Accurate estimation of the risked-weighted impact of novel technologies is critical to optimizing an R&D portfolio.? ?Overestimations of impact, especially in the nearer-term, of yet-to-be-delivered technologies tend to be far more common and significant than underestimations.? ?This optimism bias can lead to R&D decision makers missing or delaying investment in valuable opportunities.
In my experience, the most common issues leading to overestimating the impact of novel technologies under R&D include:?
- Underestimating, often by a considerable margin, the “debugging” time and cost required to achieve commercially attractive functioning of an innovation.
- Overestimating the utilization rate (i.e., consumer uptake) as a basis for appraising benefits.
- Underestimating the need to make adaptive adjustments in the upstream, downstream, and connected systems in which a new technology is placed.
- Underestimating the cost and effort to gain acceptance of associated changes in tasks by users, manufacturers, and support groups.
So how can these issues be avoided or at least mitigated??? There are two main avenues:
- Explicitly specify guidelines and best practices for how to assess and report “difficulty” (i.e., degree of challenge to meet the commercial goals within an acceptable timeframe), including clear definitions and levels of technical, implementation, and uptake difficulties.
- Explicitly calibrate against similar projects regarding the effort and time required to achieve the commercial and uptake targets.? ?Note that it is critical to consider survivorship bias in among the R&D efforts used for comparison since efforts by others which failed or suffered serious stumbles tend not to be well documented or even openly discussed.?? That is, do not just calibrate against success cases or simplified and sanitized histories of how technologies were developed.
?How do you avoid biases in your R&D prioritizations and portfolios?
Great post, worth a share!
Also, oftentimes, an operator implementing/facilitating a particular project may be spread thin across multiple concurrent research efforts. Let's not forget unexpected malfunctions and waiting for parts to arrive.