Trying to cut costs in a company by trimming the IT budget is like trying to cool a room by blowing cold air at the thermostat: The harder you try, the worse things will get. But since you’re reading the thermometer built into the thermostat, you’ll be sure you’re improving the situation.

I can almost hear the yeah-buts. “Yeah, but you can’t just write a blank check to IT.” “Yeah, but you have to hold the CIO and CTO accountable for IT spending.” “Yeah, but IT has to generate a measurable return on investment.”

Yeah, but we’ve replaced management by objectives and management by walking around with management by platitude.

Last week’s column inspected quicker/cheaper/better — the process triad — and found that the “quicker” consists of two independent measures, cycle time and throughput. If you don’t pay attention to both, process redesigns can do more harm than good.

This week’s column focuses on “cheaper,” which turns out to consist of three measures, not one. To make a process, or for that matter an organization less expensive to operate you can reduce fixed costs, variable costs, or total costs. The distinction between fixed-plus-variable cost reduction and simple cost-cutting is the difference between smart, well-considered business improvement and felonious stupidity.

I’d stop here, figuring it’s just too obvious to need further explanation, except that I’ve seen, and sometimes endured, quite a few IT cost-cutting exercises during my increasingly long tenure in this business, and every single one ignored this, the most basic concept in cost accounting. (For that matter, I recently corresponded with one of the industry’s most respected authorities on total cost of ownership (TCO), who waved it off as an irrelevancy. And people pay more for his opinion than they do for mine. There ain’t no justice, or even worse, perhaps there is.)

To understand why looking at costs from a fixed-plus-variable perspective is so important, look what happens when you don’t: You cut total costs, usually by laying off staff since labor is the most expensive line item in the IT budget. Layoffs reduce costs quickly and convincingly. Of course, they also reduce your ability to deliver working product, which in the case of IT means working software.

Simple-minded reductions of total costs, in other words, reduce delivered value. Almost without exception, this reduction in value exceeds the reduction in costs, which means cost-cutting leads to a net reduction in benefit. It’s the thermostat problem.

Separate fixed and variable costs and you get a very different result.

Fixed costs are your overhead — the cost of turning the lights on. Reducing overhead has no effect on your ability to deliver results.

Variable costs, also known as unit, incremental, or marginal costs, are what you spend to produce one more doohickey. Reduce variable costs and you don’t affect your ability to deliver results either. Reduce fixed and variable costs and you reduce expense while leaving value untouched. You increase benefit, where mindless cost-cutting reduced it.

The popular subject of offshoring illustrates this concept nicely. Why send programming work offshore? To reduce variable costs. Whatever your measure — cost per line of code, cost per function point, or cost per hour of programming effort, it’s lower in India and the Philippines than here.

Between networking costs, the infrastructure needed to coordinate activities on multiple continents, and the additional responsibilities associated with monitoring vendor performance and managing a multicultural, 12-hour-time-shifted business relationship, offshoring has higher fixed costs than using staff resources or on-shore systems integrators.

I was talking about fixed-plus-variable analysis with a client, recently — a former teacher of high school algebra. “It’s nothing more than the equation for a straight line,” I mentioned, “y = ax + b.”

“That’s all you need to say,” he responded. “I understand completely.”

Which means, I guess, that to understand the relationship between the cost of information technology and its benefits, you’re better off with a high school algebra teacher than a high-priced TCO consultant.