CHALLENGE

Philipp Postrehovsky, Marketing Manager, mogo, was looking to increase the number of loan applications his team harvested via Google for their Canadian online cash-advance brand. In the tough economic climate, there promised to be more searchers than ever looking for quick cash.

But attracting *quality* leads at a profitable rate was still a formidable task. While testing had long been a part of the brand culture, trialing strategies and tactics became an even bigger focus for Postrehovsky and his team at the start of the year.

"This was where writing and testing different types of ad copy came into play," said Postrehovsky.

CAMPAIGN

After conferring with their SEM services provider, they decided that their performance data clearly showed that some ideas -- like week-parting (i.e., targeting Wednesdays with keyword buys more than Mondays) -- were more test-worthy than others, such as day-parting (targeting certain hours).

So the team decided on three test strategies:

Test strategy #1. Weekly A/B splits on copy

Since they were focused on the Canadian site for Google, it made sense for them to concentrate on improving their click-through rate and, subsequently, their Google ‘Quality Score.’ This score measures how relevant a keyword is to ad text and to a user's search query, and affects a brand’s PPC ad position and cost-per-click.

To make those improvements, they A/B tested one ad copy concept every week and scheduled weekly analytical meetings with their SEM vendor (see hotlinks below) to capitalize on the results.

From more than 30 A/B copy tests, these were four standout examples:

- Removing the words "5-Minute App" from the search results body copy: "mogo Loans: $5 Per $100. 5-Minute App. Borrow up to $1500"

- Adding "{New}" to a headline vs. not using it

- Headline test: "Stop Using Payday Loans" vs. "mogo Loans {New}"

- Adding "mogo" to the headline vs. not using the brand name

Test strategy #2. Week-parting

Why week-parting instead of day-parting, exactly?

"We dove into the data and noticed that when we segmented conversion data by day, we observed that certain days of the week just had higher-than-average conversion rates," Postrehovsky says.

By contrast, their conversion rates didn’t differ by hour.

They knew going in that it was a relatively easy idea to test, as Google AdWords has a setting allowing marketers to vary bids based on day of the week.

- Mondays and Tuesdays had middling results. So, they used default bids for those two days, meaning they kept their keyword buys at the already established default setting. (See hotlinks for more information on terms.)

- Then, the team ramped up buys by an average of 30% on its best conversion days of Wednesday, Thursday and Friday.

- mogo’s worst days, Saturday and Sunday, were dialed down by 50% from the default bids.

"We didn’t have set daily budgets; we spent on whatever leads we could acquire within a certain cost-per-click rate," he says, adding that his company’s maximum CPL rate was $80.

Test strategy #3. Expand keyword funnel

Week-parting wouldn’t necessarily open up the lead funnel all by itself, and Postrehovsky and his team wanted to increase volume as long as the loan applications were being filled out by a qualified crowd.

But they knew that the Internet can be a confusing place for some viewers, and risked attracting consumers who were looking for money -- but not necessarily of the cash-advance or short-term loan variety.

That’s why they needed to test a series of generic keyword possibilities on a small scale before applying them to an entire account/campaign.

Here are three notable keyword trials they ran:
o "loans" in Google AdWords’ Broad Match option (see hotlinks below for definition)
o "cash loan" in Broad Match
o "bad credit" in AdWords’ Phrase Match option (see hotlinks below for definition)



RESULTS

The three test ideas were the chief drivers that helped mogo double their sales leads -- completed online loan applications -- during the last nine months.

More specifically, the conversion-to-lead rate was lifted from an average of 6% to 9% in that span of time, Postrehovsky says.

"Our highest [conversion-to-sale] rate out of any customer acquisition channel comes from paid search. After the lead is generated, conversions go anywhere from 35% to 40%. So, it definitely has been big for us."

Postrehovsky says that his team’s A/B testing improved clickthroughs and the brand’s Google Quality Score, while the lessons they learned were "core strategy to the success" when it came to increasing lead volume.

While the statistics from the A/Bs were often both intriguing and conflicting, the team re-learned the fact that every single word of copy has an impact on an ad’s success.

Here were the winners from the copy tests detailed above:

- Removing "5-minute App" increased clickthroughs by 21.4%

- Adding "{NEW}" to headline lifted clickthroughs by 30%

- "Stop Using Payday Loans" beat "mogo Loans {New}" by 21% for clickthroughs

- Adding "mogo" to the headline increased clickthroughs by a whopping 123%.

Postrehovsky adds that the weekly A/B split meetings with their SEM provider helped his team "stay on top of changes and results. [It] definitely took time, but we mapped out a process for testing that worked well for both agency and client."

Then there were the AdWords options tests to widen the funnel:

- Postrehovsky says that "loans" turned out to be a "very profitable" winner in Broad Match, while "cash loan" didn’t produce leads at an acceptable ROI.

- "Bad credit" came in a winner in the Phase Match option. This test surprised them, as they were skeptical that any generic terms would fare better than the drilled-down ones employed in the past like "Toronto payday loans."

"We are always looking to add new ways to get more leads," Postrehovsky says. "The strategy of looking at a wider [keyword group] definitely worked. Even though the keywords are higher up the funnel and not a specific product keyword, they worked really well."

Useful links related to this article

Creative Samples from mogo’s search marketing tests
http://www.marketingsherpa.com/cs/mogo/index.html

Webpage describing Google Broad Match, Phrase Match and other options:
http://www.wordstream.com/keyword-match-types

Site that explains default bids, among other Google SEM-related issues:
http://www.wordstream.com/expanded-broad-match

Google’s page explaining how ‘Quality Score’ works:
http://adwords.google.com/support/aw/bin/answer.py?hl=e
&answer=10215

Cossette Communications Group -- its Magnet Search Marketing Service orchestrated the weekly tests:
http://www.MagnetSearchMarketing.com
http://www.cossette.com/www/default.php

Mogo:
http://www.mogomoney.com