Great – we’ve persuaded you that split-testing can improve conversions on your landing page. Let’s get down to the specifics of how you go about doing that.
Effective split-testing does not involve randomly changing things on the landing page and seeing what happens. Instead, you need a hypothesis about what is working and what isn’t, so you can develop a test strategy that will yield useful results.
Start with a clear understanding of exactly what you want visitors to do: Give you a valid email address? Make a purchase? Download an app? There’s no point in testing anything until you know what “X” – a successful result – is.
Next, take a careful look at the landing page as it exists now, and ask yourself two questions:
- If I arrived at this page, what about it would persuade me to do X?
- If I arrived at this page, what about it would give me an excuse to not do X?
Specifically, you need to figure out what a visitor needs to know before proceeding. What information does he or she need, and in what order, to be persuaded to do X? “Information” here might be product specifications, a value proposition, or the terms of a particular offer.
Keeping those questions in mind, analyze each element on the page to see whether or not it moves the visitor toward conversion. Is all the information the visitor needs on the page? Are there elements that don’t move the visitor into the sales funnel?
Use the results of this analysis to develop a hypothesis about what isn’t working on your landing page. Examples of such a hypothesis are:
- “The page doesn’t clearly spell out how the customer will benefit from accepting the offer. A bullet list will clarify the value proposition and encourage visitors to do X.”
- “There’s no clear call to action. Placing a large button with a succinctly worded CTA directly next to the list of benefits will make visitors more likely to do X.”
- “The product photo is too static. A photo or video of someone using our product will more effectively communicate the benefits customers can expect if they do X.”
Once you’ve got a hypothesis, you’re in a position to develop a series of A/B tests to determine the specific changes you should make to optimize the effectiveness of your page.