Workforce Analytics

Accept clients use the Client Portal, which shows the operational record rather than a summary of it:

• Who is booked for each shift, by site and by role
• Live attendance — who has clocked in, and who has not
• Timesheets, with scheduled against actual hours
• Worker profiles, including qualifications and history at your site
• Staff requests, raised and tracked rather than emailed
• Compliance views for the workers on your site
• Ratings and feedback your team has submitted
• Selected invoice and account information

Access is controlled by client and by site, so a site manager can be given their own operation while a head office user sees every site on the account. For multi-site businesses this is the difference between local detail and central comparison, and both are normally needed by different people on the same day.

The operational figures come from AcceptPulse rather than being recalculated in the portal. That sounds like a technical detail and is actually the point: it is why the number you see and the number your account manager sees are the same number, and why a query does not turn into a reconciliation exercise.

Changes go the same way. A requirement that moves — more people on Thursday, a shift pulled, a start time changed — is raised and tracked in the portal rather than in an email chain, so both sides are looking at the same request with the same timestamp on it. Most arguments about what was asked for turn out to be arguments about which message was the final one.

Most disputes about agency invoices are not disputes about rates. They are disputes about hours — what was booked, what was actually worked, and which of the two is being charged — and they happen weeks after the shift, when nobody can remember.

Where booked shifts, clock-in records and timesheets sit in the same place, cost accumulates visibly as the week runs. Finance can see labour cost building rather than discovering it later, and a discrepancy surfaces while both sides still have the context to resolve it. A query raised on Wednesday is a conversation; the same query raised three weeks later is a credit note and an argument.

Timesheet approval follows the same logic. If the hours being approved are the hours recorded at the clock, approval is a check rather than an investigation, and it can be done at the end of the shift week instead of becoming a monthly administrative event.

For interactive calculations before shifts run, explore our rate calculator and workforce cost calculator.

Six figures, reviewed weekly by site, will tell an operations manager almost everything worth knowing about agency labour.

1. 1Booked against attended, by shift pattern. Not an average — the split by pattern is where the problem lives.
2. 2Shifts that were never accepted, as distinct from shifts where someone failed to arrive. Different causes, different fixes.
3. 3Repeat workers as a proportion of the total. A falling number predicts an attendance problem before the attendance figure moves.
4. 4Leavers in their first two weeks. This is an induction and first-day measure, not a recruitment one.
5. 5Overtime covering agency gaps. The cost of unreliable supply, sitting in a different budget line.
6. 6Hours booked against hours worked against hours invoiced. Three numbers that should reconcile and frequently do not.

None of these requires a data analyst. What they require is that the figures are defined the same way each week and come from the same source, which is why definitions matter more than dashboards.

Workforce problems show up in the data before they show up on the line, because the behaviour changes first and the output changes once the slack is gone.

Four signals tend to arrive before a visible staffing failure, often by a fortnight:

• Acceptance rates falling on a particular shift pattern.
• Repeat workers declining week on week.
• More leavers in their first fortnight.
• A rise in shifts being filled at short notice.

By the time the line is actually short, the signal has usually been present for a while and nobody was looking at it.

This is pattern recognition rather than prediction, and the distinction is worth being honest about. Nothing here forecasts what will happen. What it does is make an emerging pattern visible early enough that someone can ask why, which is a more useful thing than a forecast and a considerably more defensible claim.

Last year's peak is the best forecast available for this year's, and most operations cannot use it — because the record was never kept in a form that survives twelve months. Three things are worth having:

• What you actually needed by week, not what you booked. The gap between the two is the overbooking you paid for, and it is usually consistent enough to plan against.
• Where the failures clustered. Which shift patterns, which weeks, which sites — peak failure is almost never spread evenly, and the cluster is the thing to fix.
• When acceptance started to fall. That date is roughly when the local labour market ran out, and it is the date to start ahead of this year.

Those three turn a peak plan from a percentage uplift on last year's headcount into something you can defend in a meeting. They are also worth putting to any supplier: one who cannot produce them has not been keeping the record, and no amount of goodwill will reconstruct it afterwards.

We will be straightforward about our own position. The systems that capture this were not running through last year's peak, so for most sites we cannot hand you last November on request either. They are running now, which means this coming peak is being recorded as it happens — and next year the conversation starts from evidence rather than from what anybody remembers. If a supplier tells you they can already produce all three for a site they took on recently, it is worth asking to see it.

Comparing two sites on a single attendance percentage is usually misleading, because the sites are not comparable in the ways that drive the number.

Before concluding that one site is better run than another, check what differs underneath it:

• Shift patterns and start times
• Public transport access at those start times
• The local labour market around the site
• The proportion of night work
• How long the site has been operating
• How much of the workforce is new in any given week

A site starting at 6am in an area with no early bus service will underperform a site starting at 8am on a bus route, whatever the management does. Comparing the two on a single attendance percentage tells you about the bus timetable.

The comparison that does work is like-for-like: the same shift pattern across sites, or the same site over time. And where a genuine difference remains after the structural factors are accounted for, it is worth looking at induction, supervision and the first-day experience before concluding anything about supply.

Each of these three measures on its own can be read several ways. Together they usually give one answer.

• Poor attendance, good retention. A scheduling or transport problem — people want the work but cannot reliably get there.
• Good attendance, poor retention. The work itself, or the way people are treated once they arrive.
• Poor attendance and poor retention, on one site while others hold. The site rather than the supply.

These point somewhere; they do not prove a cause. A pattern that holds for three weeks is a reason to investigate rather than a finding.

Reading the three together takes minutes and prevents the most expensive mistake available in temporary labour, which is treating a workplace problem as a recruitment problem and replacing the same people repeatedly. Learn more about analyzing retention and worker feedback.

Six questions worth asking about any supplier's system. Put them to us as well — we would rather be compared on these than on a demo.

1. 1Is the attendance data live from a clock, or entered afterwards from timesheets? The two produce different numbers and only one of them is useful on the day.
2. 2Can you see bookings, attendance and timesheets in the same place, or are they three separate reports that have to be reconciled?
3. 3Does the portal show the same figures the agency works from, or a summary prepared for you?
4. 4Can access be controlled by site, so local managers see their operation and central teams see all of it?
5. 5Can you raise and track a staffing request in the system rather than by email, so there is a record of what was asked and when?
6. 6Are the definitions behind each figure written down anywhere? A dashboard without definitions is decoration.