Eat tasks, not people: productivity for Indian small businesses

Andy Kessler’s book Eat People (Portfolio/Penguin, 2011) takes its title from its most provocative rule. It is not literal. “Eat people” means replacing human labour with productivity and technology: finding work a machine, software or a better process can do more cheaply. In an interview about the book, Kessler said most technology puts someone out of work, gave ATMs replacing bank tellers and online booking replacing travel agents as examples, and told entrepreneurs to look for jobs they can make unnecessary (Black Enterprise, May 2011). His case is that this is how an economy gets richer, and that better jobs follow for the people who build the new tools. (See idea 7 on our Eat People page.)

This article takes the useful part of that rule and applies it to Indian traders and manufacturers, where labour costs and the job question look different. The numbers are hypothetical.

What productivity means

Labour productivity is output divided by the labour used: units, or rupees of value added, per worker or per hour. It is the standard measure, and Kessler builds the book on it: in a reviewer’s summary, rising output per worker is what creates wealth (David Seah).

Productivity is not turnover. A business can double its sales by doubling its staff and be no more productive than before. It is also not the same as cutting jobs: output per worker rises just as much if the same team makes more.

Why the maths is different in India

Kessler writes about the United States, where labour is expensive, so a machine that replaces a task often pays for itself quickly. In India the same machine can replace work that costs far less, so it takes longer to pay back.

So at Indian wages, saving on labour alone is often a weak reason to automate. The stronger reasons are usually elsewhere.

The bigger gains: quality, speed and consistency

A machine that makes fewer mistakes saves material, rework and customer complaints. These savings don’t depend on wages, so they count just as much in India.

Add that to the labour saving in the first example and the ₹4,00,000 machine pays back in a few months, not a year and a half. Speed counts too: a distributor that bills and dispatches the same day gets paid sooner (see the stock and cash cycle), and an OEM may choose a supplier for consistent quality more than for price.

Eat the task, keep the people

Kessler’s view is that displaced workers find better jobs over time. That is a reasonable long-run argument about economies, but it is cold comfort to a helper with a family, and in a small Indian business the owner often knows every worker personally. There is a productive middle way: remove the low-value task, and move the people to work that earns more.

Here output per worker rises and nobody loses a job, because the business grew into the hours the machine freed. It only works if there is profitable work to move people to, so plan it before you buy the machine.

Where on the ladder

• Job workers (3.1) and contract manufacturers (3.2): you sell hours, so output per hour is your margin. See machine hour rate and capacity and yield.
• Component suppliers (3.3): OEMs ask for yearly price cuts (price-down); productivity gains are how you give them and keep your margin.
• Distributors (1.3) and shops (1.2): billing software, phone ordering and scan-and-pay remove counting, writing and chasing, the tasks Kessler would call ripe for eating.
• New-product founders (4.2): a product that replaces a manual task for thousands of businesses is “eating” at scale. Sell it on time saved and errors avoided, not on staff cut, and it is an easier sale in India.

A fair word on jobs

Kessler is deliberately blunt, and his Wall Street Journal column published as the book came out sorted whole professions by how replaceable he thought they were (WSJ, 17 February 2011). India’s position is different: a very large workforce, many people in informal jobs and middlemen roles, and new jobs that often need skills that take time to learn. Treat the rule as a question to ask about tasks, not a target for headcount. Raising productivity is still the only lasting way to raise wages and profits together.

In short

• Measure output per worker or per hour for your main product.
• Judge a machine on payback including quality and speed, not just wages saved.
• Before you automate a task, decide what the people doing it will do next.
• Look for tasks that need many hours but little judgement; they are the ones technology will take, whether you act or a competitor does.

Related: Markets that get cheaper with scale, make or buy and ROI and payback.

The rule is Andy Kessler's, from Eat People (Portfolio/Penguin, 2011), as he explained it to Black Enterprise (May 2011). The explanation, the India notes and the examples are ours.