Digital Marketing Research in The Age of “Test Everything”

“Test everything.”

In a recent online chat about best practices in pay-per-click advertising, this was one piece of advice that I saw repeated again and again by the participants. Some individuals qualified the statement with more detail, but most shared the same general idea: incorrect assumptions about your digital marketing strategy can cause you to miss out on valuable business; to avoid this, you should verify your assumptions through testing. Lots of it.

I didn’t disagree. Whether it’s search engine optimization, pay-per-click advertising, or email marketing, testing is one of the best ways for digital marketers to improve performance for their clients or employers.

The problem with testing everything is that it’s impossible. Testing requires time and money to gather data, and nobody has an unlimited supply of either of those. Because of this, it can only ever be half of the optimization puzzle. The other half is research.

Research helps marketers to be more effective in a shorter amount of time and with less money by revealing which assumptions are safe to make as they kick off their campaigns. For example, some basic research during the research or “discovery” phase of a campaign might tell you that 35-44 year-olds are your client’s best customers, thus minimizing the need to test other age groups (at least at first). Other market research might show that 90% of customers travel less than 25 miles for your client’s product or service, showing you- without any tests- that your campaigns can start out with a focus on this geographical range.

Where’s the love?

So, why did so many of the chat’s participants leave out any mention of research? It may not have been their intention. However, in my experience, the digital marketing community does have a tendency to glorify testing while snubbing research.

I can’t prove definitively that this tendency exists (nor is that the point of this article), but comparing the number of results for the search query digital marketing “a/b testing” with the number of results for digital marketing “client research” or digital marketing “discovery phase” gives a good indication of which topic is more popular among authors.

Neglecting research isn’t limited to digital marketers, either. Anyone at a marketing agency that performs research can attest that there is no shortage of managers and business owners who want to skip the discovery phase and jump straight into campaigns.

This is dangerous. Although some business “gurus”- especially in the world of startups- like to tout a “ready, fire, aim” approach to running companies, the best marketing strategy for most businesses is still researching before acting (i.e. testing). The following are two (broad) reasons why research should always inform a digital marketing plan.

1. Research saves time and money

By providing you with baseline information for your initial targeting, research allows you to skip early tests and use proven targeting, ad copy, and/or images from the start.

For example, let’s say you’ve just acquired a new client who wants you to create a Facebook ad campaign for the company’s line of waterskis. The ad will run from April to September in the United States. For geographic targeting, you have two options:

Option A

Jump straight into testing all 50 states

Option B

Do some research, find the U.S. Coast Guard’s 2011 National Recreational Boating Survey, and target the 25 states it identifies as having the highest rate of boat ownership

With Option A, you will spend some of your budget on states that provide a poor return on advertising spend (ROAS) because you will be targeting all states, but not all states buy lots of waterskis. When consumers from these states click on your ads, it will cost you money, and this will leave you with less money to spend on ads in states that provide a higher ROAS (assuming that your budget isn’t “unlimited”). You will eventually turn off your ads targeting the low ROAS states, but not until you gather enough data to convince yourself that doing so is justified.

Compare this to Option B, in which, as long as there is some correlation between the rate of boat ownership in a state and the state’s consumers’ propensity to buy waterskis, your campaign will start off with a higher ROAS than in Option A. In other words, the research performed in Option B allows you to start your campaign with more profitable targeting, whereas the strategy in Option A forces you to pay for clicks as you figure out which targeting is more profitable.

Clever readers might say that they could skip research by targeting, for instance, the 25 states with the most coastline, but wouldn’t finding these states require some research? Furthermore, even a logical approach like this has its problems. For example, this group of 25 would include Hawaii, California, Texas, New Jersey, New York, and Maryland, who, according to the Coast Guard’s report, are among the 15 states with the lowest rate of boat ownership.

2. Research teaches you things that testing cannot

Some facts about your industry and target audience simply don’t reveal themselves through testing. Think about your ad copy. Over the course of several months of testing, you may begin to discern what tone and terms result in the most clicks and sales, but you won’t really know why until you research the psychology of the target audience. (Note: laddering and client interviews can be great ways of conducting this type of research).

Nor can testing tell you whether your tests are worth performing in the first place. For example, imagine running a display ad campaign for engagement rings on male fashion blogs. Your rationale for running the ads on these sites is that men who are interested in style are probably also old enough and mature enough to be thinking about proposing to someone. However, if it turns out that these style-conscious men’s significant others are actually the primary decision makers when it comes to engagement rings, this approach would probably be barking up the wrong tree. As a result, you might spend months and thousands of dollars optimizing a campaign that was doomed before it ever started. As in number one above, a little research beforehand (whether conducting consumer or client interviews, reading market research reports, or using another method) could have saved you time and money.

Putting it all together

Testing can make your digital marketing better, but performing tests without guidance from research is like treasure hunting without a map- even after a lot of digging, you might end up empty-handed.

For good results fast, I recommend crafting a research-based campaign and then continually testing it. In this way, you will use both of the optimization tools at your disposal and get the most mileage out of your efforts. Don’t worry – there will still be plenty to test once your research is done.