[Zope] RE: [ZCommerce] RE: Philip leaves Arsdigita (was: Re:[Zope]kerberos ? + LDAP + ecommerce + ZEO replication etc)

[Walter Ludwick]

Richard:  YES!  It is *precisely* this ability to "transform a 'customer'
into a 'member'" that defines the OpenACS magic that i seek.

Chris:  Did that answer your question?  Because, i don't know if what's
needed is just "an interface that mimics the tcl interface  to the Oracle/
Postgres tables that make up the ecommerce application " (sounds simplistic
to me, given the deep magic that Richard has so aptly described - but what
do i know?).  I don't know this is "core" or "apps," or whose job it
*should* be.  I just know that it would make for a pretty radical (i.e.
"from the roots") change in what the Zope proposition has to offer, and that
this sort of change simply won't happen without *strong*  "encouragement and
direction" from DC.

Michael:  I think you're correct in that all anyone's asked DC for in this
thread is a proper analysis of the opportunity.  Of course we'd all like to
see this new magic enabled without having to do anything ourselves, but i
think we've seen ample evidence here already that a fair number of people
are willing to give generously of their time and expertise to the cause.  It
is indeed hard to do such ambitious things without a cohesive vision for
what is to be accomplished (i *almost* feel that these occasional
penetrating questions from Chris and Paul might get us to that cohesive
vision, if only the whole process could be made a bit smoother and faster,
but...  :-)

At any rate, i am certainly grateful to Albert for both the "mind-bomb" idea
(to borrow a Winer-ism) and the solid technical information he has provided
as support.  This is indeed an interesting thread, which, as Jason pointed
out, really puts the vagueness of this eCommerce term under the spotlight of
critical scrutiny.  This is the most vital discussion i have yet seen on the
list since subscribing many months ago (guess i should have said something,
but...) Clearly, eCommerce means different things to different people, and
if nothing else, this thread seems to be moving us toward a common
definition - or framework for definition - which might help us to have more
focussed conversations within the subject area henceforth (something that
has been done to good effect on the CMS list, another that i have been
lurking).

Albert, you are right:  i oversimplified eCommmerce requirements horribly by
saying that its all about "Personalization."  Still, you seem to understood
what i mean by it (tho i barely understand 40% of what you say;  the rest is
greek to me, alas), so hopefully my use of shorthand is forgiven.

To expand on what eCommerce means to me, let me first orient myself within
Jason's "e-context" classification scheme, which strikes me as both easily
understandable, and instructive.  My enterprise context is an amalgam of
Jason's cases A and B, which is to say that we are
    -> [A] "Large well established business [existing market infrastructure]
wanting to move some or all of their non-web business into a web-based
system.. sales management admin.. total content and management"
HOWEVER, given our morass of legacy challenges, and the pressing nature of
competition from the [C] class you've defined (i.e. eBusiness pure-plays),
we find ourselves having to work within the interrim context of:
    -> [B] "Similar well established business who wants to start a new
spinoff web-based e-business. Basically marketing cataloging sales etc.."
While this may seem a misfit case according to Jason's scheme, i'm actually
finding it to be quite common, in the wake of the "DotCommie" invasion
(though most of the serial entrepreneurs responsible for this carnage have
moved on, the separation of eBiz from traditional value streams remains to
be integrated).  To put a finer point on it, we make our money on services,
not merchandise;  specifically, services to MEMBERS - kinda like health
clubs, only not so facility-centric (most of the member services we provide
are information-oriented).  The most familiar analogy i can provide would be
the Community College, oriented to carrying the Continuing Education focus
thru to its logical conclusion, which is cultivating customer relationships
of the Lifelong Learning kind.

Perhaps this "e-context" helps to explain why i see Personalization as lying
so close to the heart of eCommerce.  If it sounds like a niche interest,
then perhaps i've not explained it properly (although Jerry does seem to
grok the shape of this opportunity zone i am driving at):  it's about
knowing thy customer, and building enduring customer relationships (member,
subscriber, other) on the basis of ever-increasing customer knowledge, and
using it to discover ever-better ways of demonstrating care and respect for
the person.  The most powerful way to do this is by saving the customer
time.  As Jerry put it:  "Wouldn't the frequent flyer be tickled if the site
noticed after a few visits that VisitorX always hits the weather page, and
offers to add a link to save a few clicks?  Making this sort of thing
painless to implement will get everyone closer to the goal."  Of course
Jerry is right.  In fact, the bar in the eCommerce world today has been
raised (largely on the back of Doubleclick.com and the advertising budgets
that they and their kind have commanded) to at least this minimum level.
But this kind of "personalization" (pretty easily deliverable thru the kind
of collaborative filtering services that Charlie describes) is essentially
superficial, not fundamental.

Fundamentals are what enables the kind of interactions that prompts Doc
Searls and pals on the Cluetrain to say "Markets are Conversations," such as
the following (all pages ripped straight out of the Amazon playbook):
    -> AFFINITY PURCHASE PROMOTION:  At the risk of igniting a whole debate
about patent abuse (but let's not go there), the "one-click" purchasing
model was certainly a ground-breaking innovation (whether precedented and/or
obvious, or not) brought to the marketplace by Amazon.  What makes this such
a money machine for Amazon is their ability to suggest follow-on purchases
(i.e. "Would you like some fries with that?") out of their own catalogue,
not someone else's.  What makes it feel personalized to me is the way Amazon
asserts responsible - while profitable - stewardship over the data i leave
with them, as opposed to giving that over to some clickstream-counting
purveyor of eyeballs.
    -> ALLIANCE-BUILDING:  Obviously Amazon had a huge supply-side challenge
to address, which they could never do without opening their information
architecture (or parts of it) to publishers, allowing them to move large
volumes of product thru at a higher overall profit than thru the bookseller
trade.  On the sales side, their Amazon Affiliate program was another
ground-breaking innovation, appearing as early as it did in the history of
eCommerce, and giving all those unprofitable websites out there a chance to
make a bit of $.  What makes this feel personalized to me is that Amazon is
not sitting there expecting me to fall into their orbit;  instead, they are
somehow extending their offerings into many of the other spaces i frequent
on the web, and giving me useful information (and purchase opportunities)
in-context.
    -> CUSTOMER COMMUNITY-BUILDING:  Philip Greenspun likes to call this
"social shopping" (note the word order within this string), whereby user A
can comment on a page and/or write a reviw, making this information
available in-context to user B, who can then click thru to all the other
content contributed by user A (including personal data, if desired -- even a
photo!  Now that's personalization!), thereby scratching the two customer
itches of pre-purchase research, and social networking, at the same time.
(All of these, as PhilipG has pointed out, are features of the ARSDigita
Community System that Amazon has custom-built over half a decade.)  It is
interesting for me to reflect on the fact that the ratio of books i actually
buy at Amazon to reviews (mostly submitted by peers, not pros) that i read
there has got to be 1:75, at least. If my useage pattern is at all typical,
does this make them more of Community and/or (information) Content site than
an online merchandiser?  I don't know, but i'm inclined to think it's their
uncanny ability to be all of this at the same time that got them where they
are today.

As to the technologies it takes to enable all this, again i cannot say, but
it does appear rather obvious to me that none of this stuff is merely
grafted-on to the Amazon website, but is instead built into their
enterprise/ information architecture at a deep level -- futhermore requiring
dynamic exchange between numerous transactional data stores (running the
gamut of customers, products, orders, credit card approvals, packing slips,
etc.), residing in RDBMS's in at least 3 different places (website,
financial service provider, and warehouse).   And to do all this in a way
that feels "Personal" is... wow.  Fundamentally challenging, let us say.

Sorry to be dwelling at such length on Amazon;  they just seem to have had
so much to do with defining expectations about "personalization" in the
eCommerce space.  In fact it is very hard to talk about eCommerce at all
without also talking about Amazon -- just as it is very hard to talk about
the benefits of any given architecture for web development without also
talking about how it supports eCommerce.  Can't be all things to all people,
of course;  at the same time, you must recognize that popular expectations
become reality, and that's what you've got to deal with.  There is certainly
a significant opportunity here, but also a threat of becoming irrelevant in
the conversation that is this marketspace, without that compelling
"eCommerce story" (which Paul has already acknowledged needing) to tell.

Without presuming too much, i would like to propose a simple SWOT analysis
for Zope's place in this problem space, which might run something like this
(just to throw something quick-and-dirty out there for spitballing
purposes):
    -> STRENGTH:  Along with other strenths cited in my last message (good
licensing terms, architecture, accessibility, "good people," etc.), Zope is
about Content management with a capital "C" for catalogued, we all seem to
agree.  Searchability of every object in the system is built-in, RDF and
Dublin Core are deeply supported, and (as PaulE and KenM pointed out in the
OSS4Lib interview already referenced several times) metadata is generated in
Zope quite naturally, especially when content is developed in the "WikiWay"
that lies at the heart of all zope.org development processes. (Paul, i must
say that i think your focussed approach to cultivating these particular
strengths - while letting the larger community make what it will of Zope -
is nothing short of brilliant.)
    -> WEAKNESS:  Optimized as it is for reading, the ZODB can't handle the
sort of CRUD transaction volume that goes with the turf of eCommerce, and is
the proper domain of RDBMS, where Oracle is clearly king (our legacy systems
are written to the Oracle back-end, and there are no open API's.  We look at
the prospect of migration pragmatically;  of course we'd like the idea of
free Postgresql licenses, but must look at total cost of ownership for a
*complete* solution.  Until you have that, let's talk about your API's.),
and where Zope doesn't yet have much to offer.
    -> OPPORTUNITY:  The market for delivering online content (as opposed to
merchandise orders requiring fulfillment) in a profitable way.  As i see it,
the answer here lies not in treating items of information as SKU's (i.e.
information merchandising -- a path that meets its ultimate end down the
dark alley of micropayments;  with all due respect to Nielsen and
Negroponte, i share Clay Shirkey's distaste for this), but rather in support
for a finely-nuanced subscription model, most commonly found in learning
communities with a strong knowledge content component.  EDUCATION (where
Python is emerging as a favorite first language in CompSci courses;  check
postings of the last few months on the edu-sig list
[http://www.python.org/sigs/edu-sig/] for abundant evidence) suggests itself
in many ways as the perfect vertical market for Zope;  within this realm,
the burgeoning market for online continuing-ed offerings (where the focus is
more on what works than on academic formality) may be the ideal entry point.
    -> THREAT:  In the online education industry, WebCT is the king of
courseware-centric solutions, but it is oriented more to teachers and the
institutions that employ them, not curious web-surfers looking for
information, who might be open to "infectious involvement" in communities of
learning/ practice.  A solid community-hosting solution -- open source,
preferably, in this realm -- with effective eCommerce hooks built-in, could
prove a formidable competitor.  Does that sound like OpenACS to you?
Admittedly they are weak in the problem space of sophisticated content
management (i.e. metadata management, RDF, Dublin Core, etc.), but watch
them climb that learning-curve as this WorldBank Gateway project (which at
its heart is a metadata exchange challenge) kicks into high gear.

Do i beg the question too much?  Of course i'm shooting this off the cuff
from my own subjective viewpoint, and any SWOT analysis worth its salt must
be derived from as many perspectives as you can possibly engage.  At any
rate, i hope that there's enough merit in my argument (supports are
available on request;  i'm just tired of footnoting at the moment, busy as i
am with writing another academic paper, as it happens :-(

    
    |/|/alt
    Walter Ludwick
    wgl4@student.open.ac.uk (MA-ODE, IET, OU-UK)
    
    
PS:  Jason, i sense you are right that, beyond the strategic marketing
survey i have proposed, looking at ACS as a requirements document may be the
logical next step.  I also agree with Albert (i think you said this too)
that, by analyzing what would be involved in a port, you can hardly go wrong
- as long, i would add, as this analysis is done *in partnership* with
developers and business solution-buyers committed to the goal of bringing
these desired functionalities into the Zope world one way or another (not to
leave Paul out there on that limb all by himself).  This would be a
decidedly different approach from the one that ARS Digita has become know
for (i.e. first order of business is:  client coughs up a minimum of US$1m,
giving ARS Digita engineers at 6 mo's [or more, at cost-plus] to learn the
business before client can expect to see anything in production), something
i like to think DC would be more open-minded about -- which is why i'm still
(until Monday at least;-) thinking in terms of 'plugging ACS goodness into
Zope,' essentially, and not the reverse.  |/|/