A building contractor decides to automate his quoting process. At the moment, quotes involve a site visit, handwritten notes, a phone call to the supplier for prices, a spreadsheet, a Word template and an email, with the details retyped at almost every step. He buys an automation tool and asks a freelancer to connect everything together.
Three months and a sizeable invoice later, the process is automated. It’s also just as slow, because the real delays were never the typing. They were waiting for supplier prices, chasing missing measurements and getting the owner’s approval on every quote, none of which the automation fixed. Errors now travel faster too: a wrong measurement flows straight through to the client’s inbox.
Automation can save small businesses enormous amounts of time, as we explored in What to Automate First. But one of the oldest lessons in business improvement is that automating a poor process tends to give you a poor process that runs faster.
One Like insight: Simplify the work first. Then let technology speed up what’s left.
Don’t automate, obliterate
In 1990, Michael Hammer, a former computer science professor at MIT, published an article in Harvard Business Review with a provocative title: “Reengineering Work: Don’t Automate, Obliterate”. He argued that companies were spending heavily on computers to speed up processes that had been designed for a paper-based world, and that much of the investment was wasted because the underlying work was never questioned. Instead of automating existing steps, he urged businesses to ask whether those steps should exist at all.
Bill Gates made a similar point in his 1995 book The Road Ahead, in an idea still widely repeated: automating an efficient operation makes it more efficient, but automating an inefficient one simply multiplies the inefficiency.
Hammer’s ideas fuelled the “business process reengineering” movement of the 1990s, which had mixed results: some large redesigns failed, partly because they ignored the human side of change. But his core insight has endured, and it is especially relevant now that AI and no-code tools make automation cheap and easy. It has never been simpler to automate a process without first asking whether it makes sense.
For small businesses, the good news is that processes are usually simpler and easier to change than in large organisations. A short conversation and a whiteboard can often remove more wasted effort than any software.
Map, simplify, then automate
Before automating any process, a simple three-step approach can save time and money.
1. Map it as it really is. Write down every step of the process from start to finish, who does it, how long it takes and where things wait. Include the workarounds people actually use, not just the official version. A whiteboard or sticky notes work perfectly.
2. Simplify ruthlessly. For each step, ask: why do we do this? Could we stop? Could we combine it with another step? Could we do it once instead of several times? Where does work wait, and why? Often the biggest gains come from removing approvals that aren’t needed, collecting information once instead of repeatedly, or using a standard price list instead of calling a supplier for every quote.
3. Automate what remains. Once the process is as simple as it can sensibly be, automate the repetitive, rule-based parts: copying data between systems, sending standard emails, generating documents from templates, setting reminders.
Then measure the result: time taken, errors and customer experience. If automation hasn’t improved them, look again at the process, not just the tool.
As we explored in The Wobbly Table Problem, small, continuous improvements to how work is done often matter more than big technology projects.
[Infographic placement] Simplify, then automate. Three columns showing the same quoting process. Before: eight steps, with retyping, waiting for supplier prices and owner approval highlighted in red. Simplified: five steps, using a standard price list, digital site-visit form and approval only for quotes over a set value. Automated: the five steps, with data flowing automatically from form to quote to email. Caption: Based on Hammer (1990), “Reengineering Work: Don’t Automate, Obliterate”.
Starting with one process
Choose the process that frustrates your team or customers most, such as quoting, onboarding new clients, booking, invoicing or handling enquiries. Spend an hour mapping it with the people who actually do the work. Count the steps, the handovers and the waiting. Then challenge each one before spending anything on automation. You may find that half the improvement comes from simplifying alone.
A few questions can help:
- Which of our processes involves the most retyping, waiting or chasing?
- Have we mapped it with the people who actually do it?
- Which steps could we remove or combine before automating anything?
- After automation, how will we know whether the process is genuinely better?
Which brings us back to the contractor and his newly automated, still-slow quoting process. The software worked perfectly; it simply automated the wrong things. A standard supplier price list, a digital site-visit form that captured every measurement, and an approval limit so the owner only signed off larger jobs would have cut days from each quote. Automation could then have handled the rest in minutes.
Sources
- Michael Hammer, “Reengineering Work: Don’t Automate, Obliterate”, Harvard Business Review, July–August 1990.
- Bill Gates with Nathan Myhrvold and Peter Rinearson, The Road Ahead (Viking, 1995), paraphrased.
