviation…
The std dev represents how compact or spread out the sample data is.
This site explains it better…
http://www.robertniles.com/stats/stdev.shtml
But for an overly simplistic summary.. the higher the standard deviation… the greater the spread in all of the numbers (which is what you are looking for in this case).
This kinda rings of the old “buy low, sell high” mantra of the stock market… A higher standard deviation means that you have more “room” to buy in low and sell off higher.
Great stuff!!
You probably have just quadrupled the number of BestXsites.info registration.
I Don’t want to sound to anal but there is one small rectifcation to your math.
The actual bid is one cent more then the max of the next one, so for the second example the real numbers will be
Bid 1: $16.98
Bid 2: $12.01
Bid 3: $10.01
Bid 4: Bid 5+0.01
hope I made myself clear…
I’ve written a Standard Deviation calculator APP in C#/.Net. All you do is plug in your numbers and hit calculate. Being that it is .NET you need to download the .net framework. If anyone would like a copy of my calc just let me know.
-Tony
nice infos, thanks!
Excellent information. I just found a keyword with the following bid prices:
46.01
45.99
28.11
2.20
If I calculated correctly, that gives me a standard deviation of 20. Hope I can do something with it…
Ughhhhhh my brain now hurts. Hopefully after some coffe and re-reading this about 5 times I’ll figure it out.
It’s called Excel (or open office) and it will crunch your numbers effortlessly. ’stdev’ yum!
Finding a standard deviation is a breeze with Excel or OpenOffice Calc.
My question: Aren’t AdSense advertiser bid prices NOT fixed, but instead automatically adjusted to $0.01 more than the next lower bid price?
Also, aren’t we looking for a differential between the Google vs MSN bids? What good is the STDEV among bids on the same search engine?
Comments on this entry are closed.