Crush AdWords: A Successful Campaign is in the Maintenance (Part 5 of 5).

I’d like to talk a few minutes about the true strong suit of AdWords (tracking & testing), and how most people misuse it. Many people write a few different ads, put them up and let Google rotate them out and decide which ones are the most effective. While that’s better than not doing anything, you’re still wasting a lot of time (and a good bit of money).

A better approach is to use AdWords to test different Elements of a single Ad against each other, to optimize that ad, before trying to test a bunch of different ads together.

WHAT DOES THAT MEAN
Okay, let’s return one last time to our weightloss example. And let’s say we want to create ads that will take people to our “swimsuit season” landing page. Typically, we’d write several different ads and put them up against each other. In fact, we’ve already written several different ads, haven’t we?

Lose 10 pounds quickly
get ready for swimsuit season
in just 20 mins a day!

Turn Heads in a Swimsuit
Knock ‘em dead at the pool
or beach this summer

etc, etc, etc.

Here’s the problem: you’re comparing apples and oranges here. You might learn which ad works best, but you won’t be able to draw any conclusions as to why. And if you don’t know WHY a certain ad pulls better, it is going to be difficult to tweak your landing page to improve conversions.

START WITH ONE AD
Don’t get me wrong, we’re going to use all of our ads that we created. But first we’re going to get started by picking one and testing the hell out of it. Let’s take the first ad above, as an example:

Lose 10 pounds quickly
get ready for swimsuit season
in just 20 mins a day!

STEP 1: Does the Order Matter?
We’ve got three basic points in here: 10 lbs quick and easy; Swimsuit season; and 20 minutes a day. The first thing we’re going to do is try to figure out which one pulls best as a headline. So we’ll write 2 variations on the ad.

Just 20 Minutes a Day!
Lose ten pounds quick and easy
and get ready for swimsuit season

Get ready for swimsuit season
Lose 10 pounds quick and easy
in just 20 mins a day.

It’s the same ad, but we’ve written it 3 different ways. Now, run a test, and let Google rotate the ads equally. Google has a setting where they will display the ad that pulls the best more frequently, but that will skew your test results. We’re looking for hard data here.

(Now, you might be saying that “20 minutes a day” doesn’t seem like a terribly strong headline. However, keep in mind that the keywords we’ve selected have to do with wieghtloss (or even looking good in a swimsuit). The reader is already going to be thinking about this, so it’s not like a print ad, where we have to grab their attention, explain the benefit and select the audience. They’re LOOKING for this information, so that part of our job is already done. )

So, run this ad until you get a few hundred impressions apiece to see which one pulls better. Chances are you will have a clear winner. If not, then just pick the one your gut tells you you should go with. For our exercise, let’s just assume that the original pulled the best, and go with that one.

This ad will be considered our “control,” and our goal from here on out is to beat it.

STEP 2: Tweak it Three Times.
Now we want to see if we can tweak this particular headline to make it more effective. So let’s come up with three similar, but different versions.

Lose 10 Pounds Quickly. (the original)

Lose 10 Lbs. Quickly

Lose 10 pounds, fast!

Lose Those Last 10 Lbs, Quick!

Let me raise one point here: you are going to be surprised by the difference in the response rates in two headlines like the first two above. Even though they say the same thing, for some reason, just the way it looks will make people react differently. Sometimes the spelled-out version works, sometimes the shorter. This is why testing is so important.

Leave everything else in the ad the same and test these four headlines. Since you know you the original works for you, run that one 70% of the time, and substitute the others out 10% each. You’ll have to test for longer before you can get results that are statistically meaningful, but since you’ve got one that works, that’s not really a problem. Remember, from here on out, we’re simply trying to beat the control.

Now, if you have one that is consistently hitting a better conversion rate than the original, pause the other two, and split the campaign 50/50 between the control and the contender. If the contender outpulls the control consistently over time, guess what—you have a new control!

STEP 3: Tweak the Body
Now, with your new head in place, it’s time to bring the body inline. Make a few simple changes. Add a strong call-to-action. Maybe even change out a benefit. And test against your control.

Finally, you’ll end up with an ad that pulls pretty strongly. And having tested the elements independently, you’ll have a pretty good idea of what works and why.

STEP 4: Tweak Your Conversions
Once you have “optimized” your control ad, it’s time to look at your landing page. Do these same “tweaks” to it and see if you can improve your conversions. After a few hundred clicks, change your headline, and see if that improves the number. After another few hundred, change it again. Then reorder the benefits in your body copy, to find the order that pulls the best. Tweak the images and the call-to-action. See how high you can push your conversions. When your numbers reach a plateau…

STEP 5: Try a New Ad.
Now bring in one of your other ads. Reorder the elements, as we did initially with our control, and run the different versions of this new ad against your control. If one of them outpulls the control, pause the others and run a 50/50 split. If they all outpull the control, then drop the control and run an equal split among them.

Follow the steps above until you end up with a new control that consistently outpulls all of the rest. Tweak it to optimize its effectiveness. Tweak your landing page to improve conversions. Then bring in another ad and do the whole thing over again.

The goal here is to find an effective ad and run it. Then, while you’re running it, see if you can come up with one that works even better. This should be an ongoing process, one that evolves almost continuously.

There are options that can make this part easier. You can use a service like Aweber.com to create and split-test landing pages in real-time, in much the same away AdWords will split-test your ppc campaigns. Or you can pick up a product like Omniture or Google’s Website Optimizer to help or even automate the process. But that could be a whole article series (or book) in and of itself.

I hope you enjoyed the series. More importantly, I hope you got something out of it. If nothing else, the point I hope you take away is that, simply setting up an AdWords account an slapping up a few ads is a haphazard way of managing a ppc campaign, and in the long run you will be leaving a lot of money on the table. But with planning, time and a little effort, you can put some basic principles to work and…well…crush AdWords.

PHOTO CREDIT – Source: SXC.hu