Conversion Rate Optimization for Sites with Low Traffic

By Jenny DeGraff | May 21, 2014
More Articles by Jenny


funnelConversion rate optimization (CRO) is a simple idea: to make every webpage as effective as possible at getting customers to “convert” from one stage in the desired process to the next. You accomplish this by simply split testing different versions of a page to see which one results in the most sales, leads or other KPIs.

However, to conduct a scientific test, you need enough visitors to increase your conversion rate and enough conversions to generate good results. Most conversion optimization specialists agree there should be a minimum of 100 visitors to each landing page and at least 20 conversions for a test to achieve statistical significance (a recommended 95%).

But what if your website doesn’t quickly generate this amount of traffic and converts even less? This scenario happens to marketers quite often, and leaves many wondering if split testing can be effective or if CRO is right for their company.

The answers is: Of course it can and CRO is right for everyone. You just need a slightly different approach. Here’s a few ways that you can make CRO and testing more effective, even if you don’t have enough traffic.

Test the Page with the Most Traffic

Although you may want to start testing a page that directly impacts conversion rates, like the checkout page or request a quote page, you just may not have the traffic. Home pages typically receive the most traffic of any page on your site. Start there by testing elements that will yield results that can be used to improve other lower-traffic pages on the site (test, iterate, improve, repeat). Also keep in mind that improvements made at the beginning of the process trickle down into more traffic to your lead or sales generating webpages.

Don’t Bother with Multivariant Tests

This is pretty simple, every variation added to a test decreases the amount of traffic each page receives. This translates into increased time it would take to get a valid result. Keep it simple with an A/B test.

Make Large Changes

Because you can’t multivariant test, an alternative is to test all the elements on one challenger page. A more dramatic variation will be more likely to produce bigger differences in conversion rate, which means that you need fewer page impressions before you can announce a winner.

Test against Micro-conversions

Micro-conversions are goals, just like the sales and lead generation “bottom-line” macro-conversions you’re most likely currently optimizing for. But instead, these are the conversions that typically make up your macro-conversions. These are the smaller, lower-commitment or top-of-funnel steps, such as visits to key site pages, comments on your blog, file downloads, or adding items to a wish list. Because these goals generally have much higher conversion rates, tests optimized for these goals take less time to conclude.

Don’t Forget Secondary Dimensions

Similar to micro-conversions, secondary dimensions such as bounce rate, time on page or click-through rate can sometimes be impactful enough to optimize. While reducing your bounce rate won’t necessarily increase your sales directly, you may find that reducing bounces increases engagement and other types of conversions.

Be Patient, But Start Now

Wingify has a nice conversion calculator to help you determine how long it might take with your amount of traffic to declare a winning page. Ideally, a test shouldn’t take longer than 90 days as environmental factors, such as an upcoming holiday or market downturn, could dramatically alter the results. But as long as you are split testing with equal traffic distribution, external factors will play a roll in both page’s conversion rates, so go for it! The more you test and track your conversion rates, the more aware you will be of what makes your website most effective. The more aware you are, the more you can improve.

How have you had successes in low-traffic CRO? I’d love to hear your tips below or @JennyDeGraff.

Share this article

Share on LinkedIn Share on Twitter

Blog Search


Subscribe today!

Over 4,000 marketers are receiving actionable insights and tips for improving their digital marketing in each monthly issue - why not join them?

 

  • brianlang

    I’d argue that even setting a minimum of ~100 visits and ~20 conversions is too low; with those numbers, to reach 95% statistical significance and 80% statistical power, you’d need a ~78.5% relative lift to conversion rate, and that doesn’t account for bias from returning customers (primacy / novelty effect).

    I’m absolutely on board that CRO can always be used. I’d say separate to your recommendations, if you do struggle with low traffic, perhaps aiming for 95% statistical significance isn’t the best option, and instead one should consider lowering this threshold, as it will have a significant impact on reducing the amount of data needed to make a decision. Of course by doing so you will be able to make more decisions in general, some of which will ultimately be wrong – as a result, your ability to tweak this threshold will be dependent upon the cost of making a wrong decision.

    Thanks again for the article!

    Brian