|
Covert Personalization: Two Winning Models
|
Personalization, in the context of the Web, is all about making your
site familiar and appropriate for the visitor - both the visitor who's been
there before and the visitor who hasn't.
Overt vs. Covert Personalization
Broadly speaking, there are two kinds of personalization that need to
be distinguished: overt personalization and covert personalization. Overt
personalization takes place when a Web site asks a visitor to register,
then reuses that information to greet the visitor later - "Welcome back,
Alexis" -, or to expedite the checkout process. Overt personalization
includes saving shipping and billing information (with permission, of
course), and asking a visitor for his zip code so that you can offer him
logo merchandise for his local sports teams more prominently on the
page than that of all the other teams. Overt personalization is interesting
and necessary...and a topic for another day.
Covert personalization is a bit like the old adage about the weather,
"Everyone talks about it but no one seems to do anything about it."
In fact, that's somewhat of an overstatement. I believe that most sites
aren't making any substantial efforts at personalization, however, when
personalization is done properly, it's difficult to tell that it's going on
at all.
Amazon Has My Number
My favorite personalization anecdote took place on the Amazon site.
My husband is a historian who frequently orders books and reads
reviews about history and Church history books. Of course, I'm more
likely to pull up reviews of computer and business-technology books.
Consequently, the collection of data that Amazon has about the profile
associated with the computer we generally share to connect to the Web
is focused on those narrow fields. One day my husband brought up the
\Amazon home page to find only Church history, American history, and
computer books featured. Until he realized that they had our number, he
couldn't believe that this was what Amazon was selling.
Personalization, when done properly, works just like that. They're never
going to sell us a Harry Potter book or an Oprah's-List book, so they
shouldn't be wasting the screen real estate showing us those books.
They still usually do, but as you'll read, there are other considerations
than previous behavior that play into what goes into a personalized page.
Personalization Options
Most of the industrial content-management solutions, such as Blue Martini,
V/5 E-Business Platform by Vignette, Fatwire's Update Engine, and
BroadVision's One-to-One Enterprise include some personalization in the
content and product displays. Merchandisers or editors can indicate what
visitors will see based either on the software's own categorization of the visitors,
or on real-time click-stream analysis. For example showing a visitor that is looking
at baby swings bouncy chairs, as well.
If you have a home grown content management solution, or if you have a
commercial one but you're not satisfied with the personalization and analytics
that it provides, then you should consider implementing a personalization engine
or recommendation engine to process the wealth of data you're collecting. This
data can be converted into an improved visitor experience - with higher basket sizes,
higher closure rates, better clickthrough and closure rates on e-mail promotions,
and increased customer loyalty and long-term value.
Two Models
Two companies that are providing entirely different, but very interesting, ebusiness
solutions are Blue Martini and Net Perceptions.
Blue Martini's suite of software is used by companies to interact with their
customers over Web sites, in call centers, and through wireless devices. Blue
Martini also includes a Web store interface, a call center component, a
customer-collaboration tool (so customers can shop together), a merchandise
management component (for getting your products into the store), a customer
management component (with order history), a content management component
(for both product and non-product-specific content such as promotions, warrantees,
etc.), tools enabling business users to change and update the system and related
content, and a data-mining module for micromarketing. All in all, it's a very
impressive suite of products for serious contenders in many industries including
the retail space, financial services and manufacturing.
Net Perceptions' solution complements either a commercial or a home-grown
content-management system, with varying degrees of customized integration required,
depending on the solution you're using. Net Perceptions has a two-part solution:
analytics and collaborative filtering, which create the intelligence for a
merchandiser to make business decisions and the action for the merchandiser to
execute his wishes in real time.
Blue Martini
Blue Martini ties operations to analysis to improve customer interaction.
The micromarketing module of the Blue Martini customer-interaction system
has click-stream analysis capabilities, customer demographic capabilities,
product data, and content data, against which a business user can apply
business rules. The business rules are applied to the various data sets to
show customers the most appropriate products at the right times based
on their demographic profiles and click-stream analysis during their current
shopping session. Some of the demographic profile information is collected
during the sign-in process. In this model, the customers are clustered based
on demographic considerations, the products and content are clustered, then
the business rules (created by business users such as product merchandisers,
marketing personnel, etc.) are applied so that Blue Martini shows the appropriate
products and content to the appropriate audience.
The business rules can tell the system to show content based on who a visitor
is (demographic data), on what a visitor's looking at or has looked at
(click-stream analysis), on other factors such as merchandising (this shirt always
goes well with these pants), on day of week, or on time of day. The tools provided
permit real-time measurement of the effectiveness of marketing campaigns in increasing
average basket size, increasing closure rate, ultimately increasing the value of a
customer. More than just measurement, however, the analysis or measurement
of this information is turned back into actionable changes to the system to improve
the user's experience.
NetPerceptions
NetPerceptions takes a different approach. First, the NetPerceptions analytics
component sifts through the customer and order history files of your database,
clustering customers based on what they've purchased, and making recommendations
as to what you should be offering to each cluster based on what others in that cluster
have purchased. (No doubt this is the philosophy Amazon is taking in continuing
to show me Harry Potter books, even though I've never purchased fiction from them.)
Demographics can have an effect, but are not required. For each product-to-product
recommendation, Net Perceptions gives you a confidence level so you know what
products sell with what other specific products. A simulation of the actual combination
of product affinity and collaborative filtering then provides what rate of sales can be
expected from recommendations based on this combination of business intelligence
and action. This is useful for both cross-selling - selling batteries with a
flashlight - and for collaborative filtering to increase the order total with unrelated
products - such as offering throw pillows to a shopper who put a woman's sweater set
into her cart.
The analytics component also indicates which clusters (of customers) have the
highest long-term value, and what products are generally purchased by long-term
value customers. Therefore, in the future, you can use those products to attract
other long-term value customers. Net Perceptions looks at the profitability of both
the customer clusters and the product clusters to help make suggestions to the
merchandisers as to what offers should be made to customers. The association
rules manager allows the merchandiser to select combinations of products that
will be displayed to customers when they have selected a single product to look
at or to place in their shopping basket. Collaborative filtering can then perform
click-stream analysis (so that even new visitors will be shown the "best odds"
promotions). Then - and this is the brilliant part - the Net Perceptions Simulator
will perform calculations running the promotions the merchandiser has selected
against the current customer database to predict how well each model of
promotions would do in advance of anything being shown on the Web site.
The merchandiser then has an estimate of profitability based on the current
customer base and on each promotion or combination of recommended products.
Reporting after the fact is also available, but profitability modeling before
the fact makes this product truly unique among its peers.
Net Perceptions has found that the average basket size after implementation
is as much as 40% larger than before implementation for its merchants, and
that up to 25% of customers accept the recommendations that the
NetPerceptions engine has selected.
Personalize Or Else
The days of competing on price have created an environment where nearly
everything on the Web, from computers to cosmetics, has become
commoditized. Merchants can no longer afford to lower prices any further
to make the sale, and data from sources such as BizRate tells us that
price is not one of the top three factors in creating customer loyalty, in
any case. Personalization is one of the factors that makes a customer feel
at home in your store. The larger your inventory, the truer this is.
With your competitors only a click away, desperately offering discounts,
promotions for other companies, and loyalty points - all in a bid to avoid
being road kill on the "information superhighway" - you need to include
personalization in your arsenal of weapons to make your site stand apart
and your products irresistible.
(For more information on how to make the most of your ecommerce site,
check out the Ecommerce Guide's article
Designing Your E-Commerce Site for Service.)
Alexis D. Gutzman is an E-commerce Technology Author and Consultant
and author of The HTML 4 Bible, FrontPage 2000 Answers!, and ColdFusion
4 for Dummies. She can be reached at agutzman@internet.com
To read more articles related to e-commerce, visit The E-Commerce Guide
 |
PRODUCT OF THE WEEK |
|
|
|
Free Insider Report
Receive free monthly marketing updates, tools, tips and strategies to help your online business succeed. Get immediate access to our bonus ebooks AND 150 back issues.
|
Site Navigation Guide
TOP
Copyright © 2000-2006 by Startup Internet Marketing Services* All rights reserved
58 Porteous Ave. Kirkland Lake, ON, CA P2N 2N9 705.567.5850
This site is powered by Host4Profit.net
|
|
Free Stuff
