The name of ResponseLogic, Inc.'s newly released product is ADAPTe. While
the "e" is probably intended to represent the expert systems technology
at the product's heart, they are sending another message by putting the
"e" after the name. While definitely an e-commerce company, ResponseLogic
has its roots and focus in business. Their technological approach is,
as President Jim Scott explains, "wrapped in the best practices of traditional
systems technology is primarily rules-based, where the marketing specialists
set the rules. An inference engine observes the behavior of shoppers and
generates recommendations for cross-sells and up-sells. With the inference
engine focusing on behavioral aspects the marketer is freed to devise
rules based on business objectives.
Expert systems technology is sometimes thought of as an older technology
than "collaborative filtering," on which many other personalization products
are based; collaborative filtering technologies attempt to identify one
customer as being part of a group with similar behaviors. Its best known
application is on sites where a shopper sees a list of recommendations
based on previous purchases, and after a purchase also sees cross-sell
recommendations of the form "others who bought this also bought "
filtering is often successful, and is part of the ResponseLogic inference
engine, but it has some notable weak spots. As CTO Dev Sainani points
out, if a merchant's database is not statistically valid collaborative
filtering may end up de-personalizing the shopping experience. Also, Dr.
Sainani notes that "everyone has had the experience of buying a gift for
a friend and then seeing recommendations for months afterward that are
based on the friend's preferences and not yours."
Response Logic's solution to the second problem is embedded in its permission-based
approach. While using large direct marketing databases of anonymous profile
information, ADAPTe displays itself to the user as a personalized shopping
assistant. It builds individualized profiles only for shoppers who request
such assistance. One of the features of the shopping assistant is that
it allows the shopper to notify it about gift shopping. So, when you are
shopping for something special for Aunt Minnie ADAPTe will not mix those
purchases with your own personal preferences; but it will remember them
the next time Aunt Minnie's birthday rolls around.
The company intends to charge for the product based on success. Pricing
will be based on the number of successful matches of shoppers to rules
by the inference engine; these represent presentations of "the right product
to the right person." In the future the company hopes to develop pricing
models that are tied to the sales lift provided by the product.
not only makes the recommendations, it also manages the content to be
displayed, tracks customer actions, and reports on customer demographics,
segmentation, and buy versus browse behavior. Like other products in this
space, ADAPTe can work with the data it collects, anonymous third party
profiles and data on individuals that comes from a company's own databases.
This is a growth market right now, and the inevitable shakeout is far
in the future. Companies like ResponseLogic are still staking out their
turf and laying foundations. In the short run they need to develop identity,
niche, and proven successes. While there are already powerhouse companies
in personalized marketing, there is plenty of room for startups to make
inroads. Perhaps more so in this area than in most areas, in fact, because
the sale is much less dependent on persuading technologists than on persuading
ResponseLogic has been forging some interesting partnerships, including
one with a high profit direct marketing agency and one with a Latin American
technology investment firm, that should give it both visibility and experience.
While specialists can make general statements about the efficacy of one
technology over another, we think there is no clear way to differentiate
without seeing how products work on your data. Any of the vendors on your
short list should be able to take a sample of your data and give you an
indication of the kinds of recommendations it would make. This is a very
competitive market, and you should accept nothing less.
vendors like ResponseLogic, where the infrastructure investment and impact
are relatively small, we believe that the IT professionals should be comfortable
in letting Marketing drive the selection process (they will anyway). What
is most important about products like these is whether the marketers can
control them easily, and whether they can see the results of their actions.
IT can provide a useful supporting role in the earlier stages by looking
for potential interface and functional problems that might not be readily
apparent to non-technical folk, and in validating the ease with which
corporate content management systems will interface with the product's.