Subject: Re: Wal-mart drives software industry
From: Ian Lance Taylor <>
Date: 26 Feb 2002 15:37:16 -0800

 26 Feb 2002 15:37:16 -0800
Tom Lord <> writes:

>        Ian and I have spent the last couple of years with big
>        Enterprise I/T.  We had some of the same na´ve preconceptions
>        you do, but were painfully disabused of them by the experience.
> Perhaps you could elaborate.

Gumby and I probably have different slants, but here are some of my
views.  These are obviously broad generalizations, and there are
exceptions.  In fact, one characteristic of enterprise IT is that
every company is an exception in some fashion.

Enterprise IT people are totally swamped by dealing with their
infrastructure, which has accreted since the dawn of the information
age.  They have no time to rationally revise it, and it's too big
anyhow.  They have no interest in purchasing anything which they
aren't already using.

New directions in IT are driven from above.  The IT people are
negative buyers--they pretty much always object to new things, because
new things obviously mean more work for them and they're already
behind on everything else.  When the higher-ups insist on acquiring
something, the IT people will try to get something as similar as
possible to what they already have--for example, if they have
Microsoft products already, they will prefer to get another Microsoft

Enterprise IT people try hard to dictate a standard Windows based
nenvironment which will be used on desktops throughout the company.
The head of IT will say that this has been rolled out everywhere.  The
people in the trenches will explain that nobody actually uses the
standard platform--that it has been tweaked everywhere, and people
persist in loading their own applications.  Things like 802.11b
wireless networks are generally installed in departments without IT's
knowledge or approval.  Desktop users typically see IT as an obstacle
rather than a member of the same team; often enough, the feeling is

Many shops are Microsoft only.  Microsoft buys business in the
enterprise IT space, so they are by far the cheapest solution, at
least for the first few years.  Also, you can hire people with MCSE
degrees who arrive predictably knowing how to operate Microsoft tools.

I never met any enterprise IT people who were interested in purchasing
in the technically best solution.  That just wasn't a decision
criterion.  Even studying several different possibilities was not
normally done.  Enterprise IT doesn't have to look for solutions;
solutions arrive regularly in in the form of salespeople.  Several
companies have particular time slots during the week in which they
torture sales people; when I was on those types of sales calls, the
sales people would just shut up and I would draw on the whiteboard--at
least the IT people would listen to me without heckling.

When a vendor walks in the door, the choices are: 1) fixes an
immediate problem and the vendor is reputable according to, e.g.,
Infoworld, in which case, buy the solution; or 2) does not fix an
immediate problem or the vendor has no reputation, in which case,
don't buy it.

Every enterprise IT department thinks that their company is so large
that their problems are essentially unique.  And, in fact, they're
generally right.

I think there are two typical routes for a startup to sell into
enterprise IT (no doubt creative people can come up with others).

The first is to hire a ``rainmaker;'' this is someone who regularly
plays golf with enterprise CIO's, and can get the CIO to order the IT
department to try things out on his personal recommendation.  Of
course the rainmaker has to believe your solution will work, and can
be profitable--i.e., that recommending your solution will not make the
rainmaker look like an idiot.  There aren't very many rainmakers, and
they get a base salary of $250,000 on up, plus sizeable commissions.

The second is to get industry analysts--the kind of people who write
for the Gartner Group--to believe in your solution and to mention it
to people.  This will get you into the computer press, at which point
you need to get a few of the weekly magazines to declare you best of
breed in your niche.  Then you have a shot at selling to companies who
have whatever problem it is that you solve.

> The growth and structure of Walmart implies that it has been managed
> with a deep knowledge about what was and is coming from technology
> suppliers: what's planned, what's possible, what's practical.  Are you
> saying they just made a series of lucky guesses?  Are you saying they
> wrote up specs themselves, dropped them off at the right doorsteps and
> the technology companies said "Oh!  Supply chain management!  Customer
> habit tracking!  Real-time retail financials reporting!  Wow, we never
> thought of that before and, what do you know -- it's practical to
> implement!"

I think you misunderstand.  Walmart simply says things like: we will
buy from the supplier who can give us the lowest price at predictable
volume.  Here is our purchasing interface: if supplier A supports it
and supplier B does not, from our point of view supplier A will have
the lowest price.  Since Walmart is very hardline with their
suppliers, but has enormous volume, they create a highly competitive
market.  This is a situation where capitalism is at its best, and the
solutions appear.  Walmart doesn't care what the solutions look like,
or even what the solution space as a whole looks like.  They just know
what they want in the end.

> Perhaps, but they apparently haven't dictated the retreat of the
> rising tide and the intellegence to not make such mistakes must come
> from somewhere.

Not at all.  It's quite possible that Walmart has said things like: if
supplier A can make the tide retreat, and supplier B can not, we will
buy from supplier A.  In a case like that, it may be that nobody will
have figured out how to get the competitive edge obtained by making
the tide retreat.  But Walmart doesn't care; as soon as somebody
figures it out, they'll start buying from them.