May 01, 2008
How To
SUMMARY:
With Mother’s Day, Father’s Day and graduations approaching, that surge in Web traffic is too important to waste with a bad landing page. But testing during this time can be dangerous. Done wrong, you’ll waste one of the most profitable times of the year.
See how one marketer optimized a landing page to deliver a 35.4% higher conversion rate during one of their highest annual peaks in site traffic.
|
|
Many consumer marketers depend on traffic surges before a holiday period. Some might not survive without those traffic booms, which makes marketers uneasy about testing holiday landing pages. What if something goes wrong?
The experience of Sam Ee, Art Director and Principal Designer, MIVA Direct Inc., and his team proved this fear unfounded for their ALOT toolbar applet brand. The truth is that consumer motivation is different during the holidays, so your landing pages need to be different, too. Otherwise, you’ll end up wasting conversions.
Ee and his team ran a Thanksgiving campaign to get consumers to download a free recipe toolbar. During the campaign, they used more than 30 paid search ads relating to holiday recipes linked to a landing page offering the toolbar.
By looking at site data from the previous year, Ee knew Web traffic for holiday recipes would explode just before Thanksgiving. He wanted the page to capitalize on this traffic, so his team ran a multivariate test in the weeks before, while the traffic surge was still building. The page he started with was optimized several weeks earlier for a different consumer.
“It was actually a good time for us to try this multivariate testing, because we know that the holiday time is popular in terms of recipes,” says Ee.
10.62% of visitors to the old landing page downloaded and installed the toolbar. After testing, an optimized page was placed about 11 days before the surge, seeing a 14.73% conversion rate leading up to it and a 16.43% rate during the highest spikes.
Testing landing pages around the holidays can be effective, but you must have a plan. An optimized page has to be placed well before a traffic surge to avoid devastating your conversion rate.
But don’t be afraid. Here are five steps to safely optimize a holiday landing page:
-> Step #1. Study past traffic
First, study last year’s site traffic around the holiday you’re targeting. If you’re like Ee, you’ll notice that the graphs are a bell curve with a sharp point where the surge peaks.
Make note of when the traffic starts, when it surges and when it peaks. It’s key to test while your traffic slowly builds and use an optimized page during the surge and the peak.
At the peak, Ee’s surge delivered three to four times more traffic than normal.
-> Step #2. Start with a strong landing page
Take a look at landing pages you’ve already tested (if you haven't tested anything, start now!) and use one of your strongest performers. If you have a few options, or any doubts, don’t be afraid to test them again.
Before the Thanksgiving holiday, Ee and his team tested an outside system that automatically monitors websites and drives traffic to the landing pages that deliver the highest conversion rates. Next, they tested four landing pages for themselves to make sure they were getting the correct answer.
The page they found to have the highest conversion rate included an image of chocolate cake (see hotlinks below). The outside system, however, had found that the strongest performer had an image of pizza. After looking deeper into the figures, Ee’s team noticed that the outside system sent less traffic to the chocolate cake page, which skewed the results. So, here it proved beneficial to test their landing pages themselves.
Once you’ve tested your options, select the page with the highest conversion rate. This will give you a strong base to start testing from. You can only go up from here.
-> Step #3. Select the elements to test
Next, select portions of your landing page to test. The number of elements will depend on your resources and technical ability. The more you test, the stronger the page will be.
Some of the elements that Ee’s team tested:
o Main image
o Color of the call-to-action button
o Footer text describing the benefits of the toolbar
o Other text about the tool bar
o Title text at the top of the page
Every test element had at least three variations, which resulted in 1,024 possible landing pages. Ee and his team narrowed this number to 16 pages using proprietary multivariate testing software.
Don’t feel that you have to test hundreds of pages. Testing even a few elements will help raise your conversion rate, Ee says.
-> Step #4. Run the test and pick a winner
If you picture the graph of your holiday traffic as a bell curve, you want to be testing just as the curve begins. This will allow you to target consumers with a holiday frame of mind without risking a test during intense traffic.
If you prefer, you can test during the surge, but that could be risky. It’s safer to have your optimized page in place beforehand.
To target the Thanksgiving surge, Ee and his team ran multivariate testing between Oct. 24 and Nov. 7. The resulting optimized page featured an image of a turkey and was placed on Nov. 7, long before the big jump in traffic on Nov. 18-Nov. 20. Then, traffic began tapering off around Nov. 23.
Ee’s team ran the test for two full weeks to be able to compare the weeks and make sure any spikes in the conversion rates weren’t anomalies.
Note: When you’re testing, don’t change the landing page’s sources of traffic or it may skew your data.
-> Step #5. Save everything for next year
Save all of your data from the campaign for next year. This will help you know what to expect, and you can start the process over again by conducting more tests on the same landing page.
Useful links related to this article
Creative samples from ALOT's optimization test:
http://www.marketingsherpa.com/cs/alot/study.html
Widemile - helped with optimization:
http://widemile.com/
ALOT:
http://www.alot.com
MIVA Direct Inc.:
http://mivadirect.com
