It should be clear to all readers that I think John Sumser
is great. I said it here and here
and here.
I was flattered by John’s article yesterday.
As John said, this is an important debate. His article is worthy of reply. So here we go:
Most debates turn on misunderstandings or lack of clarity. There
can also be an honest difference of opinion. In this case it appears we have
both.
As for the lack of clarity, it is best
to start from the beginning. In the article that started all this,
I said:
Here’s a good rule of thumb for people thinking about buying
a piece of technology: the more training it takes to use it, the worse it is.
And by “worse” I mean you will be saying “This stinks!” within a month of
implementation.
Here is John’s translation of that phrase (from yesterday’s
article):
It is no longer possible to afford software that meets user
requirements. The needs of users have become so sophisticated that the only
chance an installation will work is if it includes massive amounts of training.
In other words, software is dead.
I think that translation is a stretch. However, upon
reflection I understand that my original statement is both trite and unclear.
Some background:
In my opinion “bad software” is software that puts the needs
of users second to the objectives of the buyer and seller. In short, bad
software requires a cabal between the buyer and the vendor. This cabal is based
on the establishment of an ecosystem that is expensive, self-serving and
ultimately destructive to user’s needs. This ecosystem contains the software
vendor, the solution providers who perform implementation and customization, the
consultants who tell buyers what to buy and how to use it, and the buyer. All
successful enterprise software of the past has developed this ecosystem. ERPs
are just one of the biggest culprits in that they have made billions of dollars
keeping the ecosystem alive. The ERP ecosystem has done a lot of damage.
ERPs have made design decisions based on the likelihood that those decisions
would be expensive to implement (which would drive revenue to solutions
partners, who in turn will recommend the software that brings them the most money),
tough to operate (which would require specialized training for certain
individuals, who would then be “plants” inside the organization to constantly
advocate for continued use of the software), and difficult to integrate (which
keeps other software companies out of their business and provides market upside
for bringing out extended application sets). It’s pretty much a sham. The rich
(the software companies, solution partners and super-users) get richer, and the
poor (anybody outside the hallowed halls of “privileged access” who need the
data and services that the software is supposed to provide) get poorer.
That is the long translation of my original statement: The
more training it takes to use it, the worse it is. In other words, if you have
to spend a lot of money to make it accessible and worthwhile to the people who
actually have to use the software, then it isn’t serving the shareholders
interests. It is serving the cloistered ecosystem of the software provider. The
buyer / vendor cabal.
That’s why I don’t agree with John’s translation. My meaning
was not “It is no longer possible to afford software that meets user
requirements.” I was trying to say exactly the opposite: It is no longer
possible to afford software that DOESN’T meet user requirements. I think ATSs
are just a small microcosm of the larger enterprise software world. In my
article yesterday.
I made a quick shot of showing that I was starting to cut out software that
required me to join the cabal. I am the user of the analytics and distribution
technologies, and I wanted something that was cheap, easy and fast. And I got
it, thanks to advances in software technology.
Now to the honest difference of opinion. On John’s final
statement in yesterday’s article:
The problem is that software can not be used to automate
rote microscopic procedures on an individual basis. It's simply too expensive.
By suggesting that software can indeed drive additional organizational benefits,
Jeff is toe-ing the Silicon Valley company
line. Unfortunately, there are no examples of profitable businesses in this
arena. It's theory and poppycock.
I can’t say that I have ever been accused of toeing anybody’s line. I love Silicon Valley, but it is filled with just as many dopes as anywhere else in the world. If I didn’t honestly believe in the future of technology and business I wouldn’t be hanging out here. I could earn a lot more being a lawyer, doctor or plumber. That said, what I am expounding is clearly the Kool-aid that a lot of people have drank, in the past and the present. So John’s statement is worthy of a careful analysis.
I start with the fact that it doesn’t make sense to define
the only future opportunity of software as being the “automation of rote
microscopic procedures on an individual basis.” I believe a better way to think
about the future of software is to divide the market opportunities into three
buckets: administration (e.g. payroll), services (e.g. Amazon) and content (e.g.
this blog). (A nod to my friend Bud Hyler of Logical Marketing, who came up
with original framework.) Only the administration category relies on the “automation
of rote microscopic problems.” It will be the smallest software market in the
future by far. But even in the administration category it is a fantasy to think
that there is no upside for future software vendors.
For the question is not whether it is too expensive. The
question is whether the future automation of rote microscopic procedures on an
individual basis can be profitable. And since there are many examples of that
being the case (contrary to John’s statement), the likelihood is that there
will continue to be investment in software companies that are seeking to do
exactly what John says can’t be done. And since history shows that about 10% of
the companies taking such investments will become profitable companies or the
targets of acquisition by other profitable companies, it is likely that the
future will see a lot of profitable software companies automating the rote
microscopic procedures on an individual basis.
Who thought that Google could get rich earning $.10 a click?
Or that eBay could get rich charging a small percentage of the sale of some
dolls? Or that PayPal could get rich holding money for a short time before it
gets transferred to a vendor (often in very small increments)? All of these are
big businesses, and all of them are earning money off of administration even while
each in its way is a services company. For instance, Google’s AdSense program
makes money through the administrative tasks of finding places to put ads, taking
in money when someone clicks on those ads and then distributing the money to
the appropriate parties. Pretty dull stuff, and not that lucrative if you have
a bunch of people in a back room doing it. But put high-quality software behind
it and you can earn billions one little click at a time.
eBay and Google make profit off of small transactions. And
in each case, the first transaction was incredibly expensive. Google took in a
big chunk of change from VCs to get its ad business started. Expensive is not the issue. Scale (and the
profit that comes with it) are the issue. That scale is only possible with
software. And to think that Google and eBay have taken over the only two good
ideas that exist and therefore left the well dry for future software entrepreneurs
has no basis in logic, history, business theory or technology realities. That’s
like saying that since Ford mass produced the Model A that airlines aren’t
possible. After all, who could need another form of transportation when cars
are so convenient and quick. And think of the money that it would take to build
a jet airliner!
Here’s an idea. Voice clicks. Some vendor is going to make
a lot of money out there doing “voice clicks” in your car. As you drive around
you say “Hey, car! I am damn hungry and I want a burger.” The car is going to
match your needs to the burger vendor which was willing to pay the most to the
company that makes the recommendation. Finding a burger joint is about as rote
a microscopic individual procedure as you can conceive. It also happens millions of times a day. All
somebody has to do is figure out a way to charge $.05 for each transaction of
that type to make billions of dollars. Is that going to be Google or eBay? Is
it more dependent on search or auction? Only the future can tell. But it is
going to happen (assuming that I am holding the elephant’s trunk and not
something else).
But even if you bought into the argument that software is
dead and not coming back because all the administrative (i.e. “automating
microscopic transactions”) stuff has been taken care of, you would be totally
overlooking the software opportunities in content and services. These two areas
are both in their infancy. And as “services” start taking care of more “administration”
there is going to be even more money made in software.
I agree that the old software model is dead. It is good for
software, for users and for the shareholders of companies that buy enterprise
software that the old “pay a lot to get a little” model of enterprise software
is dying. But this conversation started with John saying “Software is dead and
not poised for a second coming.” As I say in the “About” section of this blog: “I
believe that the next five years will see greater changes in the business and
technology of talent than the sum total of changes that have happened to this
point.” I don’t know whether that means that I can see more of the elephant or
not, but I do know that the elephant is still the biggest thing in the room, and
everybody wants a ride.
BTW – By the use of the term “cabal” I am not insinuating
some sinister intent on the part of buyers. I think buyers get screwed by senior
execs who still don’t understand the concept of Cost Benefit Analysis. Heck, I
wrote a whole story about it.
Comments