"We were shooting in the dark trying to improve our online registrations," remembers PhoneHog President Phil Nadel.
The service, an offering from Gulfstream Internet, sounds almost too good to be true: In exchange for viewing sponsors' online offers, you'd get a calling card that lets you make long distance calls for free, for as long as you'd like.
While PhoneHog's marketing and design team certainly gave the online registration form their best effort, they also figured, how hard can it be to get visitors to sign up for an offer this good?
As predicted, results were great. More than 2,000,000 consumers registered using the online form over the course of three years.
But, no form gets 100% visitor-to-submission conversion rates online. This began to bug Nadel. He wondered, could his team yank PhoneHog's conversions any higher?
On the other hand, if he made the form too easy and compelling, would he risk generating a whole lot of accounts with lousy lifetime value?CAMPAIGN
At first the team tried a few tweaks based on their intuition. "It was very cumbersome and we were only able to test a few variations at a time," says Nadel.
But results were significant enough that he oked the expense of hiring a test technology firm to help roll out additional tests more quickly and easily. (Link to firm below.) Then starting in April 2003, PhoneHog went whole hog with registration form tests -- testing a total of more than 100-different tweaks, including:
- Right-column graphic:
Would a photo of a happy person holding a phone work better than a cartoon of a PhoneHog pig holding a phone? Or would it be best to junk the graphic altogether in favor of a bulleted list of benefits?
- Headline tests:
As advertising genius David Ogilvy said, 80% of the people seeing your ad only read the headline and not the body copy. Nadel's team wondered what headline would be so compelling visitors would keep on reading and then fill out the form below.
The team tried a half dozen headline tests in all, hoping to beat their control which was 'Let Our Advertisers Pay for Your Long Distance Calls' (link to samples below.)
- Putting the form flush-left versus centered
- Colors: Spot use of red, blue, yellow
- Whimsical cartoony-logo versus more serious-looking logo
- Links for more information versus no links
- Wording on the submission button (Example, "Join" versus "Send me a Free Calling Card")
Nadel's team tracked initial conversion winners as well as lifetime value for each test by watching the first 90-days of account activity like hawks.
He explains, "It does become quite predictable -- x% of lifetime revenue is earned in the first quarter. There's a chart we've developed. We can even do it within the first month, that gives us a good estimate, but first quarter gives us the best idea."
"What surprised me most," says Nadel, "is the seemingly inconsequential variables that have an impact on conversion rates." After testing about 100 variables over six months, Nadel's team were able to increase registration conversions by 38%, which adds up to thousands of new users. Plus, they were able to increase average registrant lifetime value by 60%.
There was a 55% difference in conversion rate between the worst performing pages and the best. However the data we found most fascinating was the fact that the absolute best-converting pages didn't always produce the best customer lifetime values.
"So far we've chosen one winner and are still testing new versions against it," says Nadel. "We didn’t choose the one with the absolute highest conversion because this one had a higher lifetime value."
"For example, the less information you ask for, the lower the lifetime value. We have found that when you ask for more fields, you have a lower conversion rate but a more engaged member, they're more responsive to it. That's the number one variable that's affected conversion one way and lifetime value the other way."
More results details:
- Tests that didn’t show much variance in results one way or the other: spot colors, and cartoony versus serious logos.
- More information equals more conversions … as long as it's a quick bulleted read and it's all on one page. The higher up on the screen that information was, the better. Plus, the bulleted list of benefits at the far-right side handily beat out the photo of a human being and the pig cartoon.
However, when PhoneHog tested making additional information available via links on the registration page, results sank. (This matches what we've heard from other landing page testers, perhaps because links distract people from the registration path- at-hand.)
Nadel says, "Instead of linking to information, it's better to put it right there."
- The actual form layout, number of questions and order of questions were the most important variants to test, producing 10.6% improvements in conversion rates. Page layout was second- most important at +7.2%. Copy order was at +6.9% and introductory copy was at +6.2%.
This means marketers should spend more time futzing with form layout and look and less on copy, colors, logos etc. This will be a radical change for many sites where the market lets the IT department slap a form under their hand-crafted copy.
- Forms centered in the middle of the page definitely outpulled forms that were lined up flush-left. This was one of the tests that surprised Nadel because he didn't expect a little change like that would make such a difference.
- The headline PhoneHog had built their business around for three years was beat hands down by another. Shows it's always worth testing your top benefit copy.
- You probably should be copywriting the actual word(s) on your submission button carefully. Don't just say "submit" or "go". PhoneHog found a comparatively long phrase that echoed the main benefit copy worked best.
Last note: you may have noticed there's one key test PhoneHog didn't do -- testing two-email boxes versus one to see if it reduced inaccurate entries. Turns out that they've already got a method that works well to make sure emails are correct.
The first thing every new registrant sees is a grey formal- looking pop-up box asking them to confirm their email is correct. Works like a charm. (See sample at link below.)Useful links related to this article:
Samples of several PhoneHog tests and the email address checker pop-up
Optimost, the technology firm that enabled PhoneHog to run so many tests fairly easily in such a short time: http://www.optimost.com
PhoneHog current registration form