MadPPC

If there is one major PPC question I see the most it’s:  When is there enough data [or clicks] to tell whether this ad or campaign are tanking?  Or, when is there enough data [clicks] to start chopping keywords?

These are very important questions, that I feel should should be learned the hard way, but hell I have a blog.

The very simple answer your likely to get is:  There is no magic magic number.  And to a certain degree that is very true, but leaving someone hanging on just that doesn’t help.  In fact somewhere there should be a general number, however that number is based on other factors.

Factor #1: Payout: Payout will be your number one factor in telling when you have spent enough.  Obviously when your payout is a $1.50 zip submit, you will be able to determine conversion rates much quicker than if your payout is a $40 diet lead.  Well, that is unless your a big dog whom spends many thousands a day.

Factor #2: Avg CPC: Really goes hand in hand with Payout.  Say for that $1.50 zip submit your paying an average of $0.30 per click.  That means to be profitable you need to be converting better than 1:4.

Note: 1:4 being, 1 to 4, NOT 1 in 4.  Huge difference.  1 IN 4 would say that you need to convert once IN every 4 tries.  1 TO 4 would say that you will convert once TO four failure times – so your total amount of tries is 5.  It’s a poker thing…

Anyways, while at the same time with the $40 diet lead, you may be paying an average of $1.50.  You need to convert better than 1:26 and 2/3.

This way the zip submit would have more than 5 times as many clicks than the diet offer.  So the diet offer should be collecting 5 times as much data before making a decision.

Factor #3: Tracking: How many factors are you tracking?  FB you track just the ads, Other PPC platforms your tracking keywords, some people are tracking down to the time of the day they get a conversion, etc..

I guess a general rule of thumb would be: The less your tracking, the less time you need to wait, because your not looking at as many possibilities and how those possibilities convert.

Factor #4: Budget: Have a solid budget in mind for just testing.  For some people that could be $50 or less and for other it could be a couple $1000.  This depends on the campaign and your budget.  As an example, say for the zip submit we want to get at least 100 clicks of data to analyze, so you budget out $30 (.30 x 100) for the test, and agree to not stop the test until you hit that mark.

And / Or

Factor #5: Time: No matter what the budget there is no substitute for pure amount of time, to see how you offer plays out over say a week.  The broader the data, the better you can make an overall determination on the campaign.  It’s true some days your campaign will do better than others, you must try and remove the statistical anomalies from the data the best you can.

And / Or

Factor #6: Rate: The best way I can convey the concept of rate, as I call it, is it’s part time, budget, and tracking.

Especially if your bidding broad on your keywords, you know that some of your keywords get many more impressions and clicks while others, not so much.  So the longer tail keywords need more time to mature than your better performing keywords.  This mostly deals with when you have to cut keywords or set negatives.  It can be worth your wild to give some of your better long tails some time to see how they do.

So at the end of it all I’m normally winging it.  For bigger payouts it can take a while to get a good amount of data.  I have to encourage you to take time, and spread the testing out across several days to get a good idea, but there is no one perfect way to do it in all situations.  Do what feels right, think about some of these factors, and if you fuck up – learn from it.

No go forth and make monies.