Another way to incorporate "learning from failure" into a company's culture might be to recast the entire notion of failure by creating a culture of experimentation. Over on the Fortune Business Innovation blog, Dominic Basulto cited an article by James Cash of Harvard Business School that highlighted the importance of creating a culture of business experimentation within a company, pointing out examples from a broad spectrum of companies: Wal-Mart, Capital One and General Electric. Creating this culture of business experimentation is harder than it sounds, since you can't always know whether or not an idea will work until you put it into action, and most companies are not prepared for the downside risk of failure.

Cash's article contains some interesting comments on failure tolerance and the need for a “culture of experimentation. For example, Cash says:

"It's not an easy feat to create an environment that walks the line between so-called failed experiments—where the discipline of data collection, analysis and iteration results in learning even if the experiment itself doesn't produce a desired result—and the frivolous waste of resources, where ideas are tested in an undisciplined manner."

What I'm thinking is that you could undercut the negative psychological energy around the concept of "failure" by incorporating a culture of experimentation. Here's my guess at what this culture of experimentation might look like:

Experiments would be fashioned from specific innovation challenges framed to meet certain strategic objectives. The resulting concepts would be prototyped in successive iterations, each designed to get closer to the "finished" product or process. At each prototype level an evaluation would take place using criteria developed out of the original objectives. The decision to kill the concept or develop it further would be made. If the decision was to kill, then a full report on what worked and what didn't would be made.

Here's what's most important:

1. Set up a specific process for experimentation and identify experiments as such from the start so that projects that don’t go through the process aren’t allowed to “bring down” the process.

2. Start the process with a specific challenge tied to strategic objectives – this kind of disciplined approach will ensure the relevance of both the process and the results.

3. Make evaluation criteria as specific as possible and relevant as well (which it will be if it’s tied back to the strategic objectives). This is critical to a disciplined approach, and it’s also critical to avoid blame and finger-pointing when failures do occur.

4. Subject all prototypes to rigorous analysis. The ability to build on what’s working and change what’s not is based on the understanding that would come from such an analysis.

5. When a project is killed, document all possible learnings, including things that can be learned about the process itself.