How to improve efficiency without hiring more staff

The hiring conversation in a growing SMB usually starts the same way. A department is drowning. Deadlines slip. Someone senior is working weekends. The team asks for another person, and on the evidence available they are right to.

What that evidence rarely includes is a breakdown of what the existing week is actually spent on. And when you get one, the picture is almost always the same: a significant share of expensive people's time is going on work that requires no expertise at all — moving information from one place to another, chasing it, checking it, and typing it in again somewhere else.

That is the gap between headcount and capacity. Hiring adds headcount. Fixing the process adds capacity — and it does not add a salary, National Insurance, a desk, or six months of ramp-up.

Where the hours actually hide

In our experience across UK SMBs, recoverable time clusters in five places:

• Re-keying. The same data typed into two or three systems because they do not talk to each other. Order into the CRM, then the accounting package, then the spreadsheet the ops manager uses.
• Chasing. Time spent finding out the status of something — an approval, a delivery, a signature, a document — because there is no place where the status simply lives.
• Assembly. Building the report, the pack, the reconciliation. Hours of gathering to produce fifteen minutes of reading.
• Rework. Doing it twice because it was wrong the first time, and the error was only visible three steps downstream.
• Handoffs. The queue between two people. Work sits waiting far longer than it takes to do, which is why lead times look terrible even when everyone is working flat out.

None of these show up on a job description. All of them show up on the payroll.

The five-step method

1. Measure a real fortnight

Not a survey of what people think they do — a fortnight of what actually happens, logged in blocks of work rather than by job title. Two weeks is enough to see the shape, including the monthly spike if there is one. The output you want is simple: for each recurring block of work, how often it happens and how long it takes.

Expect resistance, and expect it to be worth it. This is the only step that turns an argument about workload into a conversation about numbers.

2. Split judgement work from movement work

Go through the log and mark each block one of two ways. Judgement work needs a person: a pricing call, a difficult customer, a quality decision, a design choice. Movement work takes information that already exists and puts it somewhere else, or checks that it arrived.

Movement work is your recoverable capacity. In most SMB back offices it is a larger proportion of the week than anyone expects — and it is concentrated in exactly the roles you were about to hire alongside.

3. Delete before you automate

This is the step that gets skipped, and skipping it is expensive. Before automating anything, ask of each movement task: what happens if we simply stop?

You will find reports produced for someone who left in 2023. Approval steps that have never once resulted in a refusal. Checks that duplicate a check the system already performs. A form filled in because a bank asked for it during a loan application four years ago.

Automating a pointless task makes it permanent — it becomes infrastructure, and infrastructure is much harder to question than a habit. This is why process re-engineering comes before automation in every engagement we run: redesign the work first, then make what survives run itself.

4. Automate top-down by frequency × duration

Rank what remains by how often it happens multiplied by how long it takes, then work down the list. A five-minute task performed forty times a week beats a two-hour task performed monthly, every time — and it is usually easier to automate, because high-frequency tasks tend to be the most standardised.

For most SMBs the top of that list looks like this:

5. Spend the recovered time on purpose

Reclaimed hours do not stay reclaimed by themselves. If nobody decides what the time is for, it refills with other admin within a month and the improvement becomes invisible — which also makes it very hard to justify the next one.

Decide before go-live: this is sales follow-up time, or quality time, or training time, or the capacity that lets us take on the next twenty accounts without another hire. Name it, and put it in the diary.

When you should just hire

This is not an argument that hiring is always wrong. It is an argument for knowing which problem you have. Hire when the constraint is genuinely judgement capacity — more skilled decisions needed than skilled people available — or when you need a capability nobody in the business has, or when demand has structurally stepped up rather than temporarily spiked.

Fix the process when everyone is busy but the output has not risen, when the same complaints recur every month-end, or when the person you are about to hire would spend most of their week on movement work. In that last case you are not buying capacity; you are buying a more expensive version of the same bottleneck.