We’ve all heard Albert Einstein explain the definition of insanity as “doing the same thing over and over again and expecting different results.” Yet we continue to run ads with the same messaging month after month hoping for better conversion rates. The problem is most marketing teams are forced to make decisions based on an executive’s gut instincts rather than data. To prevent insanity, A/B testing should be at the forefront of every online marketing strategy. Lucky for online advertisers, A/B testing in Adwords is now easier than ever.

Audience is King

Tired of sales complaining about lead quality? To reach your audience you must first understand them. Avoid the notion that just because you are in marketing you understand all target markets. For your next marketing/sales meeting have a brainstorming session and gather new offers. Since most sales or product-focused employees work directly with your customers they usually have great insight. At this point write down any and every idea. Do not worry about form, length or style, simply aggregate as many well thought out ideas as possible.

Be Organized

Take your list and prioritize the winners. At best half your ideas are viable for ad copy. Start thinking in terms of a text ad. Here’s where your marketing instincts come in handy. Choosing which messages to test requires marketing savvy and experience. Here are a few common A/B test examples:

  • Benefits vs. Features
  • Branded vs. Non-Branded
  • Question vs. Statement
  • Punctuation vs. No punctuation
  • Price vs. No price

Now open your most trusted tool (Excel) and create an offline version of your keywords/ad groups. Write two new ads for each adgroup. Keep in mind, a true A/B test only changes one variable. In other words, either duplicate the description lines and test the headline or vice versa. I prefer to duplicate the description for both ads and test different headlines.

Create your Experiment

In Adwords head to your settings and find the Experiments Beta section. Note if you’re using enhanced CPC bidding, you’ll have to shut it off to start the campaign. In the experiments section create your campaign, run a 50/50 split, and start the experiment.

Next head to your ad group and create the two new ads. Select one ad as “control only” and the other ad as “experiment only.”

Congrats! Your experiment has started. To see the results wait at least one week and segment your ad group data by experiment. Sort by click through rate and look for arrows indicating statistically significant results. Read more about making account changes based on experiments results.

A/B testing is the cornerstone for a healthy marketing strategy – It allows us to gain meaningful insight and intelligence. It is the only way you can discover and deliver audience-driven messages that speak to potential customers. Test results can impact your entire business and should be considered the most valuable piece of marketing data. Make sure to share your findings with sales, executives, and your internal brand advocates. Watch your marketing team become heroes and your conversion rates skyrocket.