Markets that get cheaper with scale

The first rule in Andy Kessler’s Eat People (Portfolio/Penguin, 2011) is “If it doesn’t scale, it will get stale.” In an interview about the book he defined scale as something you invent once and sell a million times, and contrasted professionals who serve clients one by one with chip and software companies that do the work once and sell it millions of times (Black Enterprise, May 2011). A reviewer sums up his picture as rising volume, falling prices and a market with a huge appetite, giving computer memory as the example: once costly and used sparingly, then cheap and in everything (David Seah). (See ideas 1 and 11 on our Eat People page.)

This article shows the arithmetic behind that idea and applies it to the manufacturer and new-product rungs. The numbers are hypothetical.

Why cost per unit falls with volume

Every business has fixed and variable costs. A fixed cost stays the same whatever you make, within normal capacity: a mould, a design, software, rent. A variable cost comes with every unit: material, packing, piece-rate pay. Cost per unit is the fixed cost spread over the units, plus the variable cost of one unit.

The bigger the fixed part, the more volume matters. That is what Kessler means by scale. A job worker (3.1) whose main cost is hours of machine time sees a much smaller fall; an own-brand manufacturer (3.5) with its own tooling, or a software start-up, sees a large one.

When a lower price pays

Because cost per unit falls with volume, a lower price can earn more profit, but only if the volume really comes. Kessler’s advice leans towards cutting prices to grow the market. The arithmetic tells you how much volume you need first.

So a 40% price cut needs volume to double just to stand still. Ask honestly whether the market will do that. (This is the same logic as contribution and break-even.)

Which markets get cheaper with scale

Kessler’s rule works best where three things are true, and it is worth checking each for your product in India:

1. Most of the cost is fixed. Design, tooling, software, brand building, R&D. If most of the cost is material that you buy at market price, scale helps less.
2. Demand grows a lot when the price falls. Many Indian buyers are very price-sensitive, so a lower price can open a much bigger market, for example among smaller towns or smaller shops. But some products don’t sell much more however cheap they get: nobody buys twice as much salt.
3. You can reach the extra buyers cheaply. Volume that needs an expensive sales force or deep dealer margins eats the saving. See the dealership economics and CAC.

When all three hold, the business that gets to volume first can price below everyone else and still earn more. That is also the warning in Kessler’s eleventh rule: if you don’t use your low cost to flood the market, someone else will.

Where on the ladder

• Commodity manufacturers (3.4) already compete on scale, but on material-heavy costs, so the gains are smaller and the competition is on price. See OEM vs commodity.
• Own-brand manufacturers (3.5) gain most from one design sold in large numbers.
• Product start-ups (4.2) and platforms (4.4) are the purest case: build once, sell many times. See unit economics and budgets and market size.
• Traders (level 1) can’t scale a product they don’t make, but a distributor or platform seller can scale the system: one warehouse, one app and one delivery network serving more shops. See operating leverage for why this cuts both ways when sales fall.

In short

• Split your costs into fixed and variable, and work out cost per unit at today’s volume and at twice that.
• Before cutting a price, work out the volume you would need to earn the same profit.
• Look for products where most of the cost is paid once and buyers respond strongly to price.
• Remember the risk: high fixed costs make profit fall fast if volume doesn’t come.

Related: Eat tasks, not people and When technology replaces the middleman.

The ideas are Andy Kessler's, from Eat People (Portfolio/Penguin, 2011), as he explained them to Black Enterprise (May 2011) and as summarised in David Seah's review. The explanation and the examples are ours.