Quick question: how much conversion improvement is enough? Would hitting your industry’s average conversion rate convince you to rest on your laurels? If you doubled your old performance, would you stop there? Hopefully not. This case study is an example of why you should never consider your conversion work done, even after impressive gains.
Case Study Summary
Client: Ifbyphone, Inc., a provider of automated telephone applications.
Overall project goal: Increase the quantity and quality of their phone-call and click-to-call leads.
- Complex offering
- Poor prior experiences with landing page optimization
- Wide variety of audiences
- Difficulty in attributing phone calls to a marketing source
Success measurement #1: Percentage of landing page visitors who contact Ifbyphone by phone.
Success measurement #2: Lead quality, as measured by the phone call length (longer calls were correlated with higher close rates).
2008: The Original Landing Page
The original landing pages had a number of issues restricting their success. The most egregious problems, though, resulted from having poor hierarchy (both visual and informational) and too much clutter.
Original landing page
Top issues with the original landing page
- It’s very hard to find the offer. Compared to other elements, the offer description has the weakest visual treatment, and doesn’t have much chance of standing out on the page.
- It’s hard to understand the offer. As often happens with complex or technical offers, the messaging was quite information-dense and difficult to scan. Over time, the company’s messaging became simpler, clearer, and more refined. This had a profoundly positive impact on not only their landing pages but other aspects of their marketing as well.
- The primary call to action (“call to speak with us”) is overwhelmed by other elements. For example, the bright red right-column area was a persistent CMS element on every page. It was visually very strong, which distracted attention. It also offered multiple alternate actions the visitor could take, which was confusing.
- Long scrolling content doesn’t repeat the main call to action. While a long scrolling page (see below) isn’t necessarily a conversion-killer, it requires special handling that this page failed to include. For example, at the very bottom of the page there were three links, but they led to completely different service pages. Once a visitor scrolled past the initial view, there were no instances of the main call to action at all.
The long, scrolling page:
Behind the scenes: a PPC campaign rebuild
While we reviewed the landing pages, another key portion of the project was being completed—a rebuild of the company’s PPC campaigns. By our evaluation, the campaigns were wasting almost 60% of spend, driving poor-quality traffic and missing many opportunities to reach the right audiences. We slated this cleanup first because it was easy to fix. Just as important, though, we wanted to ensure our landing page test would run against the right audience.
Late 2008 – 2009: Landing Page #2
With the PPC cleanup complete, we launched into the landing page redesign. Designs are the result of multiple small decisions, and this one was no different. For example, in an early draft we included a large, futuristic phone graphic in the upper left that became a visual anchor for the entire page layout. Later, through a series of requested revisions, we replaced the hi-tech phone with the company’s “branded” phone image, the old-style black phone you see in the design above. As you’ll see later, for landing page #3 we would be given a freer design hand.
Landing page #2:
Key changes in this design
- The offer is much easier to see. While the black phone is a strong visual element, it offers no alternative action paths. The large red headline and gray background behind the features area call attention to the offer. Other page elements are balanced so they don’t overwhelm.
- The offer is easier to understand. We went through several content drafts with the client to achieve two things: an initial message “roll-up” that was easy to scan, and description blocks that had clear headlines and spoke in a more natural, less technical tone.
- The primary call to action stands out more clearly. We included a “click to call” call to action element at the page top, and the direct-call phone number within the benefits area. Later, we would add even more emphasis to the direct-call number.
- The call to action is repeated several times. As you can see below, the overall page length was shorter than the original. Even so, we repeated the click-to-call and direct-call call to actions so at least one was visible at all scroll points.
New scrolling page with multiple calls to action
A/B test setup
Once the content and design for all the new landing pages was approved, we launched our first test sequence. In this phase we ran four separate A/B tests concurrently for different products. The “A” pages all used the original template, while the “B” pages all used the new template and rewritten content. Both versions received PPC traffic from the newly updated campaigns.
The desired action for this project was a phone contact—either via a direct phone call or the click-to-call tool. In order to measure success, we took a number of steps:
- Ifbyphone assigned a unique toll-free phone number to each page. Coupled with the clients’ call-tracking capability, we had a reliable measure of each page’s call count.
- For measuring call quality, we used Ifbyphone’s tie-in to Google Analytics. This Ifbyphone feature automatically posted a virtual pageview to Google Analytics whenever a phone call reached a certain length. The pageview URLs we chose included landing page and phone number data, to make the results easy to interpret. With this data we could calculate the ratio of total call count to high-quality calls for each page.
- For fun and a proxy: Click-to-calls were measured by the Ifbyphone system, but just for fun we chose to track them in Google AdWords as well. Since AdWords didn’t yet have the capability to import Analytics Goals, we had to come up with something else. Using the “call me!” click action as a trigger, we loaded the AdWords tracking pixel with an AJAX call. This allowed AdWords conversion tracking without having to refresh the page or load a new URL. We now had a proxy metric to help us cross-check data and optimize the AdWords campaigns.
A/B test results
The A/B test results showed the new design was a clear winner. Measured across all four tests, the call per visit percentage more than doubled, total call count increased 62% and average call quality (as measured by the length of the call) nearly doubled. This could have been a place to stop, kick back and bask in the victory. But we didn’t.
The A/B test let us verify our thinking and make the case for landing page testing to the client management and board. We moved next to a multivariate test, to try out some element variations that hadn’t made the initial cut. A couple of the results were surprising, and we got to develop some fun, geeky tracking judo I’d love to go over. But in the interest article length—and because we saw only incremental conversion gains, I’m going to save that for another time.
Results for PPC rebuild + landing page redesign: 3X conversions, 2.3X lead quality
After the project: more work
Once this first project was completed, Ifbyphone took the new landing page approach and spread the love by building it out across their remaining products. They also took a step that many companies miss—they extended their landing page learnings into the rest of their web site, refining their messaging and design throughout.
January 2010: Landing Page #3
In early 2010, Ifbyphone came back to us for another landing page redesign, saying, “we think we can still do better.”
A few things had changed:
- The competitive environment had changed, with an increasing number of players on the field. This, by itself, was a compelling reason for a redesign.
- The company’s messaging and product offerings had been through multiple refinements. This let us concentrate on pure visual treatment for the page.
- Also, the design restraints were removed. We could bring in new colors, imagery—basically do what we thought best from a conversion standpoint.
Here’s the new template we came up with:
And, as executed with real content (for a different product), by the Ifbyphone team:
- Colors and imagery. We extended the color palette for greater versatility, and added a strong human face as a visual focal point.
- More compact display of the content. In this version, the detailed content was provided in expansion areas, made visible when a “more” link was clicked:
- Multiple presentations of the calls to action. We repeated the direct-call call to action no less than 3 times, and the click-to-call twice. Each had slight variations in the wording and visual treatment.
- Further clarification of the “click-to-call” call to action. We’d refined this particular element for the previous landing page, but suspected that some visitors would still not understand what it did. So, we took this opportunity to make it even clearer.
- Inclusion of a video. We’d tested videos in the multivariate test, but they didn’t perform well as a main content focus. We still thought they had value, though, so for this redesign we included them at a smaller size in the right column.
A/B test results: an additional 2X conversion improvement
The new design was A/B tested on a single product, and outperformed the prior page by a little over 2X—from 4.66% to 9.78%. Lead quality remained stable.
The extended project results: 6X conversion improvement for Ifbyphone’s campaign.
Overall Key Takeaways
- Conversion optimization is a long-term initiative, not a one-off checkbox item.
- For best results, pay at least as much attention to your messaging as to your layout and design.
- Sometimes tweaks just aren’t enough. Dramatic changes can yield dramatic results.
Can we stop now?
This extended project provided a number of valuable conversion lessons in messaging, content presentation, tracking and design. But the most important conversion takeaway from it all is this: Don’t stop. You’re not done yet.
Opinions expressed in the article are those of the guest author and not necessarily Search Engine Land.