Up, Up & Away: Simple Changes To Delta.com Boost Online Revenue
Marketing News
by Deborah L. Vence, Contributing Editor
September 15th, 2007
THOUGH MILLIONS OF customers visit Delta.com every day to research and book flights, about 50% of customers were dropping out before completing the booking process. Moreover, they were dropping out at a stage where they already had seen the airline's pricing and made their flight selections. When they left Deltas site, then, the door was open for competing sites to gain their business instead of Delta. The airline needed to find a way to streamline its Web site and stop this trend in its tracks.
The overall objective was to allow customers to book flights as efficiently as possible. "We wanted more of the people who attempted to shop on the site to buy tickets as a result of having had a better experience," says Brian DAmato, general manager of strategy and measurement for Delta.
With this goal, Delta was swimming upstream against the current. The fact is, consumers continue to conduct more travel research online than book flights online, according to a July report by The Consumer Internet Barometer, produced by The Conference Board Inc., a global business research and membership organization in New York, and TNS (Taylor Nelson Sofres plc), a global market insight and information group in London.
For instance, the report indicated that 16.8% of men and 18.8% of women used the Internet more in the second quarter of this year than in the same period in 2006 to book travel arrangements, while 20.8% of men and 23.4% of women used the Internet more to research travel but not book it. Some of that growth has slowed from 2006, when 17.8% of men and 16.2% of women stated that they booked more travel plans online than in 2005, while 22.0% of men and 24.9% of women researched more travel online.
To help increase online travel bookings, in mid-2006 the multibillion-dollar airline hired New York-based Optimost, a company that specializes in online multivariable testing techniques for landing page optimization.
"Delta does a tremendous amount of volume online. In the travel industry, there are a significant proportion of bookings online. They wanted to figure out with the booking process, how to streamline and improve usability so a higher percentage [of people] would purchase a ticket," says Mark Wachen, CEO of Optimost.
Optimost helps companies improve their conversion rates on Web sites. Its Web page testing technology enables clients to create and test unlimited versions of copy, images, offers, layouts, forms and colors on a Web site. It creates various versions of the pages and puts them up for live traffic to interact with on a site. Then Optimost keeps track of the results.
Optimost determines which values, or page variations, to test. In general, Optimost's methodology enables a company to evaluate hundreds, thousands or even millions of possibilities of a single Web page. By creating versions of the page that mix and match the different page variations, Optimost can test a much smaller subset of the page variations and get predictable data as to how much each individual page variation is contributing to performance. Through this, Optimost is able to figure out the optimal Web page configuration that will maximize companies' online conversion rates based on a subset of thousands or millions of possibilities.
In other words, while A/B testing, or split testing, enables a company to find the best solution out of two possible Web pages, Optimost's multivariate testing solution enables a company to find the best page out of 2 million Web pages.
Implementing Optimost's technology for Web page testing involves a few lines of JavaScript code to replace the area of the Web page to be tested. Optimost has the browser provide the optimized portion of a Web page as the page is viewed, which can happen as often as hundreds of thousands of times per day. Optimost determines which variations of the content to display and responds with more JavaScript code instructing the browser to deliver that content. In this way, each visitor to the Web page sees a different version of the content.
Optimost has been conducting an ongoing analysis of Delta's Web site using a Web page test plan to analyze millions of page variations at one time to see which ones work best.
The airline's online process to book a flight comprises five steps: Enter flight information, select the flight, enter personal information, verify information and purchase, and finally, confirm the flight. Analysis of the site traffic had revealed that Delta.com visitors were dropping off after reaching the page that requires them to enter their personal informationóthat is, after they had seen pricing and made their flight selections.
"It's a journey, not a destination. It's something that you are continually doing. We built a template [for Delta,] and have millions of versions [of their Web pages,]" Wachen says. "With Delta, we've been impressed with their appreciation for their key customer touch points. It's not just about the experience [flying with Delta,] but the experience on the Web site, too."
In Delta's case, changes to the Web site have been made only to certain pages, rather than the entire site in order to make sure that the changes actually work. Modifications have been made to Deltas passenger information and its SkyMiles program enrollment pages, DAmato says, adding that Optimost will continue to test more pages on Delta's site this year.
During the testing process, which officially began in late October 2006, Optimost determined, for example, that when customers reached the passenger information page, which requires customers to fill in the fields with their name, address, phone number and such, a high percentage of customers dropped out before completing the page.
After testing millions of variations on that particular page, Optimost determined that simply removing the instructional copy located at the top of the page prompted more customers to start and finish filling in their information. The original instructional copy stated: "In order to purchase your ticket(s), we need to process your passenger and credit card payment information. It is not necessary for all passengers to have a SkyMiles Number or to login in to complete a purchase. However passengers who would like to take advantage of SkyMiles benefits and earn miles for award travel should input their SkyMiles number below or log in at left. If necessary you can enroll now. Special guidelines and restrictions apply when your child travels alone. Learn More."
Says DAmato, "The thing that we learned was that on the page where [people] input passenger information, the instructions [included] at the top of the page actually hurt conversions."
Wachen adds, "Extra instructional copy at the top of the page made it look more daunting. Oftentimes, when people see copy they assume that they need to read it, even if they really don't have to. If someone is in a rush or is lazy, they may be turned off just by the mere sight of seeing copy that they think they are going to have to spend the time reading."
Though it's not exactly clear why visitors would be affected by the instructions--and thus fail to book their flights--"from our tests it has been pretty consistent that removing instructions from the page helps people navigate the site better," DAmato adds.
Moreover, on the SkyMiles enrollment page, Delta eliminated the instructions at the top and modified some of the wording on the page. For instance, instead of using the word, "Go," located at the bottom of the page, Optimost changed it to read, "Click here to finish" the enrollment process. In addition, Optimost changed the tagline at the top of the page to read, "Join SkyMiles. It takes just minutes," rather than "SkyMiles enrollment." In this case, removing the instructions and changing the wording together were more effective in enticing site visitors to join the program.
Wachen explains that Optimost determined the results for the Web pages by isolating the impact of the extra copy on response rates. When the instructional copy was present, the conversion rate was lower.
"It's true that people could just choose to ignore the [instructional] copy and surely some people do. However, on balance, when the copy was present, the response rate was lower," Wachen says.
Though it hasn't even been a year yet, just those minor changes are yielding promising results. DAmato says that at the end of calendar year 2007, Delta expects to bring in $10 million to $15 million in incremental online revenue, compared with the year earlier. He attributes the increase in part to Optimost's tests and the subtle design modifications Optimost implemented on the Web site, an investment that has cost the airline less than $100,000.
"[This has been] a wise investment to really take the site to the next level. Simple things like taking instructions out, making the [site] easier to use--that is real money for the minor changes we're making," DAmato says.
Meanwhile, though DAmato was not at liberty to reveal specific figures on conversions, he says that conversion rates to actual ticket sales have improved by 5%.
"Optimost's tool gives you the confidence that you are going to reap the data," he adds. "We think we're going to continue to reap rewards. With more use, we expect [Optimost] will find other opportunities to tweak and improve [the site] even more.
Find this article here
