Failing Your Way to Success

Fifteen years ago, I started a business that took 12 agonizing years to fail. I invested $10 million of my own money in BeComm (later called Implicit Networks) to develop a media-rich operating system that in many ways anticipated today’s smartphones and tablets. Unfortunately, the hardware and applications needed to make use of such an operating system didn’t yet exist.

Like most failed entrepreneurs, I might have had nothing to show for all my effort but a lousy tee-shirt that says, “I ran a startup, too.” But because I learned some crucial lessons from that failure, I now have the beginnings of a thriving business instead — one that is run very differently from the traditional model for building high-tech startups.

Indeed, our approach at Balassanian Enterprises today resembles what the New York Times last month ago called the “newer model for starting businesses” — one that “relies on hypothesis, experiment and testing in the marketplace,” and that also represents “a sharp break” from the traditional approach to building new businesses.

Consider, for example, the old axiom that entrepreneurs must be unwaveringly fixated on a single goal. Most startups are built around a single product or service that is assumed (but not yet proven) to meet a real consumer need and offer a lucrative market opportunity. The CEO of that startup is likewise singularly focused on getting a fully-baked product out the door as soon as possible in order to start generating revenues while at the same time building a pipeline for future offerings. Given the limited resources in most startups, this often means that the engineers are building Version 2 of the product before Version 1 has even been tested in the market.

Yet since we know from long historical experience that up to 90 percent of new products fail in the market, is such unwavering fixation really a good bet? For the venture investors in a startup, a 10 percent chance of success is not a big problem, because they have also invested in a dozen or two dozen other startups, only a few of which have to succeed to generate a healthy return for the VCs and their limited partners. But this approach works less well for the startups themselves, the overwhelming majority of whom will go out of business within five years.

Our approach instead is to develop multiple consumer products at once — all of them designed to blend the digital and physical worlds in new ways that provide value to consumers. We employ a disciplined and repeatable process for rapidly prototyping and market testing our ideas, which results in the early and low-cost failure of some of our products but the rapid growth of others that find market support. In other words, instead of speeding like a train on a fixed track towards a pre-set target — with a 90 percent chance of racing off a cliff — we have built into our process the ability to pivot quickly and change direction, which we believe is crucial to keeping abreast of today’s rapidly-shifting  markets and technology.

Wasn’t it only two years ago, after all, that netbooks were supposed to be the next big thing? In today’s smartphone-dominated landscape, who even remembers? But in that short span of time, a traditional startup can very easily launch a product that nobody wants anymore.

Our product design approach also differs from that of many startups. Our prototypes employ only a bare minimum of features before being tested with customers. In contrast, traditional startups usually develop the most fully-featured product possible to release to the public. We’ve found that our approach reduces cost, concentrates our resources only on products that generate clear customer support, and speeds the time needed to either validate or fail a product down to weeks or at most months.

Failure, in fact, is the engine of our success. The first iteration of our luxury shopping site Digbee.com tried to predict what shoppers wanted based on their past purchases. But market testing revealed that our predictive algorithm simply wasn’t good enough to provide real value to users. So we developed a new service called Strings.com, built around a highly-scalable crawler platform that enables users to effortlessly discover and follow their favorite designers and brands across multiple retail websites. We’ll be launching Strings with a unique rebate model shortly.

Another way we diverge from many startups, especially in the software field, is that we place critical importance on developing bullet-proof intellectual property. That’s not always easy to do these days, what with the backlog at the U.S. patent office. But we have learned that our ideas often have real value, and even if we can’t always productize them, perhaps someone else can — provided they license our patents. The patents from our last unsuccessful venture have generated millions of dollars in licensing revenue for us in the last 18 months alone.

The point being that if you own the intellectual property, a product failure does not have to mean a business failure. In fact, the IP can even benefit a whole industry. The patents we secured back in 1998 for our revolutionary “pinch, swipe, and zoom” technology, for example, were later embraced by Apple in the 2006 release of its phenomenally-successful iPhone.

Finally, we organize our teams around function rather than vertically-integrate them around product. This enables us to prototype and test multiple products at once — we are in the business, after all, of creating multiple new businesses —  and it also makes for a more efficient use of resources. While the prototyping team is working on an untested product, the marketing and business development teams can be laying the groundwork for rapid growth for an already market-validated product.

This functional organization of teams also lies at the very core of our business model. When a new product idea has reached an inflection point of market support and is poised for rapid growth — like our Post on the Wall service that uses photo-collages to meld people’s virtual existence with their real-world locations, events and experiences — then we start thinking about recruiting an outside team to spin out the business. We’ve found that the ideation and innovation skills of a startup team are very different from the execution, operations, and management skills of a growth-focused team.

In any event, while our approach has many similarities to the “laboratory” or “incubator” model that the Times wrote about, it contains several key differences as well. For one thing, we invest $3 million to $5 million of our own money in the businesses we spin out, rather than require the new growth team to secure early financing themselves to scale up the business. In this sense, we try to blend the best of the “labs” model with the upside of a successful VC firm.

Our approach is certainly not appropriate for every startup. But this much I do know: The average single-product startup has at most a 10 percent chance of success. My failure has taught me that by working on several products at once — and then quickly failing all but the very best market-tested ones — I’ve got a much better shot at evening the odds for startup success.