You’re staring at your screen, stewing. When you applied for this job, the ad specifically said: “Can work to tight deadlines.” But now, with a ton of work outstanding and a deadline looming, you’re thinking: how can this be my fault, when I’m waiting on other people?

No, the real problem is Gus, who should have sent you input ages ago. You refocus your attention on the screen, where you’ve been reading an article on job automation.

You know for a fact your job can’t be automated; it’s far too complex. But now you’re thinking – “They should bloody well automate Gus.”

Over 70% of US workers believe they are safe from job automation. Some have got a point: a recent OECD study tells us that, compared with what we thought five years ago, we may have grossly overestimated the number of jobs at risk of automation. How come?

It’s not necessarily that our jobs are too sophisticated. In many industries, like relay runners, we progressively transform data and add value. But our work is inter-connected, threaded with intertwined processes: some offensively mundane, others requiring a high degree of human involvement.

For this reason, automation talk is changing, moving from automating jobs to tasks. The challenge thus becomes untangling processes, to pull out the tasks most suited for automation.

It’s a tricky decision: should you tackle the first step of a process? The easiest step to automate? Or maybe just try and shave off a little bit of work out of each link in the chain?

This is not just an academic debate. Companies are increasingly seeking technology solutions to automate parts of their business. But all too often, the return on investment doesn’t materialise because the solutions end up addressing only a tiny part of the workflow. Worst of all? After all that investment, you’re still waiting on Gus.

The Theory of Constraints (TOC) offers a structured way to figure out which automation solution can deliver the most value to a business’ needs. Don’t be fooled by the vaguely academic name: the methodology – developed by Dr Eliyahu Goldratt – has been put to use by an impressive range of companies.

Its core principle is that, for any system, output is determined not by how much you push in, but by a resource with finite capacity. In other words, it’s not the size of the pipe that determines how much water flows out; it’s the size of the blockage.



The constraints in large departments can seem daunting: with all these processes, surely there’ll be hundreds to tackle! But – and this is the great beauty of the TOC framework – however complex the system, it’s nearly always governed by very few constraining factors. The trick is to find which ones.

Notice that constraints are not bottlenecks: while bottlenecks can move around, constraints are fixed. When a three-lane highway turns into a two-lane road, traffic capacity shrinks. That’s a constraint: fewer cars can flow in any given hour.

During rush hour, this results in traffic jams. There, you find plenty of localised bottlenecks, like the car snailing along, keeping its distance from those ahead of it. But even if that slugger were to magically vanish, another bottleneck would soon pop up elsewhere. The only way to genuinely solve a traffic jam is to address the constraint.

Back in the office, bottlenecks are equally predictable. Batching is quite often a culprit, like when Gus chooses to wait until he has a bunch of items to process, rather than processing each one as soon as possible. More broadly, bottlenecks are formed by the clashing of measurements, or, in the words of Dr Goldratt:

Tell me how you measure me, and I’ll tell you how I’ll behave.”

Gus isn’t prioritising your work because through actions, incentives or behaviours, his boss prioritises servicing other departments first. Gus feels bad for you, but tough luck.

Identifying the constraint is therefore crucial: the best way to optimise the entire system is to get the most out of its constraint. That’s what the TOC calls subordinating the process to the constraint. If Gus really is the constraint, then you – along with everyone else – need to revisit the workflow to make sure his time is effectively spent.



Longer term, though, the aim is to elevate that constraint, and here automation can deliver enormous gains. Rather than replacing humans with machines, automation is used to hack at tasks ideally suited for machines. Your targets are tasks that both clog a process and consist of repetitive, boring work.

Certainly, it’s easier to buy an off-the-shelf product. But, though such solutions do tackle tasks fit for machines, all too often they don’t cater to a business’s constraint.

For this reason, we’re seeing companies developing in-house solutions. Goldman Sachs invests heavily in its app suite Marquee; Reuters is developing its data-sifting tool Lynx Insight; and Bloomberg is cultivating its NLP toolset.

The entry barriers for deploying automation are also dropping, owing to high-end open-source code, which allows people to experiment in building useful tools. (Check out this cool cucumber-sorting machine built for a Japanese family farm.)

Personally, what I dislike most about automation chatter is how disenfranchising it tends to be: Jobs are “taken from employees”; “stolen”; and “killed-off”. I think it’s a wrongheaded approach to take, when such bountiful gains can be made by working with employees to identify where automation can help.

This shift in mindset moves away from cost-cutting, driven by scarcity-thinking, to a growth perspective, which seeks to achieve much more with existing resources. So employees get to improve their jobs, while their companies bump up productivity. Isn’t this the real win-win scenario?

If you’re interested in reading more about the TOC, I’d recommend the following, in this order:


  1. This Introduction, which offers an excellent taster.
  2. The Goal: A business book disguised as a novel, authored by Goldratt. It tells the story of Alex Rogo, who manages an inefficient production plant set to be closed. Alex has just three months to turn the plant around, which he does with the help of an advisor named Jonah, who introduces Alex to the TOC. The book was hugely popular in the 1990s, on account of its highly applicable principles.
  3. Pride and Joy: To me, this novel, authored by Alex Knight, who worked closely with Goldratt, represents an improvement on The Goal. The book presents a much-needed adaptation of the TOC principles to the services industries, specifically to a failing hospital (and based on real work done with the NHS). I can’t recommend it enough.
Vered Zimmerman

Vered Zimmerman

Vered is an investment writer in our London office. She holds an MBA from Cass Business School and an MSc in mathematics from the Hebrew University in Jerusalem.
Vered Zimmerman