CPA = Maximum Profit per Conversin for Google?
I've seen a lot of good things written about the CPA Auto Bidding option on Adwords. It does take a lot of the tedious bid management and ROI calculating out of managing an Adwords campaign which can be really attractive especially on very large complex accounts with thousands of keywords.
However I wonder if many of the people that got improved ROI from using CPA Auto Bidding on Adwords, got that because they essentially "didn't know what they were doing" on Adwords in the first place. (read weren't taking the time to calculate their ROI and arbitrarily picking CPC bid amounts based on watching their overall spend relating to their budget.)
With CPA bidding Google will manage your bids for you so at that over time, your Cost per Conversion (1 per-click) figure will average no more than your CPA setting.
This sounds all good on the surface but I started thinking about what is going on here in more depth and I'm not sure this is the best strategy for the seasoned PPC Pro that knows how to calculate their own CPA and determine their own Max CPC. Here's why...
When you specify a CPA setting you are telling Google the MOST you want to pay. Google isn't concerned with saving you money. Their business goal is to make money. So would it not be possible that say your CPA at position 2 is lower than at position 1. Lets throw out some hypothetical numbers...
Say your max CPA setting is $10. You have historical data when you were using manual CPC that shows, at position 2, your average CPA was $5. So lets say you got 100 conversions for $500. At position 1 your average CPA was $9 so you got 100 conversions for $900.
If you switch to auto CPA bidding, and specified $10 as your max CPA, Google is going to get you as much traffic as possible so that your CPA averages right around $10. So what I'm saying is the CPA engine isn't concerned with EFFICIENCY or VALUE. When you specify a max CPA bid you are giving away a key fact to Google and telling them how much you are willing to pay them MAX. Then they get to insure they make that MAX.
So the bottom line is that before using CPA, you need to really analyze your historical data to determine what CPA is your best bang for the buck and use that. You need to know your position sweet spot and tweak your CPA until your average position hits that spot. So there is still active management needed even with CPA bidding.
In my opinion, Enhanced CPC bidding is better than CPA, except in the case you want to be closer to a "set it and forget it" strategy that CPA can give some people. With Ehanced CPC bidding, you already know what average CPC has gotten you the lowest CPA, but the Adwords landscape is not static so you have to always experiment with tweaking your CPC. With Enhanced CPC, you have a bit better control than CPA but still can leave some of the tweaking to Google. On second thought, maybe you don't have more control but just have the illusion of more control? What IS the difference of tweaking the CPA setting to get a certain average position and tweaking the Enhanced CPC setting to do the same?
I suppose one could argue, when you specify a max CPC bid, you are also telling Google the most money you are willing to spend and due to the nature of bidding, Google is going to try to spend as much of your money as it can up to that max CPC.
So as far as I can see, what IT REALLY BOILS DOWN TO, is that all other things staying static (add copy, campaign settings, etc.), the thing that really has the most effect on what your average CPA is, is your AVERAGE POSITION.
So in my opinion the way to get the most value out of Adwords and other similar PPC programs, is that you need to experiment to see what position give you the lowest CPA, or in other words the highest ROI, or the most conversion cost efficiency. Then tweak your CPC or CPA settings to get that position. So maybe CPA is the way to go, as it seems it is the same as managing a CPC without having to manually calculate your CPA/ROI to determine how to tweak your CPC?
But low and behold Google recently retired the position preference setting making it harder to keep a keyword dialed in to a specific position. Gee, I wonder why they did this?
Looks like I found one other person that is on to what I'm on to. They've done the data analysis and provided graphs which more or less illustrates my point:
I agree, CPA bidding is not perfect.
But you can always use campaign experiments, setting 2 different CPAs for your Ad Groups to see how well a higher CPA will work. At some point you will hit the point where you will be spending a lot more per conversion while only increasing conversions slightly - if you are getting a lot of traffic, and experiment should allow to find the sweet spot fairly quickly (though I wish they would allow the statistical thingamajig to show in this scenario they way they do with keyword bids etc. as it would take some of the guess work out of it).
The difference between CPA bidding and tweaking bids by average CPC is that (allegedly) CPA bidding will take lots of factors into account as well as the average position (like the actual search term rather than the keyword you are targeting - which I don't think is possible to do in real time with a bid management tool) (and time of day (though I have not seen a lot of evidence for this one) (and geography) (and network i.e. search partner or google search (I have seen some fairly good evidence for this one) etc etc...
There is an important distinction between minimizing Cost, no matter the measure, and maximizing Net Return. So long as the Marginal Cost does not exceed the Marginal Net Return, it is economically beneficial. Google's CPA feature makes no effort to maximize the Net Return.
Additionally, what is an acceptable Marginal Cost fluctuates with market conditions, a factor that automatic management of PPC campaigns cannot accomodate.