October 16, 2019
Explaining the Conversion Discrepancies Between Google Ads & Google Analytics
Constant testing has become such a mantra in digital marketing that it’s easy to forget that testing is really a means to achieve a concrete goal: learning more about your market to outperform the competition. Digital marketing is more powerful than traditional marketing channels because you can observe every change and use accumulated data to develop new market insights. Your performance data contains market research metrics that you can deploy to improve your website, print materials, even your elevator pitches.
You will need a consistent testing methodology to take advantage of these opportunities, and any good test begins with a hypothesis – the insight that you are trying to reveal. Your first question should be, “What will we learn from this test?” If the answer doesn’t reveal something insightful about your market, then your test will not be effective. Here are three hypothesis questions to ask to make sure your tests will make a positive contribution to your marketing goals.
You’ve seen this happen: a CMO wants a display advertising campaign launched yesterday. You like this image, and I like that one. Easy, let’s just test the two images to see which performs better!
Two months later, the image of the hipster guy with glasses beats the laughing group of attractive young people by a narrow margin. What have we learned? That our audience is more likely to give us their contact information based on some guy’s fashion sense? That’s unlikely. But it’s common for advertisers to test new images against each other without any idea of what they’re trying to prove.
It’s critical that every test has an express purpose, and that you can glean real learnings from every change. As you run more tests, you will accumulate market insights that will allow you to build a profile of your audience. You can then apply that profile to each new marketing initiative you undertake, making each initiative more effective than the last.
Let’s say that you have a list of three of your customers’ most severe pain points, and you want to test to find which one most moves the audience to click a search ad. So first, you test ad copy with #1 against ad copy with #2, and #2 receives the highest CTR. Then you test #2 against #3, and #3 wins. You have, in effect, rank-ordered these pain points in terms of relevancy to your audience. You can now apply this knowledge not only to the rest of your marketing, but also to other aspects of your business, like sales and product development.
Although it can be tempting to run tests without considering the broader business context, your knowledge of your business (or your client’s business) should be front and center when building hypotheses. If you ran an inconclusive test, why do you think it didn’t resonate with your customers? In conversations with salespeople or customer service, what do your customers or prospects say matters to them? What do they love about your product? What do they look for in a new product? What led them to switch from a competitor?
Let’s say you recently read a leading competitor’s CEO talk about how his user experience is a major advantage. If it’s a selling point for them, then you can hypothesize that calling out your own ease of use will help you compete more effectively. So your hypothesis is: Our prospects are more likely to try our product if they see the interface. You can test a landing page with a compelling screenshot against a stock image to see if the screenshot encourages more signups. Once you determine that the interface image helps drive more signups on the landing page, you can test image ads with screenshots of the UI, text ads that call out your product’s ease of use, and even sales language regarding the product UI. These tests will help you drive more signups and, ultimately, more sales.
A hypothesis is a statement, not a question, and you should begin every test with an expectation of which element will win based on your knowledge of the market. You should never begin a test by asking, Will we get more leads by offering a whitepaper or a free trial? Instead begin with a perspective and use it to explain why this test will be effective: Because we are offering a brand new product in an emerging market, I believe that potential customers will be looking for information about our product and its benefits, so a whitepaper offer will generate more leads than a free trial.
By starting with a clear hypothesis, you’re using your everyday interactions with your market to generate valuable and statistically significant data that you can use to improve your bottom line. Everyone has an opinion – the beauty of digital marketing is that we can determine whose opinion actually helps improve our business performance!