Commitment Counts

Of every million ideas for a high-tech business, it is estimated that only six become public companies. But adopting the right organizational model and employment practices can help improve those odds, according to a group of researchers at Stanford and MIT.

The findings are based on data gathered by the Stanford Project on Emerging Companies (SPEC), which has tracked 167 young high-tech companies in Silicon Valley since 1994. The SPEC team, headed by Stanford professors James Baron and Michael Hannan and MIT's Diane Burton, began by identifying five organizational blueprints based on employers' ideas about how work and employment should be organized: specifically, the nature of employee attachment, the processes used to coordinate and control work, and the best criteria for selecting employees.

1. The engineering model describes companies that attract employees by offering challenging work, that emphasize informal peer-group control, and that hire people for specific skills. (Of the 167 companies surveyed, 31% were founded using this model.)
2. The star model refers to companies (8% of those surveyed) that expect employees to internalize a strong professional commitment to excellence and require relatively little control. In this model, employees are selected according to their long-term potential.
3. The commitment model depicts companies in which strong emotional bonds are the basis of employee attachment. Commitment-model companies (7% of those surveyed) believe in informal peer-group control and emphasize cultural fit in hiring.
4. The bureaucracy model (seen in 5% of the surveyed companies) covers organizations with formal control procedures that attract staff with challenging work and focus on specific skills in selecting employees.
5. The autocracy model references the organizations (3% of those surveyed) that motivate employees primarily with financial rewards. Such companies emphasize the importance of close personal oversight for control and coordination — and also make hiring decisions based on specific skills.

The remaining companies in the sample had founding models that differed in one or more dimensions from the five basic types.

Because of the importance of technical talent in Silicon Valley, the star and engineering models might be expected to prove particularly effective. But when it came to predicting the likelihood of an initial public offering, the commitment model —commonly associated with large companies in stable industries — turned out to be the winner.