Aug 24, 2015 AdWords—The Best Market-Research Tool Ever
When you plan a marketing campaign—in any medium, online or off—you may not know ahead of time what message will resonate most. Google AdWords pay-per-click (PPC) ads let you test competing messages cheaply, quickly, and accurately with a live audience.
You can write two, three, or more ads that will appear alternately in front of the same audience (people who have searched in Google using the same keywords). These ads can:
- Emphasize different products or services.
- Make different offers (e.g., your introductory price vs. your volume discount).
- Emphasize different values (e.g., price, reputation, convenience, guarantee).
- Offer the same benefits, but in different wordings.
Each ad will be set to appear the same number of times as the others, in the same position on the page (on average).
The result? Your audience will click on some ads more than others (i.e., they’ll have a higher CTR%). These ads have the most promising messages.
Key Steps for Success
Make sure that you can follow through on all the offers that you make in your ads. Test or not, you are making real offers to live people! If your current website isn’t ready for this, you can set up a page that lets visitors join a mailing list, possibly giving them something small in return (e.g., a PDF about your upcoming offering). If necessary, this page can even have its own domain name separate from your main website.
Find keywords that represent the audience you want. These are the keywords that, when searched for, will trigger your ads to appear. You won’t need to fight competitors for all the best keywords–you only need enough to test your ads.
Make sure every ad is well written. Otherwise the “winner” may simply be the one ad that’s written well! Writing within the extremely small scale of an AdWords ad is a special skill.
Estimate the test’s time and budget. You’ll want to know roughly how many showings you’ll need per ad (see point 5. below), and how long they’ll take. Google’s tools permit a very rough estimate before starting; the first 2-3 days of showing ads permit a better estimate.
When one ad beats out the others, do statistical analysis to make sure it’s not just by chance. You can determine a percent confidence level, e.g., 95% sure that the top ad isn’t just the luckiest ad so far. With less than 90-95% confidence, keep the campaign running.