Reducing Agency Labour Costs

The hourly rate is the visible part of agency labour cost and frequently not the largest part. A cheaper supplier that delivers less reliably transfers cost onto your operation, where it turns up under three other headings and is rarely added back to the comparison.

Every cost in the section above is real and almost none of it reaches an invoice. That is the actual reason agency spend is difficult to control: the expensive parts are not the parts anybody measures, so the conversation defaults to the one number everybody can see, which is the rate.

Accept's answer to that is Workforce Intelligence — the systems behind our service, which record what was requested, what was accepted, who attended and who left, as the work happens rather than in a report afterwards. It turns the three costs above from impressions into arithmetic:

• Unfilled shifts become a figure you can set against your overtime bill, rather than a sense that you are short a lot.
• Early leavers become a four-week attrition rate you can track, rather than a feeling that people do not stick.
• Supervisory load becomes visible in how many attendance exceptions somebody had to chase last week.

None of that reduces a cost by itself, and we would not claim otherwise. What it does is tell you which cost is worth attacking, which is the part most operations are guessing at. Taking twenty pence off a rate is easy to do and easy to measure. Working out that your night shift loses people in their second week because induction only runs on days is harder, worth considerably more, and impossible without the records of attendance.

1. 1Reduce the number of workers you need rather than the rate you pay. Better attendance means fewer bookings to cover the same output, and that saving is larger than any margin negotiation.
2. 2Keep the workers who already know your site. A returning worker needs no induction, is productive immediately and is far less likely to leave — all of which is cheaper than a lower rate on a new person.
3. 3Book to what you actually need. Overbooking against the risk of no-shows is paying for the problem twice; the answer is better prediction, not a bigger buffer.
4. 4Look at shift patterns and start times before pay. Transport access frequently unlocks availability that money does not, and it costs nothing.
5. 5Measure cost per hour worked rather than per hour booked. They are different numbers and only one of them reflects what you actually got.
6. 6Challenge the margin openly rather than shopping the rate. A supplier who itemises can have a real conversation about margin; a supplier who quotes one number can only move it by taking something out.

There is one decision this list does not settle, and it comes up every peak: whether to raise the pay rate to stabilise staffing. The cost of a rise is easy to work out and easy to underestimate, because it applies to every hour worked by everybody — including the people who were turning up reliably at the old rate. Set that total against what the instability is actually costing you in premium overtime hours, in replacement recruitment and induction, and in output you did not get.

Do that arithmetic before raising the rate rather than after, and test the specific cause first. If people are leaving because the shift starts before the first bus, a pay rise is an expensive way to not fix it. Where pay genuinely is the cause, the sum usually favours the rise — but it should be a calculation you can show, because it is very hard to reverse.

Any supplier can tell you their service saves money. The measurement is straightforward if you set it up before you start, and close to impossible afterwards, so it is worth doing in the first week.

Take a baseline across four things before anything changes:

1. 1Cost per hour actually worked, rather than per hour booked.
2. 2Premium overtime hours covering gaps.
3. 3Replacement workers recruited and inducted.
4. 4Supervisory hours spent on attendance and re-planning rather than on the operation.

Write down the definition of each one, including what is counted and over what period. Then measure the same four the same way, over a comparable period, allowing for seasonality — comparing a January against a November tells you about Christmas, not about your supplier.

Accept does not publish a headline return-on-investment figure, and you should treat any agency that does with caution. A number like that depends entirely on assumptions about the client's overtime rate, turnover cost and administrative load, and none of those transfer from one operation to another. What we publish instead is the calculation itself: our workforce cost calculator asks for your annual hours, your margin, your administrative position and your four-week attrition rate, and works out the value against the margin you actually pay. If the answer is unconvincing on your numbers, that is a more useful thing to find out than a figure from somebody else's.

Beyond that, we will agree the four measures above with you at the start of an engagement and report against them, which is a claim you can check rather than one you have to take on trust.

Agency administration is mostly reconciliation: working out whether the people booked are the people who worked, and whether the invoice reflects either. It is unglamorous, it consumes management time, and almost all of it exists because the booking, the attendance record and the invoice come from different places.

Where booking, clock-in and timesheet data sit in one system, that reconciliation largely stops being necessary. Accept clients see booked shifts, live attendance and timesheets through the Client Portal, which means the question “does this invoice match what happened” is answered as the week runs rather than after it.

The practical value is in finance seeing labour cost accumulate during the week rather than discovering it on the invoice. A discrepancy found on Wednesday is a conversation; the same discrepancy found three weeks later is a credit note and an argument.

Six things make up a temporary worker charge rate. Only one of them is the agency's.

• Pay rate — what the worker actually receives per hour before deductions.
• Employer's National Insurance — paid by the employer on earnings above the secondary threshold, at the prevailing rate.
• Workplace pension — employer contributions for auto-enrolled workers, calculated on qualifying earnings.
• Holiday pay accrual — statutory paid leave, accrued as the worker works and paid when taken.
• Apprenticeship Levy — where the employing business's pay bill exceeds the threshold.
• Agency margin — recruitment, compliance, payroll, account management and the agency's profit.

The first five are statutory. An agency cannot reduce them, and any supplier whose charge rate implies they have is either absorbing the cost from margin — which is unsustainable and usually precedes a service problem — or has made an assumption about the engagement type that you want to know about before it becomes your problem.

That leaves margin as the only variable an agency genuinely controls. This is worth stating plainly because the sector generally does not: a quote is not a price for labour, it is a price for the service wrapped around statutory costs and statutory on-costs that every supplier pays identically.

Four things explain almost every gap between two quotes, and only one of them is margin.

• 01The assumed pay rate. A supplier quoting a lower charge rate may simply be assuming a lower rate to the worker, which shows up later as difficulty filling the shift or as people who leave when a nearby site pays more. Ask what pay rate each quote assumes before comparing the charge rates at all.
• 02What is included. Induction time, PPE, transport arrangements, on-site supervision, out-of-hours cover and replacement of a worker who does not arrive may be inside one quote and charged separately in another. The cheaper rate is frequently the one with less in it.
• 03The treatment of premium hours. A rate that looks lower at basic can be higher in practice at a site with significant overtime, weekend or bank holiday working.
• 04Engagement type, which changes the statutory picture underneath. If two quotes differ materially and nothing else explains it, this is usually why.

Six things to fix before you compare, so that you are comparing the same thing.

1. 1State the pay rate yourself rather than letting each supplier assume one. This single step removes most of the apparent difference between quotes.
2. 2Ask for the charge rate broken into pay, each statutory on-cost, and margin. A supplier who will not itemise is asking you to trust a number rather than understand it.
3. 3Fix the engagement type across all quotes, because it changes the statutory costs underneath.
4. 4List what you expect included — induction, PPE, supervision, replacement cover, out-of-hours contact — and require each supplier to confirm against the same list.
5. 5Ask for overtime, weekend and bank holiday rates, and apply them to your actual shift profile rather than to basic hours.
6. 6Ask what happens when a worker does not arrive: whether replacement is charged again, and whether the original booking is credited.

For an interactive tool to model statutory components, explore our rate calculator.