Internet

CASE STUDY

Three Steps to Maximize Search Campaign Efficiency

The explosive growth of pay-per-click (PPC) search advertising has created an US$8.2 billion industry led by Google, according to JupiterResearch. More than 60 percent of companies recently surveyed by JupiterResearch stated they intend to increase their spending on search advertising.

This growth in search spending means increased competition for select keywords, which is reflected in higher costs-per-click for most search advertisers. You may be feeling this pain now, watching your search ROI (return on investment) decline and contemplating a reduction in keyword purchases as a result.

Don’t act rashly — it’s probably unnecessary and counter-productive. If you want to get the most out of your search advertising, here’s what you need to do:

  • Step 1: Manage your keyword strategies
  • Step 2: Don’t ignore your landing pages
  • Step 3: Employ multivariate testing and optimization

Step 1: SEM and Beyond

When confronted by increased keyword costs and declining or flat returns, search advertisers have learned to re-examine their keyword buys and try new text ads. A whole industry, search engine marketing (SEM), has grown up to support this need to improve search campaign performance. Among the leaders are Efficient Frontier, SearchRev, iProspect and Reprise Media. Visit SEMPO for a complete list. These companies are experts at managing keyword strategies and improving the traffic arriving at your Web site via search. This should be your first call, especially if you don’t have an expert search analyst on staff equipped with a good keyword bid management tool.

A common mistake made by many advertisers is to stop here, however, and limit attention to just the front-end of the sales funnel, where the SEM companies concentrate. They ignore or pay only passing attention to the actual Web pages upon which search traffic lands and travels on its way to a conversion event. The whole point of generating the traffic is to generate a conversion event like a completed sign-up or sale, right? Beyond SEM lies the real payoff — the conversion event.

Step 2: Don’t Ignore your Landing Pages

Optimized landing pages play a critical role in every efficient, top-performing search campaign. Your second area of focus is making sure you deliver your search traffic to better designed landing pages.

You and your design team can start immediately finding ways to improve your landing page conversions. Introduce your designers to such helpful sites as Marketing Sherpa and Marketing Experiments to learn some basic landing page design best practices.

Designing better landing pages involves many considerations not always used in traditional Web design. Push your design team to think like your users and employ some well-established best practices.

Step 3: Employ Multivariate Testing and Optimization

The next step on your path to maximizing the efficiency of your search campaigns is testing and tuning your landing pages for optimal performance. It’s one thing to follow some standardized best practices for good landing page design; testing the many thousands of possible page variables to determine the optimum will take your search campaigns to the next level.

You may already be doing some form of A/B or split testing of Web pages. If you are, you’ve probably realized some improvement in conversions, but you know the process can take too long and require lots of resources. Further, A/B and split tests only tell you which of a couple pages tested may rank best — not what makes a page perform best.

There’s a better way. Multivariate testing provides a scientific approach to testing many, even hundreds of thousands of, possible page variations very quickly to determine what mix of page factors is most likely to generate customer conversions. It also reveals what makes a page perform best, so lessons can be applied to other pages and projects within your network.

Your Customers Tell You What’s Best

You can seat the smartest marketing, design and advertising people you have around a table and ask them what landing page design is best. You’ll have as many opinions as there are people in the room. However, the fact is they are still just educated guesses.

The power of testing, especially multivariate testing, is the ability to present the vast array of your educated guesses to real users of your Web site and quickly determine which combination is best. Web site optimization using multivariate testing will enable you to modify key areas of a Web page — experimenting with multiple versions of an offer, image and copy, for example — to determine the best-performing combination under actual user conditions. The results can be surprising, with previously untried ideas often performing best.

Case in Point

SmartSheet.com, a Bellevue, Wash., company, provides a great example of what can be accomplished by leveraging live search traffic to optimize landing page performance. The results include increased conversion rates and higher search campaign efficiency, along with significantly more new customers.

SmartSheet.com provides online task management and collaboration software to companies and individuals, serving more than 10,000 teams worldwide. The company uses pay-per-click search advertising extensively, employing numerous keyword and text ad campaigns through the major search engines, principally Google. Specific landing page templates are linked to particular keywords and customer template needs (e.g. Project Management, Performance Reviews).

Figure 1 shows the original landing page presented to users seeking help with Employee Performance Reviews.

(click on image to enlarge)

While SmartSheet.com’s search activities are well run and conversions are relatively high by industry standards, the marketing team suspected performance could improve further. SmartSheet.com worked with Widemile, a Seattle-based digital content optimization company, to optimize the performance of its key landing pages. Below is a description of the project.

Testing and Optimization Process

First, a comprehensive assessment and evaluation of the search traffic and existing page was done. It’s important to imagine the users as individuals, real human beings, and understand how they interact with the page so as to design new elements to which they are most likely to respond favorably. Continuity of user experience from keyword and text ad through the landing page to the point of conversion is essential.

The project’s second step involved designing a new page template, incorporating key best practice design lessons learned from experience. In this case, changes included refining the action path, reducing page clutter, shrinking the product shot and adding a customer quote, among others. This new template was tested against the original to establish a new baseline.

While critically important first steps, the above-described changes alone cannot deliver the optimal page. The team is still just guessing as to what elements may perform best. A disciplined testing and analytics process is needed to finish the job and reveal the best performing combination of factors.

A test was designed to answer some fundamental questions tied to page performance, including:

  1. What text is best for the call-to-action button?
  2. What font and text size should be used?
  3. What customer quote is best?
  4. What descriptive copy is best?
  5. What offer text is best?
  6. Is the four-step process outline helpful?

Multiple variations for each factor were created and presented by Widemile’s testing platform to actual users following the prescribed testing process. Page load times and site performance are not impacted by the test process. The user is not aware a test is being conducted. The precise combination of factors which make the optimal page was determined in a little over a week from start of the test.

(click on image to enlarge)

Over 1,000 possible page configurations were evaluated as part of this test, although just 16 different page experiments were presented. This was a relatively simple test. In some situations, possible page combinations can number several hundred thousand or more.

What’s Next?

In addition to learning which combination of factors makes for the optimal page, SmartSheet.com learned the relative importance of each factor, which helps explain what makes one page perform better than another. Knowing how each factor and the individual variations influence conversion, the company is able to run subsequent tests to further increase conversions.

For instance, in this case, SmartSheet.com learned that Factor 1, the call-to-action button, had over six times the influence on conversions as other factors. The company is focusing subsequent optimization efforts on testing alternatives for the higher influencing factors. Other next steps include testing and targeting specific customer segments, such as certain traffic characteristics, including keywords and text ads, and other key performance indicators.

More Customers, Same Search Spend

This project determined the optimal combination of page factors for SmartSheet.com’s best performing landing pages and provided valuable lessons to guide further tests and advance other marketing initiatives within the company. By putting into production just one of the newly optimized landing pages, SmartSheet.com more than doubled its return on ad spend — meaning that for the same PPC search spend, it’s now receiving twice the number of new customers.

Multivariate testing and landing page optimization is still a relatively new discipline, but rapidly becoming an essential part of top performing search campaigns. Don’t let rising keyword competition ruin your day or spoil your search campaign returns. Follow the three critical steps outlined above to increased search campaign efficiency and know you’ve fully optimized your entire sales funnel for maximum returns.


Robert Bergquist is president and CEO of Widemile, a Web content optimization company based in Seattle that provides multivariate testing technology and related creative and professional services.


Leave a Comment

Please sign in to post or reply to a comment. New users create a free account.

Technewsworld Channels