A job that I never knew existed: "Metaphor Designer." Here's an essay by someone who does this for a living, about what the job is, how it's done, and why it's important. A good metaphor, as he points out, not only communicates an idea better, but opens up new ways of thinking about a problem. He gives the example of a consultant who was working with a paintbrush manufacturer in the 1960's, as they were trying to figure out why a new brush design wasn't applying paint smoothly. The key insight came in the form of a metaphor: "A paintbrush is a kind of pump!"
While this might seem a little facile, it's actually quite key to how many designs happen. For example, a few years ago I found myself tasked with designing a planet-scale data storage and serving system. A key step was visualizing each datacenter as a seaport, with shipping lanes connecting them in a tree; each local data storage subsystem as a kind of warehouse; and giant armies of screaming customers around the world, each trying to send and receive data. By recasting the problem in terms of shipping logistics, it suddenly became clear how to organize and schedule data transfer (not trivial because storage capacities have grown 100 times faster than transfer capacities over the past few years), and that in turn led to several fundamental design approaches that made handling of data at hitherto-unconceived scales suddenly quite feasible.
(If you send attachments on GMail or store photos on Google, incidentally, that system is what holds your data)
But hidden in the description above was a second conscious choice of phrasing: "planet-scale storage." I coined the phrase and started getting people at the company to use it when it became clear that our storage systems lived in tiers (disk-, machine-, datacenter-, planet-scale), but also because the term conjures very different emotional responses from, say, "global storage." The latter has an implication of being the largest scale possible; "planet-scale" carries the subtle implication that the next project may well involve the Solar System. Oddly, that change of phrasing changes the way people think about things: it makes them approach problems knowing that there will be something bigger that they will have to ultimately deal with, and that causes people to make systematically better engineering decisions.
Erard isn't an engineer, and in fact most of the people for whom he designs metaphors do something very different: social welfare organizations. These showcase other aspects of the power of metaphor: not simply framing the way we think about problems, but the way we place them into the wider context of the world.
It also highlights subtle ways in which metaphors can be effective but trigger the wrong additional thoughts: for example, when working on a project to improve childhood resilience, they found a very powerful metaphor of children as orchids and dandelions. Some children are like orchids, thriving beautifully only under a narrow set of circumstances, while others are like dandelions, doing well nearly no matter what. What this ran into, though, was a cultural value that preferred the fragile and the rare (which in turn comes from the fact that the fragile can only be maintained in delicate circumstances, and so is restricted to the nobility): people saw the metaphor, understood it, and didn't see any value in making more dandelions.
This is why metaphor design requires careful testing. (I would have been a terrible test case for this; if you ask me if I'd rather make a dandelion or an orchid, my immediate answer would be that I'd far rather make something robust, and fragility is a pain in the ass, not a virtue, something you do only when there's no other choice. But then again, I'm an engineer)