Start Ups Will Fail: Quite an Insight that’s 50 Years Too Late to Be New

September 29, 2008

One of the truisms for new companies, new products, and new girl friends in high school is:

80 to 90 percent will fail.

The fellow who gave currency to this notion was Conrad Jones, who died in 1992. If you don’t know about him and his work, click here for some information about this exceptional business analyst.

Imagine my surprise when I saw the Silicon Alley Insider’s article “Calcanis: Collapsing Economy Will Kill 50%-80% of Start Ups”. You can read this article here. Citing Scott Kurnit ( and other thought leaders, the author walked me through thr trials and tribulations of starting a Web venture. I found the mini-consulting tutorial in the section “10 Specific Things You Can Do” more intriguing than the set up to this advice. For example, the author suggests “cutting spending wherever you can.” I wanted to email my former colleagues at Booz, Allen & Hamilton to clue them in on something the firm has not known since 1917, the year Mr. Booz sold his first gig to the Sears’s management team. Tip nine also caused my heart to palpitate. The advice, “Build marketshare.” I had to take a dose of blood pressure medicine to calm down.


Reflecting on this article, several thoughts went through my mind:

  1. The advice is probably going to be made into a “Start Ups for Dummies” book. It will sell millions of copies. PT Barnum and other business wizards of the past would be supportive.
  2. The recycling of old ideas as new may be what’s behind the rediscovery of human-intermediated indexing, taxonomies, and finding ways to let people and colleagues comment on one’s work. Getting input is now “social software.” When I was 20, it was called “letting someone comment on a paper or an idea.” Paper worked. The phone worked. Now we need enterprise search systems that allow a user to tag. Old wine, new plastic bottles.
  3. Data from many fields of inquiry–quantum mechanics to framing stores in strip malls–suggest that in 100 tries at anything, most fail. Some behaviors can be learned. So line up the MBAs who are investing money as the Wall Street Journal does periodically, and most of the best fail when compared to one another. “Fail” is tough to define. Ignoring the definition makes it easier to give advice.

In one sector of the search and content processing industry, I took a Decmber 2006 consultant’s report that presented the top 48 companies in this sector. More than 90 percent of these underwent some significant change in the period to August 31, 2008. There were different types of “failure”, and frankly I think the categories of change are more interesting than the fact that most of these outfits are gone or transformed to such a degree that I could not recognize them unless I had my research.

To close, let me outline the types of “failure” that this research project identified. Since I haven’t delivered the report to the client, I can’t go into the type of gory detail that readers of this free Web log desire. But here’s what I’m willing to share:

  • Repositioning. About one fifth of this sample shifted business direction. “Fail” in this case meant continued experimentation. Most of these outfits are still answering their phone, but it’s hard to determine if the investors will continue to have patience.
  • Selling out or merging for no money. Another batch of companies in the sample sold to another firm. In several instances, the sell out allowed the buyer who competed with the company to perform the repositioning mentioned in point one above. The “no money” deal means that neither company had a business and someone felt that zero plus zero would equal a payoff.
  • Disappearing. I have profiled Entopia, Delphes, and TeezIR as examples of this action. I’m not sure the companies have gone out of business because executives from these outfits pop up elsewhere either as employees, in some cases business advisors (!), or entrepreneurs with a “new” idea and “new” money.
  • Refusing to respond. One example of this type of company is Lextech, a search and retrieval company profiled in the first edition of the Enterprise Search Report. In theory, it is still in business because it’s Web site is up, but no one answers queries. This is what I call “the limbo” strategy.

What I learned from this analsyis is that one must define failure. A generalization that 80 to 90 percent of new initiatives fail is old news. In fact, when Conrad Jones, a Booz, Allen guru, wrote the first version of the “Management of New Products” in 1958 or so, he emphasized that failure rates were high, and the successful companies found different ways to succeed. Among the factors identified in that ground breaking analysis was luck, family connections, inventing one thing and having it do another unexpected thing that people demanded (this applies to Google), and mindless bulldog tenacity. I’ve not exhausted the list. Your friendly Booz, Allen & Hamilton office can get you a copy and if you are lucky, you might be able to read the original version of Mr. Jones’s seminal study, now a half century deep in the management literature. A version appeared in the Harvard Business Review in 1959, and you will have to visit a real library to get this version of his brilliant work. I am willing to wager that the Web 2.0 crowd will be too busy reinventing fire, the wheel, and the lever to sample Mr. Jones’s brilliance. I had the privilege to work with him, and he would have had a chuckly over this recycling and rebottling effort.

Stephen Arnold, September 29, 2008


2 Responses to “Start Ups Will Fail: Quite an Insight that’s 50 Years Too Late to Be New”

  1. Otis Gospodnetic on September 29th, 2008 11:37 am
  2. In Tough Times, Don’t Panic | Job Searching Blog on January 8th, 2009 9:35 pm

    […] Start Ups Will Fail: Quite an Insight that’s 50 Years Too Late to Be New […]

  • Archives

  • Recent Posts

  • Meta