In the midst of preparing Autodesk's public offering, I suddenly realised that investment bankers and accountants actually believed there was a causal relationship between the percentage of sales spent on R&D or marketing and the time-delayed sales and profitability of the venture. I had never even calculated such numbers, much less assigned any significance to them, focusing instead on what specifically needed doing, then how much could be done with the resources at hand. After recovering from the offering, I began to think about the idea that ``aggregates aren't reality''.
The process of managing a large and growing business is very much a matter of learning how to interpret abstract aggregate measures of the performance of the business and thereby deciding what specific actions to take. In the midst of struggling for survival while learning that skill, it's hard to remember that:
Aggregates aren't reality
Reality is events, not a process.
Reality is discrete, not a continuum.