|
BRADLEY: Tell us a little about your background…
STERN: I am originally from Israel where I earned a master’s
degree in computer science from the Technion, Israel’s equivalent
to MIT. After college and a stint in the army, I founded two companies,
the first was a software-development consulting firm that grew to
about 10 people. We quickly realized that consulting is a cost-plus
business with no potential for leverage, and decided to change the
company’s direction to focus on a product and a market.
I’ve always been fascinated with how we think, how we solve
problems and the challenges of designing machines to replicate human
thought processes. My second company, Rosh Intelligent Systems,
did exactly that. We developed a system to diagnose failures in
complicated equipment and sold it to customers including Applied
Materials, Xerox and Philips Medical Systems.
In 1989 I moved to the US to be closer to our customers. I founded
CardScan Inc. in 1993, the leading business card reading solution,
which has sold more than half a million units. The software that
reads and parses a business card is an example of computers mimicking
human intelligence – the same challenge that intrigued me
in my earlier companies. In 2000 I spun Eliyon Technologies Corp
(www.eliyon.com) out of CardScan to focus specifically on large-scale
information extraction from the Internet. Eliyon has been generating
cash since the middle of 2002 and is growing 30% quarter over quarter.
Eliyon’s technology is based on natural language processing
and other intelligent schemes to replicate processes conventionally
accomplished by human editors as they compile profiles of people
and companies based on information found on the Internet.
BRADLEY: What is your worst mistake?
STERN: Not asking the right question. At Rosh Intelligent
Systems we developed an expert system for troubleshooting equipment.
As part of out market research we showed a prototype to about twenty
potential customers and asked, “If we built a system that
would do this, would you buy it?” and more than half said,
“yes, we need it to reduce costs of training and downtime
of customer equipment.” Confident of our future success, we
completed the development but then encountered difficulties in selling
to these same customers. It turns out that many of them were, at
that time, busy implementing an operational system to track spare
parts inventory, called “history and service call scheduling.”
What we should have asked in our early meetings was “what
is your biggest problem and what systems are you considering buying
and implementing?” We misinterpreted polite and well-meaning
answers to what turned out to be the wrong questions. People are
a less polite and more direct when you ask for the order.
BRADLEY: What happened because of this mistake?
STERN: We designed and built one of the most powerful and
flexible diagnostic systems, but it was a hard sale, with long sales
cycle and a high degree of customization. We grew the company year
after year for five years, but never managed to turn a profit, which
in turn necessitated several rounds of financing to cover our loses.
It was no fun and kept eroding our founders’ shares.
BRADLEY: How did you fix this mistake?
STERN: After I left, we ended up selling the company to
another enterprise that offers a larger basket of “field service
solutions.” The system is still operational and is actually
the backbone of several “self service” web sites that
take customers through diagnosing and fixing home appliances and
other equipment. While this is really the ideal role for such a
technology, we built the technology before the widespread adoption
of the Internet. We were a little before our time.
the web did not exist in the early 1990.
BRADLEY: What would you do differently now that you know
better?
STERN: I continue to innovate and come up with new product
ideas that are based on machine intelligence. In that sense I do
not follow my own advice of “ask the customer what they plan
to buy.” However, I now put these ideas to real market tests
before investing too much money and emotional capital. For example,
when starting CardScan Inc., we raised a small round of angel financing,
just enough to bring a rudimentary product to market, and then started
selling. Only after initial sales successes did we go to venture
capital firms to raise more money. I now insist on taking products
to the market as rapidly as possible. In general customers would
make a purchase decision based on marketing material and a demonstration
and hence one can really test the market well in advance of big
development investments. However, that is not to say that once market
is validated that you do not need to invest in product to advancement
and polishing. For example, CardScan just released version seven
of its software and hardware, and Eliyon released its fourth generation
of product.
Another lesson I now adhere to rigorously is to bring the company
to cash generation as quickly as possible. It forces you to focus
on the only real thing – what would customers pay for -- and
it reduces clutter and confusion since you have no resources to
do things that are not absolutely necessary.
BRADLEY: What advice would you give to others about growing
their businesses?
STERN: Go to market as early as you can and focus your
energies on getting paying customers who will make you cashflow
positive. The worst thing you can do is to go through several product
development cycles before real customers actually purchased it.
No product is ever perfect and your customers are the best source
of feedback. Customers will actually shorten your development cycle
because when they pay money they will help you focus on a short
list of must-have features. If you can’t get any customer
to pay money for product in its rudimentary state, you have a problem,
but at least you’ve identified it before spending all your
resources. Without the experience I gained with Rosh and CardScan,
Eliyon would still be in development, polishing the algorithms.
As it is now, we have 25% of the Fortune 100 relying on the product
for their job candidate sourcing, and we have a rapid product enhancement
curve, constantly increasing the product’s value to those
customers.
Ben Bradley is the founder of Growingco.com (see the Darwin article).
GrowingCo customers use our data, forums, surveys and white papers
to improve understanding of customer needs, collect peer insight,
benchmark against peers and competitors, evaluate demand drivers
and understand market customer requirements. Ben can be reached
at ben@growingco.com.
|
