Niche Vendors
How do niche vendors survive in this market place? Answer: rather
well, in some cases.
Introduction
I'm turning more of my attention to the best-in-breed vendors these
days, because I think new and quite workable business models are
evolving for them.
In this piece, I want to talk about some of those models and why
they work for some and might work for others. In some cases, companies
that B2B Analysts, Inc., follows quite closely are moving towards
these models; I'll cover companies like i2 (and SSA) in later issues.
Best-in-breed has seemed pretty hopeless the last few years. You
can see the proof in downtown San Francisco, or even here in Cambridge,
in all the For Rent signs and the shadows on doors where gold lettering
once stood.
It's time to realize, though, that the rules of dot-bomb-ery forced
people to make bad decisions about how software companies should be structured and what
their arc of success should be.
The Rules
We all know those rules.
- A software company must dominate its market.
- The company must prove its dominance by experiencing stupendous
growth. (The pattern follows what my old colleague, Brian Sommer,
calls "The Witch's Hat," flat at first, then sharply upward, then
flat, then just as sharply downward.)
- Margins must be very high, and these are achieved by selling
licenses and getting others to run the lower-margin services businesses.
- The software must be based on a technology (during the dot-bomb,
it was an http interface, silly as that now sounds) that renders
competitors obsolete and gives it an overwhelming advantage in
the marketplace because purchasing the software is a natural consequence
of a technology upgrade cycle.
Following these rules gave rise to a business strategy that runs
something like this:
- Find a large area of business (like "buying" or "selling") that
needs automation and tag the market for the automation with a TLA
(three letter acronym).
- Announce that you have achieved market dominance
in this area.
- Develop software based on new technology.
- Invest heavily
in industry influencers. (We would never call this, "bribe the analysts.")
-
Use the resulting publicity and FUD (fear, uncertainty, and doubt)
to sell as many people as possible Hope that the software you deliver
works.
- Wait to become the next Microsoft.
Good as it sounds, it didn't work.
Most of the TLA areas were way too small to yield
venture capitalists a 10x return on their investment. Buyers in
those areas displayed too much common sense to generate high growth
rates. Where there really was some market interest, the price of entry
turned out to be too low. The big platform players (SAP, PSFT, SEBL, ORCL)
immediately announced
an equivalent product and used the same marketing dollars (analysts
are easy) to establish their "leadership." Their announcements
froze the market.
It turned out to be a mug's game. Either you pick a bad area and die of asphyxiation.
Or, as soon as you begin to succeed, you're strangled by the platform
players.
Playing the Mug's Game
But what if you change the rules? I think you can build a company
that is quite, quite successful: profitable, growing, and, at the same
time, establishing a very respectable annuity.
During the dot-bomb, everyone said that software was
a great business because
you could print as many CDs as you want, so your margins were terrific.
I say that
the great thing about software is that
you can sell it at list and next year sell it again
at 17% of list (or, if you're PeopleSoft, 20% of list). And next year.
And next year.
If your rules exploit this wonderful fact about software,
rather than the fact that you can print CDs quickly, I think
you can do better. So I'm suggesting some anti-rules, rules that break the dot-bomb
rules.
- Build software that deals with a business problem, not an area
of business. Hint: if you can describe
it with a TLA, it's probably not a business problem.
- Design for repeatability, but deliver custom
solutions.
- Make long-term profitability, not growth, your primary goal.
If you follow these anti-rules, there are several different
ways you can build a best-in-breed company.
Grab an Architectural Edge
This is the best strategy, because if it works, it is unassailable.
Basically, you look at a business problem that
"standard" application companies simply can't solve, because
their architecture is inadequate. You create an application that
solves the problem by using a special-purpose design.
Then you find customers who need to have
that problem solved.
This strategy can work because the platform players grew originally
by underbuilding and overselling. In the process they looked at
a lot of specialized problems and said, "Let's not bother."
Their reasoning was as follows:
- "Solving these problems takes too much time, which slows our
time to market, and it takes a lot of development talent, which
is in short supply."
- "A more complex product will be harder to maintain,
install, and troubleshoot, because it does more."
- "We can't sell enough of this. The potential customers anount to only
a segment of our total installed base."
This reasoning left oppportunities behind. Not golden opportunities,
unfortunately, because in one sense, the applications companies were right.
With the standard architecture, it was too expensive
to go after them.
But with a different architecture, you have a chance.
You have to do two things. First, you choose to be prosperous
rather than Ellison-rich, so you can give yourself the time you
need with the market. Second, the architecture has to establish
a dominating advantage in the market.
Exhibit A for this kind of company is Interface Software.
Interface is a CRM company that basically does contact management
for certain special kinds of clients. For those clients and only
for those clients, they have an edge.
Their market consists of companies that do relationship selling: law firms,
investment banks, potentially insurance agencies and many others.
Companies that do relationship selling need to be able to record and manage
a much richer body of information about their clients than can be provided
by the standard contact management applications, from ACT to SAP.
It's routine, for instance, in this market for a company to notify an employee,
A, when
B comes in as C's client (C is another employee), "Hey, B belongs to your country club;
why don't you drop him a line." With Interface, they can actually do that
automatically.
How do they do it? Most contact management systems model contacts using what
I call the
Holy Trinity--Contact, Opportunity, and Account. This model
lets them say that contact A belongs
to account B, which has indicated a desire to buy C. Interface has
added a fourth element: Relationship. Their model allows them to say
that contact A used to work for Person D, who might be an influencer
for opportunity C. A general-purpose
CRM package would never, ever model using this fourth element, because
it complicates set-up and maintenance. But companies that live or
die by relationships are willing to put in the effort to set up
something like this and maintain it. For these companies--and there
are a lot of them--Interface is their only reasonable contact
management option.
Interface is profitable, takes in (our estimate) $17 million
a year, and over the past few years has sold over half the
big law firms in the country. They have no effective competition
in this niche. Now, they're moving into other
segments including venture capital and consulting. They expect (hope) to be up to $30 million in one or
two years.
I consider this success.
A lot of companies that try to establish an architectural
edge don't succeed, of course.
Sometimes
their architectural advantage isn't significant (Click Commerce); sometimes
they've got an architectural advantage, but the natural customers are hard to find
and convince. (A company like ChannelWave, also in the PRM
space, has this problem.) Sometimes the architecture just doesn't solve
the problem. (There are any number of examples.)
But if you can solve the problem, create a barrier, and find the
customers, you may grow slowly, but you'll have a nice business.
The Custom Solution Company
Another kind of best-in-breed player competes effectively in the established
application areas because it can provide a higher-quality solution
than the platform players, even when the platform players have
roughly similar applications.
These "custom solution" companies usually provide applications to
large companies, who
consider the solution to be strategic. These large companies want and
can afford a highly customized solution, one that works exactly the
way they want. For a software project, they simply
expect there to be significant process change, a fair amount of time,
and some serious integration and modification. What they
pay the best-in-breed player for is, essentially, a promise of success.
A sign that you're dealing with an effective custom solution company
is that they'll tell you
their biggest problem in a sales cycle is not competition from the platform player, but the
internal IT organization, who want to build the application themselves
(though sometimes on top of the framework that the platform
player provides).
The best-in-breed companies succeed by showing that these IT
organizations are wrong, essentially by providing so much
specialized software, services, and knowhow that it becomes
clear that they are better than IT at putting the solution together.
Very often,
time pressure is involved. The company needs need a specialized solution
today, not tomorrow, and they don't have time to acquire
the knowhow or build scratch.
Typically, the best-in-breed companies that can compete in this
way have three things:
- A software "engine" that seems very hard to build.
- Knowhow about a hard technical problem that even business school
graduates know is hard, like operations research,
or optimization algorithms.
- A track record of delivering highly scalable solutions.
Some of the companies who now play successfully in this way
started out as TLA companies, and some still nourish dreams
of ravishing success. Such companies typically overvalue their
software and undervalue their knowhow and consulting capabilities;
they regard themselves as forced into delivering custom solutions.
In my view, they should embrace this path, recognizing that
they've built up some formidable intellectual property over
the years. When they do embrace the situation, they recognize
that their optimum strategy is to
maximize the revenue from every engagement you do
get. If they don't, they will typically lay off a lot of the
work on consulting partners.
Several companies
are succeeding with this model. A poster child right now might be Ascential, which
makes ETL (extraction, translation, loading)
software. An ETL tool is a necessary enabler for a data
warehousing project; Ascential's software edge is that it
can extract and cleanse extremely large volumes of data.
Ascential is typically called in when
a large company embarks on a big data warehousing
project. The budget for these things will be $3-4 million, and
the Ascential solution will get about 1/3 of that in software and
services.
Companies in this space are in the software business--don't get
me wrong. They wouldn't exist if they didn't have software
that has been proven to work under demanding conditions. But
the software edge is always a little treacherous. To maintain it,
they end up on a treadmill.
The large players are always building (or at least announcing)
the features that these custom companies been using to differentiate themselves. The
custom companies must continually respond, either by adding more features or by
proving that the large players still haven't
produced a truly scalable application that will be useful
in these large projects.
In Ascential's case, this is indeed happening. Many big players (SAP, Siebel, to
name just two) have announced ETL tools of their own that sound a good
deal like what they provide. So far, the differences are still clear.
Ascential is willing to cede low-end projects where the differences
in capability matter less, because they can add features that make the large projects go even faster
and more cost-effectively.
If the large players do succeed in catching up, it's a problem.
The custom solution company will often go into a stall when the
market perceives they've lost their edge.
Sometimes, they do indeed die. But often enough,
the market realizes that the edge is still there. The
pendulum swings back; the people who
tried out the platform players' software return to the custom solution
players, and a new cycle starts.
The TLA Company
A third possibility for a best-in-breed vendor is to
follow the time-honored recipe and invent an entirely
new market that one can dominate.
These days, you see more TLA companies in hardware and infrastructure
areas, but there are still occasionally new products on the
application side.
Very occasionally.
It's hard to make a go of this for the reasons that we've described
above. Most of the time, the area isn't "real," and if it is,
it often takes so much time and effort to teach the potential
customers that you go broke. If, in spite of all this, you prove
the market, then you have to cope with the platform players.
The software exception that perhaps proves this rule is Ariba. Ariba
has been selling ESM (enterprise spend management) the last couple of
years, and
I think they may actually make a go of it. They seem to have established
the space and trained the customers, and despite fierce competition from
the large players, they still seem to have a significant edge in the space.
They had a couple of advantages. First of all, ESM, or what B2B
calls "sourcing," really is an area of business that is poorly automated
now and can do with some software support. Second, they have quite
a bit of cash and some other products to sell while they try to
educate the market. And third, they their products are pretty good,
outstripping even the best-in-breed competitors.
Another TLA company that is succeeding, though it's in the
early stages, is Jim Green's
new software venture, Composite Software. They serve something
they call the EII (Enterprise Information Integration) space. Essentially,
they provide an infrastructure tool that allows people to view
cleansed information that is melded together from many different sources in real time.
Analytic
applications do this, but not in real time. The reason this problem is hard is so
arcane that I won't even try to describe it.
Under the old rules, success in a TLA area was supposed to make
you rich beyond the dreams of avarice, as you pressed more CDs and
minted money. With a TLA solution, as opposed to a solution to a
business problem, there is supposed to be broad horizontal appeal.
The software solves a variety of problems for a variety
of different companies.
The companies show no real signs so far of hitting that exalted
plateau, but in both cases the potential for horizontal appeal does exist.
Conclusion
Over the next couple of months, I'm going to talk about
several software vendors in more detail. Most of them, I like
and think can do well, even in the current climate. They
are, I believe, living proof that best-in-breed is not dead.
Bear in mind, though, that I am willing to call "successful" companies
whose story would have made them a laughingstock during the dot-com,
companies that still would be very unattractive to most venture
capitalists. For me, they're successful if they're profitable, likely
to survive and grow, and building up an annuity. It's not much,
I know.
But it's all I ask for little B2B Analysts, Inc., after all.
To see other recent Short
Take, click here for a listing.
|
