SKUBIQ
LOGIN SIGN UP

One Pool, Many Sites: How SKUBIQ-LMS Moves Trained Labor Between Warehouses Against Live Demand

Black line-art illustration of a warehouse aisle in one-point perspective with tall racking of plain cartons, a wall-mounted board split into two site columns of operator badges, one badge moving across between them, a wall clock, a dashed travel route line and a worker pushing a pick trolley

Two warehouses, one week where the second site is short two pickers and the first site has three qualified operators with a quiet Thursday. The instinct is to send them. The reality is that sending them is a decision with a cost, a compliance check and a travel time attached, and most sites make it on the floor, without a rule, at the moment it is cheapest to say yes.

That is what multi-site labor pooling is: a shared pool of trained operators that more than one site can draw on, with the rules for travel, swaps and surge cover written down before the week goes wrong. It is not a second headcount. It is the same people, counted once, visible from every site at the same time.

This guide covers what a pool actually contains, the three questions that decide whether a borrowed operator pays for herself, why the pool has to be a qualification list rather than a headcount list, how travel time quietly becomes the largest line in the decision, what breaks when cross-site swaps are left to self-service without rules, the numbers that tell you a pool is working, when the pool is the wrong answer, and a 90-day plan to stand one up across sites - including what SKUBIQ-LMS and SKUBIQ-WMS each contribute.

What multi-site labor pooling actually is

Labor is the line that decides whether a warehouse is profitable or not. Element Logic, writing on warehouse workforce management and labor planning, puts labour costs typically at 50% to 70% of a warehouse's total operating expenses, and Pebb's guide to warehouse workforce management carries that figure into the wider staffing argument. It is also the line that is hardest to shrink, because the work does not disappear when the demand dips - it just gets paid for anyway.

Most sites respond to a shortfall in one of two ways: they pay overtime, or they call an agency. Both are legitimate. Both are expensive. A third option is usually sitting in the building next door.

Racklify's encyclopedia on labor planning puts it plainly: forecasting and scheduling the workforce means ensuring the right number of people with the right skills are in the right place at the right time, and effective labor planning reduces overtime, prevents throughput bottlenecks and improves service levels. A pool is the mechanism that makes 'the right place' a variable you can actually change.

So a pool has three parts, and a site only has one of them if it is doing this properly:

  • A shared roster - one list of people, with their site, their shift pattern and their current hours, visible from every site rather than siloed in each one's own system.
  • A qualification record - what each person is trained and authorised to do, with dates on it, so a site can ask 'can this person drive that truck tomorrow' and get an answer.
  • A set of movement rules - travel, rest, working-time and cost rules that decide who may be borrowed, for how long, and at what cost, written before the emergency rather than during it.

SKUBIQ-LMS describes the same three parts as a product capability: one labor pool shared across sites with rules for travel, shift swaps and surge cover. That is the honest anchor for this article. Everything below is the operating discipline around that capability, not a claim that software removes the problem.

Line illustration of a wall-mounted labour pool board split into two columns labelled site A and site B, each column holding rows of operator badges with small shift and skill marks beside them, one badge being lifted out of the left column and moved across to the right, and a supervisor standing in front of the board holding a tablet
A pool is a shared list of people who can be moved between sites - not a second headcount to manage. Illustration generated for SKUBIQ.

The three questions every pool has to answer

Before a pool is worth anything, it has to answer three questions. They sound obvious, which is exactly why they get skipped.

Can this person actually do the work?

The operator you send has to be qualified for the equipment and the task at the receiving site, and the qualification has to be current. This is where most improvised borrowings break: the person is trained, but trained on the wrong standard, or the certification lapsed in the gap. A pool that cannot answer 'is this person qualified' in under a minute is not a pool - it is a phone tree.

How long can they be away from home?

Travel is not free and it is not only money. An operator who travels forty minutes each way loses a meaningful slice of the working day, and that lost time shows up in every productivity number you report afterwards. There is also the rest question, and the working-time question, which are legal questions rather than planning questions and therefore the ones you get wrong at your own risk.

What does the borrow cost, against what?

The comparison is never 'borrow versus nothing'. It is borrow versus the specific alternative - the overtime hours you would otherwise pay, the agency invoice, or the backlog you would otherwise clear late. Lingo Staffing's note on absenteeism and buffer planning makes the point that if you backfill with overtime instead of redeploying internally, you layer the overtime premium on top of the lost productivity from a rushed reassignment. A pool is only worth having when it beats the specific alternative you have actually priced.

QuestionWhat decides itWhere the answer comes from
Can this person do the work?Current qualification for the equipment and task, not just a job titleOne qualification record per person, with dates
How long can they be away?Travel time each way, shift length, rest and working-time rulesTravel matrix plus the attendance and time record
What does it cost?Paid hours lost, travel, and the overtime or agency rate it replacesPayroll and the rate card, per person per hour
Who owns them that day?Which site is charged, and which site gets the outputOne agreed rule, written down, for the whole group

The pool is a qualification list, not a headcount list

The most common way to build a pool is to take a spreadsheet of names and shuffle it between sites. It fails for a reason that has nothing to do with the spreadsheet: a name is not a capability.

What a borrowing site needs to know is not 'who is available' but 'who can run a reach truck, whose certification is valid until which date, and who has done this task before at a comparable volume'. Those are three attributes of one person, and they belong on the same record. SKUBIQ support is where implementation questions get answered when a site is setting this up for the first time.

There is a second reason the qualification list has to be shared rather than copied. Certifications expire. A qualification that lapses at one site and is not visible from another is a qualification that will be relied on by mistake - and that is the same failure mode as a skills matrix nobody has opened since the last induction. If the pool is going to move people between sites, the expiry date has to travel with the person.

So the minimum viable pool record carries: the person, their home site, their shift pattern, the tasks they are qualified for, the certification expiry dates, and their cumulative hours this month. CXTMS's knowledge base on labor management systems describes an LMS as tracking how every minute of an associate's shift is spent by categorising time - direct and indirect - and names the goal as maximising the ratio of direct labor to total paid hours. That is the measurement a pool has to protect. A borrowed operator who spends two hours travelling is a borrowed operator whose direct labour ratio collapses.

Travel time is the hidden cost of the pool

Travel is the line that decides whether a pool is a good idea, and it is the line most often left out of the arithmetic. A pool that ignores travel will keep making moves that cost more than the overtime they replaced.

The calculation is unglamorous. Travel minutes each way, multiplied by the number of people moved, times the number of days they are moved, is an amount of paid time that produced no output and no customer service. It belongs in the same column as the overtime it saved. If the borrow is two hours of travel for a six-hour shift, a third of the shift is travel - and if that shift was covering a peak, you have paid to move the peak.

This is why the travel rule has to be a rule and not an opinion. A workable rule caps the travel time for a routine borrow, and everything beyond the cap needs a manager to say yes with the cost in front of them. Two sites fifteen minutes apart are a natural pool. Two sites ninety minutes apart are not a pool; they are an agency with extra steps, and they should be costed as one.

Distance is not the only axis. The operator's shift pattern matters: a pool that only works when two sites happen to share a shift window is not a pool, it is a coincidence with a spreadsheet. And the receiving site's demand matters just as much - a borrowed operator who arrives at a site that has no work for them is pure cost. SKUBIQ-TMS covers the movement side of a multi-site operation, where cross-dock and hub planning lifts load factors between plant, DC and store; that is stock movement, not people movement, and the two have to be planned against each other rather than separately.

Line illustration of two warehouse buildings drawn side by side with a dashed travel route and a clock between them, a travel-time circle drawn around each building, a small van on the route and a stopwatch placed beside the clock
Travel time is the number that decides whether a borrowed pair of hands is cheaper than overtime. Illustration generated for SKUBIQ.

Cross-site shift swaps, and why self-service needs rules

The second half of a pool is the swap. Two operators at different sites agree to trade a shift. The receiving site gets cover without agency hours; the sending site gets its roster balanced. On paper it is the cheapest labour there is.

In practice, an unmanaged swap is a compliance event waiting to happen. SKUBIQ-LMS handles this as shift swap compliance - self-service swaps routed through skill, rest and working-time rules before they are approved. Read that list again: skill, rest, working time. Each one corresponds to a way a swap can go wrong that has nothing to do with goodwill.

Absenteeism is predictable, not random, once you track it by shift and day of week.

Lingo Staffing, on absenteeism and buffer planning

Line illustration of two shift cards on opposite sides of a screen being swapped by a pair of hands, a checklist column of small tick marks beside each card showing skill, rest and working-time checks, and a small shield mark beside the approved card
A cross-site swap is only self-service when the rules check it before anyone approves it. Illustration generated for SKUBIQ.
  • Skill - the person picking up the shift may not be qualified for the equipment or the task at that site. The swap looks fine on a roster and fails at the truck.
  • Rest - a swap that puts two long days in a row on the same person produces a tired operator, and fatigue shows up as errors, then as injuries, then as a productivity conversation that has nothing to do with productivity.
  • Working time - hours across sites still count against the same weekly limits and the same rest rules. A pool does not create capacity outside the law.

There is a fourth check that is not a rule but should be one: whose day does this shift belong to? If a swap moves a shift between two sites, someone has to decide which site is charged for it and which site keeps the output. Leaving that undefined is how a pool turns into an argument.

Cleverence's warehouse KPI guide recommends a balanced KPI set - dock-to-stock time, picking accuracy, on-time-in-full or perfect order rate, units per labor hour or labor cost per order, and inventory accuracy. The point for a pool is the last one: if a borrowed operator works at a different site, the output measure has to survive the move, or the receiving site will look worse than the operator actually is.

Surge cover, and the numbers that tell you the pool is working

A pool earns its keep in the week you did not plan for. That is usually a spike in one site, an absence in another, or a carrier late with a load - and the correct response is not to escalate to agency by default, but to know who can be moved, how far, and what it will cost.

Four numbers tell you whether the pool is actually working, and all four are measurable if the roster is shared:

  1. Direct labor ratio - direct hours over total paid hours. CXTMS's knowledge base states the goal directly: maximise the ratio of direct labor to total paid hours. A pool that raises travel and idle time will quietly lower it.
  2. Overtime as a share of total labor - Yourco's warehouse workforce analytics guide notes that overtime costs, measured as a share of total labor, reveal the staffing shortfalls caused by absenteeism. That is the pool's headline number: overtime share before and after.
  3. Absenteeism by shift and day of week - Lingo Staffing's absenteeism and buffer planning note is blunt about this: absenteeism is predictable, not random, once you track it by shift and day of week. A pool that is sized on predictable absence is sized correctly; one sized on random absence is overbuffered.
  4. Units per labor hour - Cleverence defines it as total units processed divided by total direct labor hours, and calls it a cross-process measure, which is why it is hard to game. Report it per site, with the borrowed hours split out.

The goal is to maximize the ratio of Direct Labor to Total Paid Hours.

CXTMS Knowledge Base, on labor management systems

Two of the numbers above need a caution. First, Pebb's workforce management guide reports an analysis from Productiv putting average annual warehouse turnover at about 49%, and Prologis reporting that 73% of warehouse operators struggle to source enough labour. If your turnover is anywhere near that, the pool is not a nice-to-have - it is how you keep a trained operator instead of re-recruiting one every other month. But it also means a pool built on people you have not retained will keep emptying, and no roster fixes that.

Second, do not confuse a bigger pool with a better pool. Yourco makes the same point about buffers in general terms: overbuffer a low-risk shift and you are paying for idle labour. A pool that is exercised twice a year is a pool that costs you management attention and pays you nothing.

When the pool is the wrong answer

Pooling is not a universal good, and a guide that only argues for it is not a guide. There are four situations where building a pool is the wrong move.

  • The sites are too far apart. If travel consumes a third of the shift, you are not borrowing labour, you are buying agency hours with extra steps.
  • The qualification sets do not overlap. If a site's operators are certified for process that the other site does not run, the pool is a list of names with no usable capability in it.
  • Demand is seasonal and shallow. A pool pays off when the peaks recur and overlap are predictable. If your sites are busy in different seasons and never in the same week, the pool will sit unused.
  • Capacity is the constraint, not labour. If the receiving site has no space, no dock slot and no equipment, a borrowed operator adds queue time rather than output.

And one more, which is the most common: the sites already run on different systems and nobody owns the reconciliation. Yourco points out that for multi-facility or 3PL operations, a standalone LMS can normalise labour data across sites running different WMS platforms. That is the technical case for a labour layer that sits above the site systems rather than inside each one - and it is a precondition for the pool working at all, because a pool cannot be managed out of four incompatible rosters.

Standing a pool up across sites: the 90-day plan

Below is a sequence that has been kept deliberately boring. The order matters: the shared roster comes before the movement rules, because rules that reference a roster you do not yet share will be enforced differently at every site.

PhaseWhat you doWhat you measure
Days 1-30 - one rosterPut every site's shifts, attendance and qualification records in one shared view. Do not move anybody yetCoverage of the shared roster; how many operators can be seen from more than one site
Days 31-60 - the travel and qualification rulesSet the travel cap, the qualification rule, the rest and working-time checks and the cost comparison. Write them downTime to answer 'can this person work this task tomorrow' from any site
Days 61-90 - one controlled borrowMake one planned cross-site move, with the cost measured against the overtime it replacedOvertime share, direct labour ratio, units per labor hour at the receiving site
Day 91 onwardExtend to swaps and to surge cover once the one borrow was defensibleAgency spend, absence cover rate, and the pool's own utilisation

Two disciplines carry the plan. The first is that every borrow is costed against a named alternative - the overtime hours, the agency rate or the late order it replaced - because a borrow that replaces nothing is a cost, not a saving. The second is that the cost of travel is logged as a travel line, not absorbed into the operator's shift, because otherwise the pool will look better than it is.

For the platform side, the pieces are unglamorous and they are already there. SKUBIQ's SaaS WMS for India covers the operating layer that a multi-client site runs on, including the 3PL functionality for managing multiple customers, billing processes and schedules - which matters here, because SKUBIQ-LMS is explicit that multi-client sites bill labour by the hour and by the task, so every shift, break and overtime hour must be measurable and every timesheet defensible. A pool multiplies the number of shifts that cross a client boundary. If the timesheet is not defensible, the pool becomes a billing dispute. The SKUBIQ FAQs, the WMS pricing page and the contact page are where a site takes those questions before it commits to a design.

What SKUBIQ-LMS and SKUBIQ-WMS each contribute

It is worth being precise about which system does what, because a pool that spans two sites and two different WMS platforms is exactly where a labour argument gets lost between systems.

SKUBIQ-LMS is the roster and the rule layer. Its published capabilities map almost one to one onto what this article has called for: one labor pool shared across sites with rules for travel, shift swaps and surge cover; shift swap compliance, where self-service swaps are routed through skill, rest and working-time rules before they are approved; rule-based overtime, break compliance and working-time handling with audit trails; and performance analytics broken down per employee, shift, site and process. It also provides the workforce planning step itself - turning volume forecasts into staffing requirements - and, for multi-client sites, the measurement that makes a shift billable and a timesheet defensible.

SKUBIQ-WMS is the work layer: the tasks, the transactions, the labour the pool is supposed to produce output from. A borrowed operator who arrives at a receiving site needs the work to be visible there, not in the sending site's system, and the output needs to be measurable against the pool's cost.

What neither system does is make the decision. The travel matrix, the cost comparison against a named alternative, and the judgement that today's absence is worth an extra two hours of travel for one operator - those stay management decisions, and they should stay ones a person makes with the numbers in front of them.

The short version

A labour pool is not a spreadsheet and it is not extra headcount. It is one shared, qualified, dated list of people with rules for travel, swaps and surge cover - and it only pays for itself when a borrow is measured against the overtime it actually replaced, travel time included.

Count the operators who can be seen from more than one site. If the answer is a number, the pool is possible. If the answer is a spreadsheet someone keeps locally, it is not. SKUBIQ-WMS supplies the work and the activity data on the receiving side; SKUBIQ-LMS supplies the roster, the qualification record, the rules and the audit trail. Everything else - the travel matrix, the cost comparison, the decision to send anyone at all - stays a management decision, and it should stay one.

#Multi-Site Labor Pooling#Labor Management System#Cross-Site Staffing#Workforce Planning#Shift Swaps#Warehouse Productivity#3PL Operations