The Practical Science of A/B Testing your AdWords [Beginners Guide]

The Practical Science of A/B Testing your AdWords [Beginners Guide]

A/B Testing. You’ve heard the term in online marketing, but do you do it?

Testing out variations of an ad campaign to increase sales and profits is still a mystery to many small businesses and marketers. Unlike more traditional advertising methods like TV, radio or print, in online marketing it’s really easy to make changes, test them out and increase your ROI.

That includes making adjustments to optimize your Google AdWords Campaigns.

If you’ve come from an offline advertising background, you may not even know that a few simple tests with your online advertising efforts can increase conversions - and your profits - by 5% to 500% or more.

In this article, I’ll tell you what exactly A/B testing is, why you need it for your Google AdWords, and practical tips for you to optimize your own Ads.

Why Use A/B Testing?

Most marketers (digital and old-school) know that connecting with your consumer is both an art and a science. You may have a good idea about what works creatively. But you need to test it out, and prove your marketing results with numbers.

In a nutshell, A/B Testing is a way to determine what gets your Ads the best results. It’s testing out, for example, your Call-to-Action like “Buy This” and “Get Yours Now”, and then figuring out what your customers are positively responding to the most. The more positive a response, the higher your ROI.

I’ve seen the smallest of changes bring companies increased revenue of 10% to 2,000%. That could be thousands of dollars for your business - at no extra Ad spendfor you. That’s the value of A/B Testing.

What is A/B Testing?

A/B Testing is like a science experiment. Yeah, remember high school biology? Only now, you’re not testing the innards of a frog - you’re testing the innards of your Ads, and improving your results to bring in more money!

In A/B Testing, you make a change in your Ad against your current one, and then determine which one gives you the optimal yields._ Here’s a simple visual of how A/B Testing for Google AdWords work:_

You have group A (the control), and you have group B (the variation).Then you test and measure for the result you’re looking for (ie sales, profit, clickthrough, leads, etc). You then use the group that gives you the best results.

For example, let’s say your business is aiming to increase sales by 10% this month. Here’s how you could run a simple A/B Test to reach your goal:

  • Make two Ads in a Google Ad Group.
  • Choose a variable to test (say your Headline).
  • Set up your metrics to measure results (in this case sales conversions).
  • Write two different variations (in this case, Headlines) in your two Ads.
  • Name your Ads something like “Hln A” (the control) and “Hln B” (the variation).
  • Run the Ads simultaneously, for a certain time, say, one week.
  • Track your “A” and “B” results.

Once you’ve run your A/B Test, determine if one Ad gives you significantly increase results. If is does:

  • Put your budget into to the Ad that brings in more sales.
  • Increase your profits.
  • Repeat.

A/B Testing can be as simple (or as complex) as your business goals, time and resources allow. But, when even small changes in your AdWords can result in conversion increases of 50% or even 500%, it pays to get into the practise of A/B Testing your paid advertising.

Plan Your A/B Test

Before you jump in and start testing up a storm, you need to limit what you’re looking for. In a digital landscape, you can A/B Test for infinity. Focus on your goals, and profits will follow.

For some marketers, who do this as their day job (and get into advanced methods like multivariate testing) it might be ok to test forever.  But – if you’re a small business owner, or a practical marketer, and you’ve got millions of other profitable tasks to work on - you need to narrow your scope.

Plan your A/B test with these 7 simple questions:

1. What’s the GOAL you want to improve?

For example, do you want to:

Screen Shot 2014-03-17 at 12.55.16 PM.png

Choose a goal to achieve. You can even put numbers to it, like “generate 25% more leads within 2 weeks”.

2. How are you going to MEASURE your results?

Set up your analytics to measure and report whatever you’re tracking. This is always something you need to do in science  - instead of beakers, though, you need:

3. What PART OF THE AD are you testing?

There are tons of tweaks and tests you can make to your Google AdWord. Are you testing:

Screen Shot 2014-03-17 at 1.00.15 PM.png

Choose one Ad part to test, to make it simple.

4. What are the TWO VARIATIONS?

Make the two Ads you’re going to run. Use “A” as your control Ad. Make the variation Ad as “B”.

Screen Shot 2014-03-14 at 3.25.30 PM.png

5. How LONG will you run the test?

Before you start, determine the length of the test. That way you’ll have a finite time to conclude, assess your results, and move on.

6. What RESULT do you need?

Decide how much you need to improve your sales, leads, brand awareness and so on.

7. What ACTION will you take?

Decide what you’ll do when you reach your results. For example,

  • If one Ad is doing way better than the other (like significantly better), then use that higher performing Ad.
  • If there’s not much difference, will you keep running the test for a longer time, keep using both ads and stop the test, or try another A/B test with a new “Ad B”.

Ok, now that you know about A/B testing, and your mise-en-place marketing is in order, let’s take a look at an actual A/B Test example.

A/B Testing Google AdWords in Action

I’ll be using a client of ours (but inserting a different name), to show how easy it is to A/B test, and how it impacts your bottom line.

The example is from a company in the party rental equipment business.

  1. Their goal was to: increase website traffic
  2. They measured results with Wishpond analytics
  3. We measured a number of parts of their Ad (I’ll show you 3 of them soon)
  4. We made the Ad variations before running the A/B Test
  5. We ran each test for one week
  6. They needed a 15% increase in web traffic in 4 weeks
  7. They used the higher performing Ad, if it reached their goal

They were advertising their business with this ad:

Ad “A”

google adwords wishpond

Here are 3 tests we ran. You can easily do these too.

1. Headline: “Party Rental Equipment” vs. “Party Tent&Event Rentals”

Your headline is the first thing people tend to see in your text ad. We tested out different variations that would be relevant to the consumer and what they were searching, their keywords and their landing page offer.

For the party rental company, their headline A reads: “Party Rental Equipment”.

Ad “A”

Google AdWords AB Test Wishpond

We changed the Ad headline to be more specific about their products.  Our headline B variation reads: “Party Tent&Event Rentals”.

Ad “B”

Google AdWords ab Test Wishpond

We ran the test for one week, with keywords, targeting, budget and everything else the same.

The Result: A 3% increase in clickthroughs to their website.

2. Display URL: Short vs. Long

As you know (if you’ve been reading my series on Google AdWords), you can change the website address in your Ad Copy. Test out if a change in your Display URL makes a difference to your advertising outcomes.

In the party rental ad example, Ad “A” has the the Display URL of “” (a site I made up):

Online Advertising Wishpond

Ad “B” gives viewers a specific landing page, “

Google Advertising Wishpond

We ran the test for one week, with no other changes to variables.

The Result: An 11% increase in clickthroughs to their site.

3. Call-To-Action: “Book online” vs. “Call us”

Your Calls-to-Action can motivate a customer to click your ad, book your service and buy your products. Test out your action asks to increase your conversions.

In the same party rental ad, we changed the CTA of “book online today”.

Ad “A”

Google Ad Words Wishpond CTA Ab Test

We tested it with: “Call us today”.

Ad “B”

Screen Shot 2014-03-14 at 4.24.25 PM.png

We ran the test for one week, with no other changes to variables.

The Result: 17% increase in website traffic.

Yeah, this seemed like a strange result, considering the CTA was to call them, not visit their site. But, they don’t have their phone number on the Ad, and I surmise the “Call us” gives a more personal connection with the consumer.

Given these strong results, we could test out other CTA’s, too. The best CTA’s are action oriented, short and inspire a quick response.

Here’s a few more examples we could use:

  • “book online today” vs. “Rent Tents and more”
  • “book online today” vs. “Book now and Save”

We could then run the tests using “Call us today” vs. the next highest performers._ (See, I told you, this testing stuff can go on forever….)_

Additional A/B Tests to Optimize Results

As you can see, there are many many variables you can test on your Google AdWords campaigns, Ad Groups and Ads.

Each change can lead to exponentially increased returns on your advertising investment. You may be pleasantly surprised at how a few simple words can take your ad results from a 5% CTR to a 17% (or higher) CTR.

A 5% increase to your CTR can increase profits by thousands.

A few more top AdWord components to test include:

  • Your offer
  • Punctuation and capitalization
  • Wording your competitive advantage
  • Location targeting
  • Ad Extensions
  • Landing Page Destination URL
  • Ad Group Keywords
  • Device targeting
  • Time of day


A/B Testing is effective. It works. Once you get used to it, it’s really not that hard to do. When you are A/B Testing, make sure you keep your end goals in mind. It’s about increasing your business ROI, and making your company more profitable without excess spending. Don’t get so drowned out by the testing details that you lose sight of why you started A/B Testing in the first place.

Start by trying out one or two variables. Test. Improve your results.

Read more about Google AdWords best practices:

*Read more about A/B Split Testing: * - How to A/B Split Test your Facebook Ads to Maximize ROI - A/B Split Testing Your Email Marketing Campaigns - How to A/B Test your Landing Page to Maximize Conversions