Technology Diffusion and Cadence
Horace Dediu of Asymco has a real gem of an insight hidden in one of his recent posts – but he was kind enough to put it in bold face to call attention to it:
“As a result, paradoxically, an enterprise technology must be sufficiently slow and expensive to be adopted.” – Dediu
Let’s unpack that statement working backward a bit into his post. The mystery Dediu is trying to unravel is why the epicenter of innovation has shifted from the enterprise, to the consumer. It used to be the case that almost all new tech first penetrated enterprises and then spread out to consumers- but clearly in recent years this process has worked in reverse. But why?
There are explanations – maybe IT is too risk averse- but they don’t stand up to the historical test (IT used to buy all the new and risky stuff). Summed up this way by Dediu:
“The bigger the firm, the slower and stupider it seemed. Were large firms employing dumb managers or did being a manager in a large firm make you dumb?”
But his alternate explanation is simple: that the buying cycle for large companies takes too long to absorb technology that changes too fast (annually or less). Put another way, the diffusion of technology through corporations is limited by the cadence of their buying and budgeting cycles – and they have to have these cycles because expensive decisions really require approvals for spends that are quite large.
The diffusion of technology among the consumers has a lot more to do with price point – below a certain price point, consumers can buy at whatever cadence meets their needs: additional functionality, fashion value, etc.
But an interesting corollary to this theory for technology diffusion is this: that once the cadence of new technology establishes a pattern, large corporations will adapt by creating ongoing device or modernization budgets, and delegating purchase decisions to the individual (e.g. a device purchase stipend). Effectively moving major capital expenditures into the category of ongoing operational expenses. We’re seeing this already with mobile devices and laptops… perhaps that approach will extend to software as well.