The A&E store team has the unenviable task of running three stores in one, supporting the network's main brands: A&E Network, The History Channel, and Biography.

To add another layer of complexity, their customer base is a diverse and vocal lot, covering everyone from Thunderbird fanatics to history buffs.Though conversion rates were good by industry standards in 2003, the team had two nagging doubts about the store; was a tri-branded site confusing visitors? And were they losing sales because the storefront couldn't serve all brands and customer segments equally?

According to a lab-based usability project, the answers were yes and yes. 

Here are four steps So Young Park and her team took to tackle the problem.

-> Step #1. Introduce dynamic storefronts

Park says, "In the past, if you went to the store, whether you typed in shophistorychannel.com or shopbiography.com, you'd get the same storefront and it was tri-branded with A&E Network, The History Channel and Biography logos at the top.

"Now visitors still reach a single store - with a multi-brand navigational menu and product database - but the branding and content on the front page depends on the URL they followed.For example, clicking on the "store" link at the Biography site gets you to the store and access to all 5,500 products from across the various brands. But the store displays Biography-only branding (logos) and the products and promotions featured on the front page also only reflect the Biography brand and offerings.

o Brand consistency:

Park says matching the logos to visitor expectations isn't really about immediate conversions..."People are more concerned about the show, series or product...they don't come into the store and say, 'I'm going to shop for an A&E title'."

"But from a brand perspective, we do want to be consistent. It was important to negate some of the confusion around being in a (single brand) content site and then coming to a tri-branded store..."

o Better integration with TV:

Since storefront content is now brand-specific, it's easier for Park's team to do more integrated promotions with on-air marketing. She explains, "We might host a sweepstakes in the store for The History Channel on-air. It's much easier to do these brand promotions on the network side when we have the ability to have "different stores", even though we're all one site."

Dynamic storefronts also let the "store" better promote viewership, as much as sales. Park says, "The different networks are premiering big shows in the same week. It's hard for one storefront that has a generic look to be able to communicate three different premieres on three networks."

She adds, "Consistency of messaging that supports viewership as well as product sales - that's very valuable."

o Targeted content and promotions:

"Whenever someone sees something that airs and they want to buy it, they can go to the store. But you can't on one storefront possibly display what you think someone might be most likely to buy."

Dynamic storefronts means better targeting..."so if you're coming from A&E, what's displayed to you on the storefront is going to be A&E specific. The rest of the store is still merchandised smartly, but not necessarily by brand."

Do dynamic storefronts impact on online promotions? Park says they just have to be more careful about plugging in the right landing pages for brand-based keyword buys (PPCSE buys drive 10% of the store's revenues).

-> Step #2. Modify site architecture and navigationThe usability tests also found some problems with site design, so Park's team took the opportunity to introduce some layout tweaks into the new "stores." Here are a couple of the things they learned...

o Some people ignore Javascript dropdowns:

Previously, the store menu had a "shop by show" link with a dropdown menu list of all the relevant shows. Park says, "People would sometimes not notice the dropdown. They weren't noticing it, were not clicking on it and were not shopping that way."Now clicking on the "shop by show" link takes you to an actual page with an alphabetical show directory.

o More people ignore graphics:

The previous storefront featured a huge center-page graphic highlighting, for example, a key premiere-related product line. It's a common tactic on many online stores.Park notes, "People came to the storefront and would completely ignore the thing. And we were really surprised because if you look at it objectively, no way could you miss it, right?"

So why did they? "...because they think they're advertisements," says Park.The same applied for smaller navigational graphics representing product categories in the menu. "People were spending time internally creating them and making them look nice - it was a big waste of time.

"The new storefront has barely any graphics that aren't actual product images. Park adds, "It's amazing what comes out when you see real people using your site in the context that they would if they were really shopping."

She warns that usability tests have to involve people using the site in the most natural way possible..."not a fake setting where you sit someone down in front of a computer, start them at your store and say, 'can you find Pride and Prejudice?' They go to search, type in 'Pride and Prejudice', and you go, 'Great, my site worked fine'."

-> Step #3. Start collecting customer profile dataPark's team want to extend the dynamic content concept, so that website and email content alike automatically adjusts to the profile of the visitor.

Customers can now, for example, select up to 22 genres of interest when they open an account. And opt-ins to the email newsletter can also select their preferred TV brand and media (DVD or video) during registration.

Park's already sending targeted emails out based on brand preference. Unsurprisingly, they perform much better than the generic tri-branded store newsletter, converting at over 10% and making a big contribution to the 12% of store revenues driven by the email list.

-> Step #4. Test, test, test

Park's team are using the dynamic storefronts to fine tune their understanding of effective promotional offers, copy and presentation through a self-styled "marketing deathmatch."

Essentially this means A/B testing to "figure out what sorts of creative and messaging work better in which situations and to which segment of people."

Park explains, "Let's say we do an A/B test on a piece of creative that normally resides in a certain spot on a page and we keep the offer identical, and the messaging identical, but we make the creative different (one's blue, one's green). 

We run it for a while and see that green is really performing much better in that spot."

"But let's say we try that same green creative on a different page with a different segment of people - it doesn't perform so well. Over time you start to build up information like people build up a profile about a customer. In our case, we're building up a profile around different spots on a page and different ways that we display offers or products."

And how have all these changes and initiatives worked out? Park says, "We saw a conversion increase of over 50% on our branded storefronts versus the default."

She concludes, "About a year ago, it was 'personalization is dead.' What we're not trying to do is one-to-one personalization - it's hard to manage and costly. But with the right technology, there are smart ways you can segment and do things more dynamically."

Note: A&E is a member of Shop.org, a forum for retailing online executives to share information, lessons-learned, new perspectives, insights and intelligence. 

Useful links related to the article:

http://www.shop.org 

http://ShopAETV.com (A&E Network storefront) 

http://ShopBiography.com (Biography storefront) 

http://ShopHistoryChannel.com (The History Channel storefront) 

http://store.aetv.com/html/home/index.jhtml (actual store)