Remote-first sounds like a tech company phrase and mostly is. But something like it has been quietly viable in multi-unit operations for years, and after 2020 it stopped being fringe. The difference between what most operators mean when they say "remote" and what actually works as a model is significant, and if you copy the former hoping it will produce the latter, the region will unravel inside a year.
Remote-first is not "traveling less." That version fails because the operating rhythm was built on the assumption of physical presence and starves without it. Remote-first is a different design, top to bottom, with different decision rights, different dashboards, different meeting cadences, and different hires.
What actually changes
In a traditional multi-unit model, the regional director is the connective tissue. They are physically present in most units most weeks. They see the shift with their own eyes. Their judgment is the working truth. Decisions escalate to them because they are the ones with the picture.
In a remote-first model, none of that is true. The general manager sees the shift. The dashboard is the working truth. Decisions get made where they land, because escalation to a leader who is not on the ground introduces delay the shift cannot absorb.
Fig. 1 · Not the same job with less travel. A different design.
The four pillars of the model
1. Decision rights are pushed down and written
The general manager has real authority on everything that has to be decided during a shift. Not because the regional director trusts them personally, but because the model does not work if the general manager has to text the regional director to approve a $200 spend at 8pm on a Saturday.
Write the decision rights down. Publish them. Update them quarterly. The general manager should know exactly what they can approve without a call, what needs a heads-up in the weekly report, and what requires the regional director's signature before anything happens. Anything in the first bucket is theirs to run without asking.
2. The dashboard is the primary source of truth
In a traditional model, the regional director's opinion of a unit is the working truth. In a remote-first model, the dashboard is. Which means the dashboard has to actually work. Daily. Automated. Accurate. Openable on a phone in one click.
What is on it: labor as percent of sales yesterday, food cost yesterday, top three items sold, comps and voids yesterday, week-to-date versus plan. Nothing more. If the dashboard grows past 12 numbers it stops being usable, and if it is not usable it is not the source of truth. Which means somebody's opinion is the truth, and you are back to a traditional model without meaning to be.
3. Peer accountability replaces manager-driven accountability
This is the biggest cultural shift and the one most operators underestimate. In a traditional model, the manager holds the general manager accountable through direct observation. In a remote-first model, that observation channel is much thinner, so the accountability has to come from somewhere else. That somewhere else is peers.
The monthly regional operating meeting stops being a leader-run review and becomes a peer-run one. Every general manager presents their numbers to every other general manager. The regional director is present but speaks last, briefly. The pressure to have a real story about your unit does not come from you anymore. It comes from the room.
Peers see through excuses that managers politely accept. A room of general managers listening to another general manager explain a labor variance is doing coaching work no manager can replicate at the same depth.
4. In-person visits become intentional
Not habitual. Intentional. Every unit visit has a purpose written down before the trip. Not "check in with the team." Something like "sit with the GM through a Friday dinner service and walk their prep list Saturday morning, so we can decide whether to change the prep model before Q3." Specific.
The result is that you travel less but every trip does more work. In year one of my last remote-first setup, I visited each unit roughly 6 to 8 times instead of the 12 to 15 a traditional model would have expected. My unit-level P&L movement was better than the traditional year prior.
What breaks first
Fig. 2 · The model handles operating metrics well. It handles growth of people less well.
Remote-first is very good at holding operating metrics. It is less good at developing people. Coaching a general manager into a bigger role is time-intensive, in-person work, and the model reduces that time by design. If you are running remote-first, you have to protect coaching time deliberately.
The rule I use: every direct report gets one dedicated in-person week per year, minimum, where the whole point of the trip is their development, not the region's operating metrics. That week is on the calendar 12 months in advance. It does not move for any reason short of a crisis.
The hiring shift
The single biggest difference in a remote-first region is who you hire. A traditional multi-unit model can accommodate a general manager who needs daily contact with their leader. A remote-first model cannot. The needy hire fails inside two quarters.
What to hire for instead:
- Prior P&L ownership. The general manager has actually run their own unit's P&L before, not been shown one.
- Comfort with asynchronous communication. They write clearly. They can hold a decision in progress without a live call. They do not treat email as a lesser channel than voice.
- A track record of asking for help specifically. Not "I have a problem." Rather "I have a problem, here is what I have tried, here is what I need." That specificity is the tell.
- Peer relationships they have built themselves. A general manager who has never developed a working relationship with a peer manager will struggle in a peer-accountability model, because they have no template for it.
The interview process is longer. References are more important than resumes. A working assignment during the interview process, where the candidate has to actually respond to a made-up operating situation over email and Slack across two days, tells you more than any hour-long in-person conversation.
Tech stack, minimally
Nothing exotic. All of it has to be tightly configured, not just installed.
- POS with clean daily data export.
- Scheduling platform with real variance reporting against forecasted labor.
- Shared operating dashboard with daily automation, openable on a phone.
- Video conferencing with recording, so anyone who missed a meeting can catch up on their own time.
- Shared document store with clear folder ownership.
- Async messaging platform with real search, so the last time somebody solved a problem can be found in 30 seconds.
The tools are not what makes the model work. The discipline around them is. A poorly configured Slack workspace is worse than none, because it produces the illusion of communication without the substance.
Where it does not work
Remote-first is not for every multi-unit region. It fails predictably in a few situations.
- New concepts still finding their operating standard. If the SOPs are not yet stable, you need physical presence to co-develop them. Remote-first assumes the standards are already known and being executed.
- Regions with a weak general manager bench. The model puts a lot of weight on the general manager. If the bench cannot hold it, no rhythm or dashboard will make up the gap.
- Turnaround situations. The first 90 days of a turnaround require in-person leadership, period. Remote-first turnarounds do not work.
- Very small regions. Under 4 or 5 units, the overhead of building a remote-first design is not worth the leverage. Just show up.
How the culture actually feels
The most surprising thing about a well-run remote-first region is how much of the traditional operating culture survives. Guests do not notice. General managers do not describe their unit as remote-operated. The daily rhythm on the floor of any given restaurant looks and feels the same as any traditional multi-unit unit.
What is different is at the leader level. Remote-first culture tends to reward writing. Decisions get made in a thread that anyone can search later, not in a hallway conversation that only the two people present remember. Meetings are shorter because most preparation happens in advance in writing. Coaching feedback shows up in one-on-one video sessions and gets referenced in the next written update. The culture becomes more legible, in the sense that new members of the leadership team can catch up on the history by reading. That legibility is a real cultural asset that traditional models rarely produce.
The transition problem
The hardest thing about remote-first is not running it. It is transitioning to it from a traditional model. Existing general managers were hired against a different set of expectations. They built their operating instincts on the assumption of frequent physical contact with leadership. Removing that contact without replacing the reassurance it provided produces a specific kind of anxiety that hurts performance in the first two quarters of the transition.
If you are transitioning an existing region to remote-first, do it over two quarters, not overnight. Start with the meeting cadence and the dashboard. Introduce the decision rights framework in month two. Reduce in-person visits in month three, not month one. Give each general manager one deliberate in-person coaching week in the first quarter so they experience the new intentional model of your presence before they experience the reduced frequency of it. Handled well, transition to remote-first can be net positive by month six. Handled badly, it produces attrition inside a quarter.
Communicate the transition explicitly. Do not just show up in fewer units. Tell the field what is changing and why. A general manager who reads a memo that names the shift to remote-first as a deliberate design choice reacts differently than a general manager who quietly notices their regional director has stopped coming as often. The design deserves to be named.
The point
A remote-first multi-unit region is not a traditional region with less travel. It is a different operating design, with decision rights pushed down, a real dashboard as the source of truth, peer accountability doing the work manager accountability used to do, and in-person time reserved for coaching depth that the model cannot otherwise deliver.
Built deliberately, it lets a regional director cover geography and hold a life outside of work in a way the traditional model does not. Built accidentally, it produces a slow-motion crisis. The choice is not whether the model can work. It is whether you are willing to redesign the whole system to let it.