Paid search was an important element in that campaign, and testing and optimization on different elements also proved to be important to its success. Ten years after kicking off the focus on SEM, 90% of leads at 911 Restoration can be directly attributed to its SEO and PPC strategy and tactics.
This MarketingExperiments Blog post features one paid search test at 911 Restoration on its PPC landing pages. The test involved dramatically reducing the amount of copy on the landing page to discover how that impacted conversion, cost per acquisition and AdWords quality score.
According to Miri Offir, Chief Marketing Officer, 911 Restoration, “The long copy was legacy content we inherited from the SEO team several months ago. After sending paid search clicks to these old long-copy pages for a few months, we decided to test shorter, more-focused copy against our long-copy control sites.”
She added, “We incorporated standard direct response, conversion-oriented copywriting heuristics to achieve this. Essentially, authoritative brevity was our goal because logically it followed that emergency oriented calls originated from users who did not want to read additional copy and instead only wanted a contact phone number.”
With the shorter landing page copy, the team expected an increase in leads generated due to a simplified user experience.
The Control generated 34 calls on 177 clicks for $7,177.96 in spend, a 19.20% CVR for $211.12 CPA with an impression-weighted QS of 6.13.
The Treatment generated 35 calls on 133 clicks for $4,982.74 in spend, a 26.32% CVR for $142.26 CPA with an impression-weighted QS of about 5.68.
The Treatment outperformed the Control in conversion by more than 37% and reduced cost per acquisition by almost 33%.
From this test, the team learned the bots that determine quality score and the humans who actually generate leads are looking for two very different things from the landing page.
Where the bots saw less information and less authority from the shorter landing page, the humans reaching the abbreviated version found the information they were looking for more quickly and called 911 Restoration to take care of their property damage emergency restoration.
Ultimately, we got almost exactly what we expected. Longer copy can work in PPC [landing pages], depending on your goals and what you’re selling. But for our emergency services, and the users trying to source them, long copy seemed to do more harm than good.
The QS dip wasn’t entirely unexpected — looking at this in retrospect — but the reasoning for the dip was. Ultimately, while we anticipated a drop in QS, the amount and reasons for the drop were not forecasted. The results of this test, being positive, were convincing enough to warrant a corresponding change in many of our other landing pages accordingly.
In terms of future tests affected by the results of this one, we have determined that there is a good foundation for brevity in most landing pages. Due to the expense of writing the copy for these pages, our next tests will determine the minimum threshold for QS while maintaining the same or lesser ratio of cost per click for leads.
She also explained the reason for a lack of statistical confidence in the test: “These are low-volume, very tightly geotargeted accounts. While we can and do combine data across separate accounts to look for patterns, we saw and expect to continue seeing a lot of variance along strictly geographic lines.”
“We typically can’t run experiments to statistical significance because we get into seasonal variance by the time we start accumulating enough clicks,” Miri said. “With that in mind, the data that we did get on this experiment led us to take the same approach with other landing pages, which have since experienced a similar increase in leads per click cost.”
David is a Reporter for MECLABS Institute. You can follow on Twitter @DavidKOnline.
You might also like
A/B Testing: How a landing page test yielded a 6% increase in leads [MarketingSherpa case study]
Landing Page Testing and Optimization: How Intuit grew revenue per visitor 23% [MarketingSherpa case study]
Landing Page Optimization: How a navigation test increased leads 34% [More from the blogs]