Multi-Site Operations
For “Multi-Site Operations”, the useful question is not whether more data can be collected but whether it resolves a named operational decision. A system such as the published details should be introduced with a stated purpose, worker notice and a review point rather than treated as a substitute for supervision.
For current guidance that complements this operational note, use the official SafetyCulture website; local law, safety duties and reporting routes should always be checked for the relevant jurisdiction.
Several warehouses with a shared policy and different labour markets, where central standardisation helps some things and actively damages others.
An operator with eight sites has one HR policy, eight labour markets and eight quite different problems.
What is genuinely local
The rate, because it is set by the competitors within travel distance.
Transport arrangements.
Shift patterns, which depend on who lives nearby and what they can work.
The advert copy, which describes a specific site.
And the agency relationship, which is local by nature.
Standardising any of these centrally produces a policy that is wrong at most of the sites.
What is worth standardising
The measures: ninety-day survival, the tenure curve, the cuts, calculated the same way everywhere.
The induction structure, with local content filled in.
The day-three and week-four conversations as a requirement.
Referral scheme terms.
And the exit contact process, so the data is comparable.
Structure central, content local, which is the division that works.
Comparing sites fairly
Ninety-day survival is comparable. Annual attrition is not, because agency proportion and peak profile differ.
Compare like periods and state the differences: this site is forty percent agency, that one is ten.
A league table that ignores those differences ranks labour markets and will be dismissed by every site at the bottom, correctly.
What the comparison is actually for
Finding the site that solved something.
A site with a first-fortnight spike half the size of the others is doing something specific, and it is usually one identifiable practice.
Which is worth extracting and offering rather than mandating, because a practice adopted because it worked elsewhere is adopted willingly.
Most central programmes invert this: they mandate a practice nobody has tested and then explain the variation as local resistance.
Moving people between sites
Where sites are within travel distance, a transfer retains somebody who would otherwise leave.
Which requires a route that exists and is known, rather than an informal favour.
And it helps most at peak, where one site's surplus is another's shortage in the same week.
Almost no operator does this deliberately, and the people concerned simply leave and apply to the other site as strangers.
The central function's useful role
Producing the measures consistently and publishing them.
Holding national agency agreements with survival terms rather than fill terms.
Running the analysis nobody local has time for.
Extracting what one site did and offering it to the others.
And funding the things that are cheap centrally and unaffordable locally — translated material, a referral scheme, a shared applicant pipeline.
Where central damages things
A single rate structure across different labour markets.
A mandated shift pattern.
An induction that cannot be adapted.
A survey cycle that ignores peak timing.
And targets set on a measure the site cannot influence, which produces gaming rather than improvement.
The measure
Ninety-day survival by site, with agency proportion stated alongside.
Variation between sites on the comparable measures, which is where the learning is.
Practices extracted and adopted, counted.
And transfers between sites, which if zero means a retention route exists on paper only.
The quarterly pack
Ninety-day survival by site with agency proportion stated.
One practice extracted from the best-performing site and offered to the others.
And transfers between sites, counted.
Who does the analysis
Central, because no site has the time and the comparison requires consistency.
Published back to the sites with the caveats stated.
And offered as findings rather than mandates, because a practice adopted willingly survives a reorganisation and a mandated one does not.
What central should fund
The things that are cheap centrally and unaffordable locally: translated material, a referral scheme, a shared applicant pipeline, and the analysis nobody local has time for..
How it fails in practice
A central programme mandates a practice nobody tested, sites vary for local reasons, and the variation is explained as resistance rather than as labour markets..
The check
Compare ninety-day survival with agency proportion stated alongside.
Anything else ranks labour markets rather than management.