November 03, 2003
Since August 1st, the marketing team at Fredericks have been running loads of tests online so they can roll out the winners during the 2003 holiday rush. Find out what they've learned in our new Case Study.
Includes a super-clever landing page idea you should steal, and interesting notes on search marketing results.
From the latest Charlie's Angels movie to the hit TV
spy show Alias, fabulous, burlesque-style, fashion is the look of
the day (or rather, night). Where better to shop for it, but
from the company that pioneered the look?
The problem is, while young women across America yearn to be in
style, it doesn't always occur to them to go to fredericks.com.
Although Fredericks of Hollywood has been online since 1996, when
Online Marketing Director Jennifer Bedolla joined the Company
last November, she knew the site wasn't nearly visible enough.
"The Hollywood glamorous pin-up bombshell, the retro look, is
back in fashion now. We're a 60-year old company - it's where it
all started. I wanted to drive more unique visitors to the site.
It's the perfect opportunity for us."
Before making a major investment in outgoing campaigns,
the marketing team first set a solid foundation by prepping in
Prep #1. Increasing site effectiveness
While fredericks.com had come a long way since 1996, it was so
ultra-clean and uncluttered that Bedolla feared many shoppers
might not find what they wanted. So, she had the Web team add
many more calls to action, navigation options, cross-selling
offers, category pages listing many products, and even a
personalized wedding registry. (Link below to 1996, 2002, and
2003 home pages.)
Prep #2. Setting firm email rules
Given the sort of merchandise Fredericks offers, Bedolla worried
about being mistaken for a junk emailer. Her primary email goal,
therefore, was to protect and enhance the brand. Sales were
secondary. To that end she:
o Ruthlessly cut many affiliates, "We went through a clean
sweep if we found them misrepresenting the site or doing
anything they shouldn't in email."
written in reassuring terms both in the privacy section and
on the page where visitors can sign up for the list.
There's no legalese, no loopholes, no doubletalk.
o Required that the contracts for all lists she rented to
promote fredericks.com included a promise that the list
owner could provide a time-stamped IP address for each
name, should the need arise to prove that any individual
had given permission.
o Redesigned the regular promotions to the house list to
focus more on the brand and glamour and less on just a buy-
now type promotion.
Prep #3. Creating an online/offline campaign calendar
The marketing team built one mega-calendar to show when various
in-store, catalog and online promotions were occurring. Their
goal was to run as many concurrent campaigns across all media as
possible to strengthen impact.
(Note: This sounds like an obvious idea, but as we learned at
last month's Shop.org conference, only a small percent of
online/offline retailers have created such a calendar.)
Next step: Testing campaign tactics
Although Frederick's fiscal year started August 1st, Bedolla
carefully reigned in spending her new budget. The business is an
extremely seasonal one, so she wanted to concentrate spending
later in the year when it would make the most impact.
However, she immediately launched a series of small test
campaigns to learn what would work best, so she could roll out
the winners later. Tests included:
-> PPC (pay per click) campaigns on many major and minor
networks, including AOL Shops, MSN Shops, Shopping.com (formerly
Dealtime), Google, Overture, Kanoodle, LookSmart, etc. When
possible, tests included broad versus narrow search terms.
Although they mainly sent automated feeds to these systems, the
team chose to never invest in a PPC campaign without also
dedicating staffers who could spend daily time massaging and
analyzing campaigns for better results.
-> Banner ad campaigns on demographically targeted Yahoo
mailbox pages, and on about 10 handpicked sites that could
promise their visitors were predominantly women over 20.
(The media buyer notes she did not use ad-network buys because
she thinks truly successful online ad campaigns can't be totally
automated. "The most important thing is to get to know each site
really well, who else has advertised for a long period of time
there, and spend time molding and shaping the campaign. You may
get clicks on network buys, but it’s maybe misleading, they might
not be converting.")
The team tested many banner creative sizes, as well as a variety
of models, and offers.
-> Requiring double-opt-in to join the house list and then
tracking the lifetime value of those names versus single
permission names on the house file.
-> Landing page tests for search results links, including
displaying a single product, displaying a list of similar
products (for example, one pair of black hosiery versus the black
-> Landing page tests for gift-with-purchase offer banner and
email campaigns, where the first action the visitor was asked to
do was to select their size and place the gift in their cart …
and then move on to shopping for enough items in the rest of the
store to qualify to receive the gift.
Now just finishing three months of tests, Bedolla says she's already found some solid winners that will make a
big difference as she moves forward to invest in holiday season
marketing. "You learn a lot from testing. I'm a big advocate of
She was able to share the following results:
- Cross-sales offers on the product pages have helped bump
average order sizes significantly. For example, about 40% of
Halloween-related promotional buyers would up springing for
accessories such as wigs and shoes to go with their outfits.
- The online wedding registry/wish list has been a solid success.
"We're seeing the average order size jump in the wish lists.
People are registering for holidays, birthdays. People are
buying multiple items."
Although the viral element is wonderful from friends and family
shopping wish lists, interestingly, "we saw a large increase in
people buying it themselves. A week or two later they come back
and make the purchase."
- Coordinating email and catalog creative and offers results in a
lift for both. For maximum results, Bedolla tries to mail an
email "sneak peek" about a week before a new print catalog drops.
- Bedolla has noticed very different patterns not only in the
buying habits of shoppers using different search and shopping
systems, but also differences in conversions for search terms and
seasonal results changes.
"People use search differently in the summer months. A lot of
people are shopping clearance. In the holidays, people know what
they are looking for, versus looking for a sale rack."
"Shopping.com and Overture do amazingly well for us. My
perception is that Shopping.com users are more blue collar.
Overture customers tend to be more specific - they know what they
are looking for and use more descriptive terms in their search.
In smaller search engines, they type in more general, vague
However buying broad keywords on big search engines such as
Google can be a money-suck with lots of clicks that don’t
convert. "There are so many men who want to look, but don’t
convert." So Fredericks has backed off from those big broad
terms on Google for now - including its name.
- Having dedicated people manage important PPC feeds has made a
significant difference as well.
Senior Account Exec Jill Nakayama explains, "I check the feed
every day - 95% of companies just do the automated feed. I
noticed as I was updating AOL Shops promotions that there was a
huge block of white space where you put in text." She used the
space to insert extra text about Frederick's offers. "It's very
labor intensive, but it's what's really making the difference in
- Regarding testing levels of email permission, Bedolla says, "We
did see a drop off in names with double, but from early results,
the doubles provide a more targeted individual and made them a
Are doubles so much stronger that it's worth the drop-off?
Bedolla says it's too early to tell, but given pending email
legislation, she's continuing to test and track closely.
- Banner tests revealed that the fuller-figured model
consistently beat the thinner model, perhaps because most
American women are not all that thin. Inspired, the team
launched an entirely new campaign specifically talking about
fredericks.com's plus-sizes offerings. It was a solid success.
- The category landing pages consistently beat single-product-
focused pages for most search campaigns.
- The biggest landing page winner was the gift-with-purchase test
that asked visitors to add the gift to their cart first and then
start shopping. Now Bedolla's team are testing more variations
on the theme (how much should the offer be worth, how much should
visitors have to buy to get it, etc.)
Useful links related to this story:
Samples of home pages and banner tests:
411web Interactive - the agency that handles much of Bedolla's
online advertising. (She asked us to say, "They're wonderful.")