Estimated Reading Time: 12 minutes

TL;DR: To scale safety management across multiple manufacturing facilities, standardize the definitions, the minimum controls, and the data model at the corporate level. Then let each site own how it executes them. Name one accountable owner inside the line of power at every location, and hold every site to the same data dictionary. Copying one plant’s binder to the next plant is the part that fails.

Here’s how to scale safety management across multiple manufacturing facilities: standardize what must be true at every site, and let each site decide how it gets there. Corporate owns the definitions, the minimum controls, and the numbers. The plant owns the execution.

Most people go the other direction first. You build a program that works at one plant; it works well, so you take the binder to the second plant…and it lands like a mandate from a stranger. Now you’re spinning too many plates across 4 sites with 4 different versions of the same program, and you’re still only one person.

The instinct everybody walks in with is the gap list. I learned what that costs the expensive way.

Key Takeaways

  • Standardize the definitions and the minimum controls, and leave the procedures to the site. Corporate sets the floor; the site decides how it stands on it.
  • A program you have to personally run at every site is a bottleneck wearing a corporate title.
  • Every location needs a named owner who sits inside the line of power, with hours in writing.
  • One shared data dictionary is worth more than any dashboard. If “near miss” means something different at each plant, your enterprise numbers don’t mean much.
  • Verify controls in the field. What you audit is what the site keeps doing after you leave.

Why the Copy-Paste Approach Breaks at the Second Site

Scaling safety fails most often because the thing being copied is the wrong layer. A binder of site-specific procedures is an output… the record of decisions one plant made about its own equipment, shift pattern, and people. Hand that to another plant and you’ve handed them somebody else’s answers to questions they never got asked.

The first new site I ever walked into

My first acquisition, back in 2001, was a 100-year-old company. I walked in and saw rusted ammonia refrigeration pipes and a laundry list of safety issues.

And I did exactly the wrong thing. I berated them about everything that was wrong. Every violation, every hazard, every gap (I was VERY thorough).

These people had built something over a hundred years and had real pride in it. I came in like I was there to grade them, and they failed.

It was awful. Absolutely awful. I destroyed the relationship before it even started, and it was incredibly hard to rebuild.

So when you inherit a second location, the gap list is the easy half. How you show up sets the ceiling on everything you get to change afterward.

What actually travels between sites, and what doesn’t

Definitions travel. Minimum controls travel. So do data models, escalation rules, and competency requirements.

Procedures don’t. Neither do inspection routes, meeting times, forms with a machine name on them, or the way you word a coaching conversation on a floor you’ve never worked.

The test is simple. If it would still be correct at a plant you’ve never walked, it’s a standard. If it needs a plant walk to be correct, it’s execution, and execution belongs to the site.

The corporate mandate that got a checkmark and nothing else

I’ve been on the receiving end of this too. Corporate handed down a mandate to train all of our employees in a back injury prevention program. No choice, no input, just here it is (nobody asked us, either).

There was pushback from day one. We only did it because we were forced to, so nobody bought in.

We went through the motions, got our checkmark, and moved on. Lots of money spent, lots of time chasing buy-in, no change in the results at all…

A few years later we were staring at our lifting injury numbers, and somebody said, hey, what about that training we did? This time we asked first: the safety committee, the supervisors, employees in the break room.

Same company. Same problem. Different authors.

We rebuilt it around what they told us, and that one actually got used.

Inside one plant, you can carry a weak rollout on relationships. Across 4 or 12 or 22 locations that carrying capacity runs out. Doing everything yourself feels like job security, right up until a one-person safety department becomes a single point of failure for the whole company.

Standardize the What, Let Each Site Own the How

Standardization means every site meets the same defined floor, using whatever method fits its operation. It doesn’t mean every site runs identical paperwork.

Think of it like a recipe: the ingredients are fixed, and the kitchen is yours. Corporate writes the ingredient list. The site cooks it, and owns how.

And yes, that back injury mandate came from corporate too. The difference is what got mandated.

The floor is a WHAT. That program was a HOW, written by people who had never stood in our building and handed down as finished.

What corporate owns

Corporate owns the enterprise framework:

  • Hazard identification and risk assessment
  • Management of change
  • Lockout/tagout and machine guarding
  • Contractor controls
  • Incident and near miss reporting
  • Emergency preparedness and corrective actions
  • Training competency

For each one, corporate defines the minimum acceptable control, the documentation, the responsible role, and the verification method. Sites can go stricter. Sites can NEVER go under.

What the site owns

The site owns permits, local hazard assessments, its own workforce conversations, emergency plans, and every procedure that comes out of them. It owns the schedule too…a plant running 3 shifts and a plant running a single shift don’t need the same cadence to hit the same standard.

The 2 models side by side

Copying the program Standardizing the standard
What gets shipped Finished procedures and forms Definitions, minimum controls, data model
Who wrote it The first site (or you) Corporate writes the floor, the site writes the method
What a new site does Adopts, or quietly ignores Builds its own version against the floor
Where it breaks The first exception nobody planned for Rarely, because exceptions are the site’s job
What you audit Whether the binder matches Whether the control works in the field
Your role Owner of every procedure Guide on the side and verifier

Where the Safety Management Cycle fits

The Safety Management Cycle is the piece you standardize across every site, because it’s a system rather than a procedure. Identify, Develop, Implement and Train, Coach and Observe, Analyze.

Every site runs all 5 stages, and what a site does inside a stage is its own call…a plant that hand-stacks totes has different hazards to identify than one running 6 automated lines.

Standardizing the cycle gives you shared language for where each location actually is. When a plant manager says the program is stalled, you can ask where instead of guessing.

A site starting from nothing runs the same safety management system implementation steps you’d run at headquarters. And integrating its JHAs with its SOPs is where that cycle starts producing something the floor can use.

Put a Named Owner Inside the Line of Power at Every Site

Every location needs one person accountable for safety execution there, by name, with time and authority attached. Not a volunteer, and not “the plant manager, technically.”

Most companies skip it because it feels like headcount. It usually doesn’t… it costs a job description edit and 2 hours a week.

So why can’t that owner be you?

The first time a corporation hired me over multiple locations, I had a dotted line to corporate and a solid line to both facility bosses. When corporate called and asked me something, I answered honestly.

That got me called into my boss’s office and screamed at. They didn’t trust me after that, and they stopped including me in many conversations.

I’d answer the same way again. But I was the only safety person either site had, which meant every honest answer I gave corporate cost me something locally.

That’s not a personality clash. It’s structural. Safety is an auxiliary position. We sit outside the chain of command, so a correction from us never carries the weight of one from inside the employee’s line of power.

Our job is to create the resources, train the people who use them, audit what’s happening against policy, and report the program’s health to executives. That’s the guide on the side, and it’s the ONLY thing that works at 12 locations.

Try to be the safety presence at every site personally, and you become the reason nothing moves. Programs stall the week you’re out, and leadership reads slow as ineffective.

What to actually hand the site owner

Here’s where this goes wrong. Companies deputize a superintendent, call him the safety guy, hand him nothing, and then act surprised when it dies.

Give them a block of time in writing that their supervisor agreed to, the corporate standard, the site’s own gap list, one form, and one cadence.

Then connect them. A monthly call where site owners compare what they tried is how a fix at plant 3 reaches plant 7 without routing through you. It’s the least glamorous 30 minutes on your calendar, and it beats every binder you’ll ever write.

Accountability is where most site-champion programs quietly die. The method for holding supervisors accountable for safety without direct authority works the same way on site owners.

What this looks like when it works

Kelly Dunn is the Director of Safety at a consulting, architecture, and engineering firm with 130 offices across the US and Canada. Different industry, identical constraint…130 locations, one of her.

So she implemented an Office Safety Coordinator plan across the organization. Instead of scaling herself, she put a named person in each office and shifted her own job from directing people to influencing them.

Distributing the work freed her up to change how she talked to leadership, and that’s the part she names herself: “This is the first time I think safety has really been involved in senior leadership. And I think it’s because I’m approaching it with that business mindset of what they want to hear.”


THE ALL-ACCESS PASS RESOURCE PAGE

Get all the FREE templates, safety management resources, PDFs, spreadsheets, and more...

One Measurement System Every Site Reports Into

Ask 4 of your plant managers what counts as a near miss. You’ll get 4 different answers, and at least one of them will start with “well, it depends.”

That’s what separates an enterprise safety program from a folder of unrelated site reports. Without shared definitions, you’re running a company-wide game of telephone, and the last person in line is a VP holding your numbers.

Build the data dictionary before you build the report.

Get the mix right, not just the mat.h

Aim for roughly 75% leading indicators and 25% lagging. Lagging data has 3 problems that get worse with every site you add: it barely moves, it’s all negative, and it hits a hard floor at zero.

Leading indicators have none of those limits, and most of that data is good news…which changes what it feels like to walk into the quarterly review. So track:

  • Control verifications completed on schedule
  • Corrective actions closed by their due date
  • Hazard reports per site
  • Supervisor safety conversations
  • Preventive maintenance on safety-critical equipment

And don’t rank plants by injury rate alone. A site with almost no reports and no participation is usually doing worse than a site with more reports and an engaged workforce, and the rate will tell you the opposite. That’s the case for building the review around leading indicators.

Compare like sites to like sites

According to BLS injury and illness rates by industry, manufacturing recorded a total recordable case rate of 2.7 and a DART rate of 1.7 per 100 full-time workers in 2024. Private industry overall came in at 2.3, which the BLS 2024 injury and illness news release calls the lowest rate in the series going back to 2003.

Use that as a broad floor, not as a peer group. Your real peer group is your own plants, grouped by comparable operations.

Food production against food production. Metal fabrication against metal fabrication. Benchmark a distribution center against a machining plant and somebody looks like a hero or a problem for reasons that have nothing to do with their program.

Where the OSHA 300 Logs actually live

OSHA’s rule on multiple business establishments requires a separate OSHA 300 Log for each establishment expected to operate a year or longer. Employees who float between locations get linked to one home establishment.

You can keep it all centrally, and most multi-site companies should. The condition is that information reaches the central location within 7 calendar days, and that you can send records back to any establishment within 4 hours on request.

Decide that early and on purpose. Retrofitting it after you’ve added a plant is a genuinely miserable week.

Your Site Standardization Checklist

Work this in order. Skipping to the rollout is the most common way multi-site programs stall.

Before you roll anything out

  • Write the data dictionary: every term, one definition, signed off by corporate.
  • Define the minimum control, documentation, responsible role, and verification method for each program element.
  • Group your sites for comparison by operation type, and decide where the OSHA 300 Logs live.
  • Get the site owner role into a job description, with hours attached.

At each site

  • Walk the plant before you present anything, and lead with what they’ve built instead of the gap list.
  • Run the hazard assessment with the site’s own people, then let them write the procedures against the standard.
  • Name the site owner and introduce them to leadership.
  • Set the local cadence with the plant manager, not for them.

Every quarter

  • Verify controls in the field. Watch the guard, the lockout, the conversation.
  • Close out overdue high-risk actions before anything else.
  • Run the site owner call and move one fix horizontally.
  • Update the standard wherever 3 or more sites hit the same wall. That’s a corporate problem, not 3 site problems.

Frequently Asked Questions About Multi-Site Safety Management

How many facilities can one safety manager realistically support?

There’s no fixed number, because the constraint is whether the sites run their own execution. A safety lead supporting 15 locations with named site owners and a shared standard has an easier job than one personally running 3 plants.

Should each plant have its own safety procedures?

Yes, and that’s the design working. Corporate owns the definitions, the minimum controls, and the verification method. The plant writes the procedure that meets that floor with its own equipment and shifts, and identical paperwork across different operations is a warning sign.

How often should I visit each site?

Quarterly is a workable baseline. When I was the Safety Director over 22+ locations, I visited every domiciled location at least once a quarter, including sites with only 2 employees (yes, that’s a lot of windshield time). Clump nearby locations into regional swings, and go to verify and be seen, not to run the program.

What if I don’t have the authority to standardize anything?

Start with the data dictionary. It needs no budget, no approval, and no title, and once every site uses your definitions, you become the person whose numbers leadership trusts. Authority follows whoever holds the only reconcilable data in the company.

What’s the first thing to standardize?

Definitions. Not forms, not procedures, not the audit tool. Until “recordable,” “near miss,” and “closed corrective action” mean the same thing at every plant, every number you roll up is a guess.

Now It’s Your Turn

Scaling safety management across multiple manufacturing facilities comes down to one early decision: standardize the floor, and let every site own how it stands on it. The program that scales is the one that runs when you’re not in the building.

Do this week:

  1. Write your data dictionary. One page, every term, one definition.
  2. Pick your worst site and ask what makes the standard hard to follow there. Not what allows them not to follow it…what makes it hard.
  3. Name one site owner and put the hours in writing.
  4. Group your plants into comparison sets and rerun last quarter’s numbers inside the groups.

None of those 4 need a budget, an approval, or a title you don’t have yet.

So picture next March, when none of it happened. Same 4 versions of the program, plus whatever site got added since. A VP asks why plant 3 looks better than plant 1, and you can’t answer, because the 2 plants were never counting the same things.

And the incident that finally lands is going to land at the site you’ve walked exactly once. That’s a design problem, and design problems are fixable.

If the layer underneath those 4 steps feels shaky, say so out loud. Most of us got handed regulations and never got handed a system…I built a free training on the Safety Influencer System that walks the whole structure those 4 steps sit on.

You got this, my safety friend!

Hi, I'm Brye (rhymes with sky)!  I am a self-proclaimed safety geek with two decades of general industry safety experience.  Specializing in bringing safety programs to a world-class level and building a safety culture, I have trained and coached many safety managers, just like you, on how to effectively manage workplace safety in the real world.   I would love to help you too.

Get started with my weekly newsletters: