I believe that
cheap behavioral data will fundamentally change the way we do online marketing. How would this look for something as straightforward as developing a landing page?
Below I’ve outlined nine studies — all of which can be done from your own computer and usually for a few hundred dollars or less — that will help you successfully develop a landing page.
Before diving in, you may want to review my method for developing a landing page backwards. I’ll use this approach here and apply behavioral science as we go.
Doing studies
I’m about to use the word “study” a couple of dozen times in several of its plural and possessive forms. That’s a fancy way of saying “study, studies and study’s.”
When I say “Do a study,” I refer to the act of stopping during development and asking yourself, “Do we have any data to help with my current design decision? If not, could we collect some?”
Some studies may only require a quick peek at your analytics. Others may require that you collect some data and may take days or weeks.
All of the studies I’m going to show you can be done from your computer-equipped desk.
Study #1: Ad and email offer analysis
Our first challenge is coming up with that offer that will get visitors to take action. Before you start scheduling brainstorming options, I’m going to make this very easy for you.
The offer only needs to keep the promise of the ad, email, social post or link that brought the visitor to the page.
It’s important to get the offer right.
“Do we have any data to help us with this decision?” Probably. We want to find behavioral data that follows as many of the rules as possible.
We make offers to our advertising network, like AdWords. We also make offers in our emails. Let’s do a study and find out which offers perform the best.
Which have the best click-through rates?
Which have the best conversion rates?
Which have the best return on ad spend?
These should guide us to the offers that we want to promote with ads and the promises we’ll be keeping on our landing pages.
We did an analysis of one client’s AdWords account and found that ads offering 20 percent off and $100 off generated the most clicks. These are clearly transactional buyers. We designed our landing page to match the offers and doubled down on the ad spend. We saw a 40 percent increase in the conversion rate of these new pages.
We like this kind of data because it is behavioral: large sample sizes of customers and prospects (not strangers), the data was taken over time, and it is recent.
The offer accomplishes the first job. Now we need to get visitors to take action.
Add a form
The most common way for a visitor to take action on a landing page is some kind of form. The simplest form is a button that takes the visitor to a registration or purchase process. The most complex allows a visitor to complete a sale.
Once we ask a visitor to do something, to take action, our abandonment rate will rise. In general, the more we ask for, the higher the rate at which visitors will leave our page.
This is where we invite a copywriter in to build value around our offer and handle some anticipated objections.
Study #2: Online customer survey
The copywriter will want as much insight into your customers as possible. They will want access to any qualitative data you have. Reviews, live chat transcripts, call logs and interviews with anyone who knows your customers will fill out their research.
To supplement this data, consider sending a survey using any of the many online surveying tools available. Ask customers why they bought, what triggered their visit and what almost made them not buy.
When the copywriter is done, you should have a number of headlines and images to choose from. Here’s how behavioral science helps you narrow down your choices.
Study #3: Five-second tests
When faced with a number of headlines and images, the five-second test can help us narrow them down to the top two or three. These can then be vetted with an A/B test.
Use a five-second test to narrow your headline and image choices.
UsabilityHub.com offers a five-second test. They provide a panel of subjects who view your page for just five seconds and then answer questions about what they saw.
This service doesn’t require a live page. Our landing page mock-ups will do nicely. The test takes a day or two, depending on how many test subjects you want. Ten to 25 is usually enough.
Once we’ve narrowed our copy and images down, it’s time to go to layout. Let’s call the designer.
Study #4: Eye-tracking tests
Your designer shouldn’t add or subtract anything from the page we have. We don’t need carousels, background video or parallax scrolling tricks. They should simply lay out a page that gets the visitor’s eyes to the right part of the page — the parts that keep the promise and call the visitor to take action.
How can we be sure our designer is doing a good job? It’s now cheap to do eye-tracking studies, thanks to companies like Sticky.ai.
Eye-tracking tests give amazing insight into the effectiveness of our layout.
Like UsabilityHub, Sticky will provide a panel of subjects to look at your landing pages. New technology allows the eyes to be tracked using regular webcams. So the price is much less than it has been in the past.
This study doesn’t require a finished page, so several designs can be tested. These studies take a week or two, depending on how many subjects you want to see the page.
All of the reports are gathered for you automatically, as well as subject questionnaires. I recommend testing several layouts for our landing page.
We should now have a formidable landing page, supported by four studies even before launch.
Launch and know
While a traditional page development process follows a “launch and see” approach, we have used a “launch and know” development cycle. We know this page is going to perform well because we use behavioral data to make our decisions along the way.
There is no better behavioral data than results. We want to be sure we’re collecting as much data from the page as possible.
Study #5: Analytics
Our landing page will be instrumented with a behavioral database like Google Analytics, Adobe Analytics or Mixpanel. This will collect our performance data.
Study #6: Click tracking
I also recommend implementing one of the click-tracking and scroll-tracking solutions, like Hotjar, Crazy Egg or Clicktale. This will validate that your layout is working well.
Study #7: Session recording
A session recording tool can be invaluable, especially if you’re asking the visitor to complete a multi-step signup process or purchase. Inspectlet and SessionCam are inexpensive solutions.
Session recordings allow you to watch your visitors struggle. This visitor clearly wanted to click and book a tee time.
Study #8: Thank-you page survey
Finally, we like to run a survey on the thank-you page or receipt page of our landing pages. Our question is, “What almost made you NOT take advantage of this offer?”
Then we let the visitor answer in their own words. There are many solutions on the market for doing on-site surveys, such as Hotjar, KISSmetrics, and even several A/B testing tools. The completion rates will be high, and information we get from this simple approach feeds new versions of our page for A/B testing.
Study #9: A/B tests
The Supreme Court of behavioral science has to be the A/B test. This data is randomized, anonymous, recent and behavioral and has a large sample size. The ideas you get from studies #5, #6, #7 and #8 drive new design and copy ideas. You can determine if your changes are helping with an A/B test.
Of all of these studies, the A/B test requires the most expertise and discipline. Don’t let that stop you.
Too much?
If you feel that this is too much research for just one page, then compare this to the cost of the launch-and-see approach. Unless you’re lucky, you’ll require several page launches before you have one performing at the level of our launch-and-know approach. Each is a shot in the dark without some of this cheap behavioral data.
Behavioral data lets us be more creative because it helps us prioritize and narrow our design ideas. All of these studies can be done from your computer, without leaving the office. They can typically be done for a few hundred dollars or less. Why not work a few of these studies into your next web page design?
Comments