So ... what is a business? In other words, let's assume we want to automate a given business to the point that only strictly human activities are provided to humans - and wherever possible, those activities are assigned to cheap humans. That sounds horrible when I put it that way, incidentally. Let's say instead that I prefer to automate the boring and understandable parts of the business to permit the business owner to concentrate on the human aspects of it - interacting with customers and being creative. It's the same thing, but man does it sound better.
But what does this actually look like? What can we study of existing business management theory and practice that we can refactor into this system?
First, the overall structure of a given business is a
business model. The business model describes how customers are found, what customers pay for, and how customers are retained. Steve Blank
says the definition of a startup is an organization constituted in order to define, test, and iterate business models until one's found that works. (There's no reason you couldn't just keep iterating and spinning off businesses
ad infinitum, but most people stick with the initial business for a few years until they can sell it to others and either start over, start investing, or start vegetating.)
I should note here that a given organization can run multiple business models. Sometimes these are dignified as separate cost centers, and sometimes they're informal. There can certainly be revenue sources that consist of one-off deals that aren't particularly reproducible (although thought should be given to reproducing them - if one customer had a need that you could address, there's a good chance they're not unique).
So a business model defines a class of customers, a marketing strategy, and so on. The specific actions taken to instantiate the business model are business processes. Business processes are expressed in terms of workflow. The resources used in a business process are provided by a business organization.
A business organization is often a legal entity (i.e. protected by law and with its own tax identification and so on). It can also be informal, but experience shows that this is often a mistake, and certainly will create problems when the business starts to scale (I should note here that the goal of a business is generally to scale, that is, to increase the number of customers in order to maximize returns for a given amount of discovery).
The business organization also includes accounting, which is keeping track of the money and other resources provided by the business to run the business processes. Note that the underlying business structures have a lot to do with how the accounting is performed - only a business entity can have a bank account, for example, and tracking what goes into and out of that account is what accounting is.
Accounting also keeps track of a lot of the other parts of business processes, especially the standard business processes that every business has - issuing invoices, doing payroll, filing tax forms, tracking orders. A lot of this gets grouped into enterprise resource planning (ERP).
But there is a whole class of business process that has nothing to do with ERP; I think of these as "research", in other words, the processes devoted to finding and organizing information. Most of these processes don't really seem to be very organized in the literature I've found so far; most discovery seems to be kept in people's heads or at best quadrille notebooks somewhere. Given that business model iteration is research, though, I think it should be better defined.
Then there's the
business plan. The business plan is best thought of, I think, as being a snapshot of a given business model at a given point in time plus forecasts of what could happen to that model if you make a given change; its purpose is twofold. First and foremost it's something to show investors as a basic summary of why they should give you money in order to make more money. But it also represents
research - putting a b-plan together forces you to ask questions. Obviously, though, you should already be asking those questions all the time as part of your regular business model iteration process - because the questions are things like "how many customers probably meet these given criteria" and "how much money do I have". Stuff you
need to know. Steve Blank contrasts business models and business plans
here.
The financial part of a b-plan consists of your actual accounting figures plus your projections of how they could evolve, either if you continue the way you're going, or if you make a given change after the investor gives you X thousand dollars. In other words, there are two kinds of business plan: there's the snapshot of what and how you're doing, and then there is something that should by all rights be called a differential b-plan that represents your forecast of return on a given specific investment. Hm. I guess that makes it a plan, yeah? The snapshot is just a report. Your annual statement, as it were - and if you have shareholders, you're required to do those anyway. If you're just a startup, you're your own shareholders, so it behooves you to give yourself that same information.
The more of that we can automate, the better off everyone is. That's my goal.