A recent academic study has shown that project managers without direct (managerial) authority over project teams are more likely to use redundant communication to get their message across.
For example, they might use both an email and an IM to transmit the same message to the same person. The redundancy doesn’t necessarily achieve better results, but the work gets done faster. See the full article here.
Old, but still relevant analysis of government IT project failure here on HBR. The six themes are familiar:
- Poor governance
- Lump sum funding
- Waterfall development
- Scope creep
- Automating too early
- Doing data cleanup too late.
It’s not a bad list, although I’m surprised to see waterfall development on there. There’s nothing inherently wrong with waterfall, and I suspect the issues relate more to scope creep rather than the methodology choice.
A recent post from the Havard Business Review quotes a Babson survey that surveyed 149 global executives, with 51% saying that experimentation is their preferred approach to assessing potential opportunities. The post goes on to catalogue various approaches to experimentation such as prototyping, A/B testing and simulation. The article implies that prototyping should rise from 51% over time, and I think that’s probably right, but there are also a few cases where firms should just skip prototyping altogether and build the final product.
Criteria for when prototyping will be useful:
- When prototypes are cheap to produce relative to the final product
- When failure rates are high
- When the cost of failure is high
There’s actually a mathematical relationship behind these three variables, but I’ll spare you the details. The first two are almost always true, for the first with a little imagination and creativity you can find a way to prototype that’s cheaper than the original design and for the second even if you think you’re design will work, there’s almost always something that can be learned from prototyping.
However, the third point about the cost of failure is worth discussing because in some industries given the long tail phenomena you are better off skipping prototyping and just producing the final product. This is true if the costs of failure are low i.e. customers will just not buy the product, rather than buy it and have a bad experience hurting your reputation and increasing customers’ support costs. In these cases prototyping is actually a stage you don’t need, you should just get the product out there, and the line between a product and a prototype becomes so blurred as to be meaningless.
Another implication from a portfolio perspective is how much time should be spent on analysis before building something (either product or prototype) starts, given the ease of prototyping, it certainly seems that the number of prototypes compared to products should be higher in almost all industries.
Insightful post from Havard Business Review on the value of friendships – spending money on others makes you happier than spending on yourself. More successful people at work tend to have more friends at work.