Feb 10, 2003 

 

“Even at the prototype stage, experts were saying the technology was obsolete.  Yet, in the face of tepid consumer response, the company stubbornly kept increasing production capacity and developing new models.  By the time it was finally killed, the initiative had cost the company an astounding $580 million and had tied up resources for 14 years.

 

The product was RCA’s SelectaVision videodisc recorder, and its story is hardly unique.  Companies make similar mistakes – if on a somewhat smaller scale – all the time.”

 

How do you allow projects to “fail fast” before they “fail big?”

 

The Feb 2003 issue of the Harvard Business Review, quoted above, contains an article by Isabelle Royer that presents research from two case studies and hypothesizes several reasons why companies launch innovations which fail in the market.  A couple of highlights:

 

Ø      The power of collective belief in a company that results in discounting countervailing data. 

Ø      The value of an “exit champion” who can be skeptical about a project’s potential for success.  In the two case studies researched by Royer, the skeptics were over-ridden, and left the company.

 

Three comments:

 

n      There are two types of errors – launching a failure and killing a winner.  The article focuses on “Type 1 errors” – launching a failure.  Hank Chesbrough and others looked at “Type 2 errors” at Xerox PARC and found there were a lot of those as well – Xerox PARC killed a lot of projects that went on to become successful companies – Documentum, for example.

 

n      The problem of selection bias. We don’t know how many other projects had these same kinds of narratives, with successful endings.  We all have heard examples of “romantic” innovation projects where the stubborn product champion prevails against the resistance of management, saving the company.

    

n      Balancing the need for skepticism with the need for alignment. The “exit champion” will find existence within the team quite problematic because, on a daily basis, the team needs to be aligned and have  everyone pulling in the same direction.  In addition, an “exit champion’s” views will be discounted precisely because of the role she’s been assigned to play.

 

Another way of approaching the need for an “exit champion,” which is intended to serve the same function, would be creating an external review mechanism, something that Hank Chesbrough at Harvard has studied and written about.

 

=======================================

After publishing this piece, Isabelle Royer, the author of the Harvard Business Review article,  wrote back with some extensions and corrections.  Here they are:

 

Here are a couple of comments on what you wrote.

Ø      The value of an "exit champion" who can be skeptical about a project's potential for success.  In the two case studies researched by Royer, the skeptics were over-ridden, and left the company.

This is not exactly what I wrote. An exit champion is NOT a skeptic except for a short period during which he/she makes his/her mind that the project should be stopped. Further, skeptics usually just speak or show their lack of enthusiasm. Exit champions do more than that. They go into action (provide financial and strategic evaluation, use sociopolitical networks) to defend an exit proposal.

Being merely a skeptic is not enough to reach an agreement upon exit among a group of managers. In the cases I studied, people who where merely skeptics and failed to go into action did not succeed in changing the course of action.

On the contrary, exit champions could make it because they had several characteristics that allowed them to succeed. These characteristics are pretty close to those of champions. That's why I called them exit champions.

n      There are two types of errors - launching a failure and killing a winner.  The article focuses on "Type 1 errors" - launching a failure.  Hank Chesbrough and others looked at "Type 2 errors" at Xerox PARC and found there were a lot of those as well - Xerox PARC killed a lot of projects that went on to become successful companies - Documentum, for example.

Totally right. I focus on one type of error only. However, I do not ignore the other type of error. That's why I indicated in the paper that you cannot ask for evidence of success at the beginning of the project because no evidence exists to estimate whether the project will be a success of a failure. Thus it's good to start a project based on belief. However, as the project unfolds it's possible to collect data to estimate the issue. The problem with the collective belief is that even when these data are negative, people don't change the course of action.

In the projects typical of Type 2 errors , projects are typically stopped too soon because of the lack of relevant data for evaluation.

Although my paper focuses on Type 1 error, I tried to point out the balance required to allow exit without cancelling projects too soon. That's why I indicated the need for top management to ask exit champions for evidence supporting their evaluation, just as they should have asked for growing evidence to champions as the project unfolded.

n      Balancing the need for skepticism with the need for alignment. The "exit champion" will find existence within the team quite problematic because, on a daily basis, the team needs to be aligned and have  everyone pulling in the same direction.  In addition, an "exit champion's" views will be discounted precisely because of the role she's been assigned to play.

You perfectly pointed out the difficulty of being an exit champion. The need for alignment is a strong pressure to continue a project whatever the situation. It's much easier for anyone to conform with others because there is no need to make one's own opinion (it's less demanding cognitively), it's more comfortable to be in agreement with others than to oppose them and it's easier to bring good news than bad news. Consequently, even when someone identifies a problem, it requires some courage to bring bad news to a majority of opponents.

I'd like to stress a misunderstanding, though. I indicated in the paper that the exit champion is NOT a role someone is assigned to play. On the contrary, the role is emergent. Someone happens to be willing to play this role.

This role cannot be assigned. It consists in promoting an exit proposal. Should Top Management assign someone to do it, they would already have made their decision. Thus the exit champion would be of no value. There is just a need for someone to communicate the decision and manage disbanding.

Another way of approaching the need for an "exit champion," which is intended to serve the same function, would be creating an external review mechanism, something that Hank Chesbrough at Harvard has studied and written about.

External review mechanisms are very important to provide new insight about a project. In the cases I studied, exit champions used external reviews to provide credibility to their evaluation. However, external reviews have an impact on the decision only when managers are open to any kind of result, be it good or bad. When a believer asks an external review and it happens that the conclusions are not in the direction she expected, she just puts it in her drawer and forgets about it. Should these conclusions have to be presented at a meeting, the consultant is just asked to revise the analysis so that anybody can understand what they want to understand. It's then very difficult to refuse this kind of request from your costumer, isn't it?

The point is: When providing the results consultants might face the same pressure "for alignment" than internal members do. This pressure might be pretty strong, if the firm is a big client of the consultant or if it threatens to ruin the consultant's reputation...