What to Centralize and What to Leave to Each Location

Why this matters

Every multi-location shop drifts toward one of two failure modes. One owner centralizes everything and slowly strangles each location, so a crew waits on a distant approval to do something obvious and customers feel the lag. The other leaves everything local and ends up running several small businesses that happen to share a name, paying for duplicate everything and getting inconsistent service across sites. Both are expensive. This is the map that keeps you out of both ditches: a principle for deciding, a list of what almost always goes which way, and the tells that you have overcorrected.

The principle

There is one rule underneath the whole decision:

Centralize what gets cheaper and more consistent at scale. Localize what needs local judgment and speed.

Some things get better the more you run them as one: buying power, a single standard, one clean set of numbers. Other things get worse the further you move them from the customer: the same-day judgment call, the relationship, the read on a local market. Centralize the first kind. Localize the second. When you are unsure which kind a function is, ask what you lose by moving it: money and consistency, or judgment and speed. That answer decides it.

What almost always centralizes

These share a trait: little local judgment, real gain from doing them once.

• The brand and the standard. What the company stands for and what "done right" means are set centrally so a customer gets the same shop at every location. This is the whole point of being one company.
• Back-office finance. Billing, collections, payroll, and the books run once, cleaner and cheaper, and keep the numbers in one place where you can actually see the business.
• Purchasing leverage and vendor deals. Negotiated centrally to buy as one large shop instead of several small ones.
• The data and the scorecard. The measures that let you compare locations and catch drift have to be defined and collected the same way everywhere, or they do not compare.
• Training design and the playbook. The system for bringing people up to standard is built once and shipped to every site, so quality does not depend on who happened to train the new hire.

What almost always stays local

These share a trait: they live or die on local judgment, speed, or relationships.

• The daily dispatch call. The in-the-moment reshuffle when a job runs long or a customer needs a squeeze. A local dispatcher who knows the crews and the streets beats a distant screen.
• Customer relationships. The repeat customer wants to be known by name by someone who remembers their last job. That does not centralize.
• Crew management day to day. Coaching, scheduling the team, handling the small stuff belongs to the local lead who is standing there.
• Local market decisions. Which neighborhoods to push, how to respond to a local competitor, the read on local pricing pressure. The person on the ground sees this first.

The gray zone and how to decide

Some functions have both a system half and a judgment half. The move is to split them, not to force the whole function one way:

• Scheduling: central owns the rules and the overflow desk; local owns the real-time judgment.
• Intake: central sets the standard and catches after-hours overflow; local handles the customers it knows.
• Purchasing: central negotiates the deal and sets the stocking standard; local pulls the daily replenishment.

The pattern: central owns the standard, local owns the call. When a function feels like it belongs in both places, that is usually because it does, split along that line.

The failure modes to watch

• Over-centralized: locations wait on approvals for obvious things, local judgment atrophies, good local people leave because they have no room to run, and customers feel the lag. The tell is a distant crew that has stopped making decisions it is perfectly capable of making.
• Over-localized: duplicate staff and cost, inconsistent service site to site, lost buying leverage, and numbers you cannot compare because each location tracks them differently. The tell is that your locations feel like separate companies.

If you see the first set of tells, push authority back down. If you see the second, pull the standard and the back-office back up.

The spine to keep

Centralize what gets cheaper and more consistent at scale (brand, back-office, buying, data, training). Localize what needs judgment and speed (daily dispatch, relationships, crew management, local market calls). For the in-between, central owns the standard and local owns the call. Hold that line and you get the leverage of one company with the responsiveness of a local shop, instead of the drag of neither.

References

• U.S. Small Business Administration (SBA): multi-location structure and operational delegation
• Trade-standard practice on centralized versus local service operations
• See related: Centralize Dispatch and Office or Keep It Local (decision tree); The Playbook That Makes a New Location Repeatable