My friends over at Big Sky Thinking had a great series of posts on the decision making traps as discussed in a classic Harvard Business Review piece (that you can buy from amazon here: The Hidden Traps in Decision Making (HBR OnPoint Enhanced Edition)
- Decision-Making Traps Part 1: The Anchoring Trap
- Decision-Making Traps Part 2: The Status Quo Trap
- Decision-Making Traps Part 3: The Sunk-Cost Trap
- Decision-Making Traps Part 4: The Confirming-Evidence Trap
- Decision-Making Traps Part 5: The Framing Trap
- Decision-Making Traps Part 6: The Estimating and Forecasting Trap
It seems to me that these traps are interesting from an Enterprise Decision Management or EDM perspective. Some of them can be applied to how we design decision services (wiki) while some can be addressed if we automate intelligently.
The anchoring trap requires that you be systematic about how data is valued. For instance, is recent data more or less relevant than older data when making a decision. As they say in their posts, awareness is key. You can also use adaptive control (wiki) to keep challenging assumptions by trying new challenger approaches.
This is also the way to address the status quo trap. You can use adaptive control to test some really wacky challengers. Because you are not targeting many decisions with these challengers you minimize the risk but you gather data about what else might work. It is also critical to “talk with data” about the impact of different approaches, not just allow people to make assertions. Finally you can try data mining for rules instead of asking experts.
The sunk cost trap is a real one but I don’t think I have anything specific to add.
When it comes to the confirming evidence trap, systems tend to be better than people as they don’t have an emotional investment in their previous decision. That said, it is easy to design decision services in a way that falls into this trap too. Once again it is important to be systematic about using adaptive control rather than gut checks and to consider data mining for rules.
The framing trap is a real issue when designing decision services because you typically are doing so as part of an single business process. This makes it important to think about reuse of the decision in other processes and systems to make sure it is not framed by its initial use. Something similar applies to business rules, which can be framed by the first decision that uses them if you are not careful.
The estimating and forecasting trap was one of the primary reasons for me blogging about the challenge for experts of actually doing better than algorithms. Remember, speak with data.
One last thing, a colleague of mine once posted on the same topic and how EDM relates. Check it out here.
Comments on this entry are closed.