We
are in the midst of a repricing revolution.
All around us software prices are in a state of change.
OLD
SW PRICING MODELS
In
the old days we had four ways to buy software:
·
Some
software, especially operating systems and utilities, came bundled
with its hardware. We
thought of it as free, although, of course, its cost was simply
included in the system price.
·
Desktop
software was sold in shrink-wrapped boxes.
It had both a list prices (rarely paid) and discounted
prices, which depended on the software’s popularity, its category,
and its competition. If
you bought lots of desktop software, you could get a steeper
discount, especially if you were willing to buy it naked (with only
one or a few physical copies of the software and its documentation,
the assumption being that organizations were looking for control and
downloaded desktop software over their Local Area Networks
(LAN’s).
·
Business
software (for businesses and computers of nearly any size) was
considered a more serious matter.
It generally cost thousands of dollars (millions of dollars
for large packages in large companies to run on big computers with
many users). Buyers
paid for the software, then paid an annual maintenance fee (perhaps
15-20%) for patches, enhancements, and support.
Customization was priced and purchased separately (and could
cost many times more than the software itself).
All
of these models had three things in common:
(1)
Customers bought the software and paid for it.
(2)
Customers owned the software.
(Actually they owned a license giving them the right to use
the software, but that amounted to the same thing.)
(3)
Customers could keep using the software as long as they
wanted to, often far longer than software vendors had in mind.
When entreaties and bribes (new features and discounts)
failed to lure users to upgraded versions of a software product,
ISV’s frequently resorted to threats (ending support; making new
versions but not old ones compatible with important new products and
standards).
A
fourth software distribution and pricing model, Shareware, was based
on the idea that people wouldn’t buy software unless they could
try it first – and that, anyway, software shouldn’t cost very
much, if anything. Shareware
vendors were often very small companies – sometimes just an
avocation for a bright programmer.
Software might be distributed in plastic baggies, with little
or no documentation and support could be nearly invisible.
Nevertheless, shareware was very popular in certain powerful
computing communities, notably among university students and
hackers, and is probably the origin of the Open Source movement.
CHANGING
MARKET NEEDS
All
of those models are still present in the marketplace, but they are
becoming less relevant over time.
New models have emerged, based on:
--
The ubiquitous connectivity of the Internet
--
An overwhelming desire to leverage that connectivity by
creating cross-company,
cross-platform, cross-application, interactions and transactions
between employees, contractors, suppliers, business partners,
prospects, and customers, all with high degrees of security and
accuracy
--
An interest in preserving investments in existing software
--
A desire to outsource hosting and supporting software
applications both to decrease
implementation time and to reduce in-house resource
requirements
--
A desire on the part of vendors to smooth their revenue curve
and improve the number of
customers who upgrade to each release of new software
--
The move to continue to provide customized software while
avoiding all or much of the
cost and time of a full-blown development project via the use
of new tools, new programming techniques, and integrating new
and existing software via web services
This
means vendors are now trying to figure out how to simultaneously
preserve or grow revenue while satisfying customers who are trying
to contain or decrease costs. There
are some very lively pricing discussions going on.
NEW
SOFTWARE PRICING MODELS
One
thread of this discussion says just create marketplaces where
sellers offer goods and buyers bid for them.
This auction market will set the market clearing price.
You will note that very few business transactions, except in
commodity types of markets (or for obsolete or excess material and
equipment) is actually using this price setting mechanism.
That is because (a) pricing is either uncertain (or no
minimums are set) or (b) not market clearing (if minimums are set).
Almost no software is sold this way today.
Another
suggests that software be rented rather than sold. (Here, I’m
talking not about an ASP service, where you’re renting the USE of
the software, together with the processing power to run it and its
support, but rather the rental of the software itself.)
I think I’ll be boring and insist on pointing out that, in
the olden days, we did that. It only works when you’re selling relatively large amounts
of software to relatively small numbers of companies or the cost of
negotiating and administering the pricing and collection
arrangements exceeds the value of the software itself.
In fact, this pricing arrangement has never entirely
disappeared from the market and has always been available to large
customers.
But
how else should we look at Microsoft’s recent interest in changing
its arrangements with its corporate customers and moving as many of
them as it can to three year contracts where they are essentially
renting the use of the software and its upgrades during that time,
with the need to negotiate a new agreement (for, Microsoft hopes,
the newest release of the software) at each contract turn.
Maybe it’s just part of Microsoft’s moving to an
Enterprise Software Vendor model (a major initiative for the company
this year), and not part of anything else at all.
The
ASP model can be another kind of rental arrangement, but one where
the customer is presumably renting not just software, but the
facilities and support of its use. In fact, for many types of software products, especially
those which may require significant customization, ASP customers
must sign a contract which is in essence (or in actuality) a sale of
the software to them although the software will be loaded onto an
ASP’s servers and run for them at his remote site.
In
a Web Services model, different participants in a business
transaction will each bring their parts of their existing
applications and software, however acquired, and join them via new
software, probably purchased, but conceivably (especially for
certain special functions) provided by ASP’s or other Service
Providers (SP’s).
Meanwhile,
software itself is becoming increasingly componentized so that it is
becoming possible, in modern development environments, to purchase
tasks that would once have been coded for each particular
application as standalone modules, to be combined with much less
time and expense. Even
if the modules require customization, a remarkable variety of
intellectual property exchanges and services are popping up to
provide such work.
There
is the thought that eventually components will work so smoothly
(we’ve only been talking about this for five or ten years now,
with a wistful look on our face) that they can be offered to users
over the net, within and outside of applications, for quick and easy
access. That’s what
Microsoft has in mind as part of its .NET initiative.
Some parts of its newest Office XP product already feature
web accessed additional goodies. How many users will ever get to the
point where they’re often adding new features to their software
remains at question and in the future.
Remember, we have lots of research that suggests that desktop
users rarely use more than 10% of the function in their productivity
packages.
And
then there is the Open Source movement with its promise of – at
least – free operating systems and some free software and the
confusion it creates in some people’s minds as to whether it
threatens the future of all paid-for software in the future.
The
Open Source model of Linux permits a level of inspection and
interaction that permits bugs to be found more quickly and fixed by
the best available minds and hands.
It’s a strange miracle, but it seems to work.
Is it infinitely scalable?
We don’t know, but we’re clearly going to find out.
I’d
doubt that it represents a threat to anything but the notion that
you can sell bad code by refusing to let anyone see what’s in the
box. Increasingly,
customers are going to demand more access and more information about
what they’re buying at the technical level and more ease of use
and stability at every level, but especially for users.
Great
code that appeals to people because of what it does will always
sell, regardless of the terms and conditions or the price (within
reason). Linux and Open
Source shouldn’t threaten really good software at all.
