How Many Employees Should I Schedule Each Shift at My Poke Restaurant in 2027?

Two ways to staff the line: fixed crew versus gross-profit division

Almost every poke operator lands in one of two camps, and the difference between them is worth several points of labor cost per year.

Option A — the fixed crew. You decide the shop runs "four on days, five on nights," publish the same grid every week, and adjust only when someone calls out or a holiday hits. This is how the majority of single-unit fast-casual restaurants actually schedule. It is fast to build (ten minutes on Sunday night), it is predictable for the crew, and it makes availability conversations simple because everyone knows their block. The cost is that it ignores what the register actually did. A Tuesday that rings $400 in sales gets the same four bodies as a Thursday that rings $1,100. On the slow day you are paying roughly two people to restock sauce bottles; on the busy day you are pushing a two-minute line out the door and losing the walk-aways you never see on a report.

Option B — gross-profit division. You set one number — the gross profit an average employee should produce on an average shift — then divide each shift's trailing average gross profit by that number to get the headcount. Nothing else decides. Not seniority, not "we've always run six," not the assistant manager quietly stacking friends onto the easy blocks. The schedule becomes arithmetic, which is exactly why it works: arithmetic doesn't have favorites and doesn't get talked out of a decision on a Sunday afternoon.

The two options are not equally good, but they are not equally *easy* either, and that matters. Fixed crews cost you nothing to maintain. Gross-profit division costs you about 45 minutes of setup and 15 minutes a week of upkeep, and it requires you to actually pull sales by daypart out of your POS — which, for some operators, is the real barrier. If you cannot get hourly sales out of your point-of-sale system, you cannot run Option B honestly, and a half-run version of it is worse than a clean fixed crew because it gives you false confidence.

There is a third posture worth naming so you can reject it deliberately: staffing to labor percentage alone. Plenty of operators chase a flat target — "keep labor under 28%" — and let that number pick the headcount. The problem is that labor percentage is an outcome, not an input. It tells you after the fact whether the shift was efficient; it cannot tell you on Wednesday how many people to put on Saturday, because you don't know Saturday's sales yet. Percentage is a scoreboard. Gross-profit division is a play-caller. Use both, but do not confuse them.

For a poke shop specifically, Option B has a structural advantage: your product is assembled to order on a visible line, so throughput is almost linear in bodies up to the physical width of the counter. Add a builder and you genuinely move more bowls per minute, up to the point where people start bumping elbows. That is not true in a full-service Italian restaurant where the constraint is the kitchen. It means the division method maps cleanly onto how your shop actually produces revenue, and the number it spits out is usually the number you should staff.

How to decide which method your shop should run

The decision hinges on three things: whether you can get clean daypart sales data, how much your volume actually varies day to day, and how big your roster is. Work through them in order rather than picking a method by preference.

Test one — variance. Pull the last eight weeks of daily net sales. Find your median day, then find your best and worst. If your best day is less than about 1.4× your worst day, your volume is flat enough that a fixed crew loses you very little; the division method would hand you nearly the same headcount every shift anyway. If your best day is 2× or more your worst — which is normal for poke, where weekday lunch and weekend dinner live in different universes — a fixed crew is structurally guaranteed to be wrong on both ends, and division pays for itself immediately.

Test two — data access. Can you export sales by hour, by day of week, for a trailing 90 days? Most modern restaurant POS systems do this in two clicks. If yours does not, you can rebuild it by hand from daily close reports, but budget an evening for it, and know that you will only have day-level granularity, not hourly. Day-level is enough to run the method; hourly is what lets you place people *within* the shift instead of just counting them.

Test three — roster depth. Division only helps if you have enough part-time employees to flex the schedule. If your entire staff is four full-timers who each need 38 hours to stay, you have no flex; the schedule is already determined by hour guarantees. In that case run the fixed crew, and use the division math as a *hiring* signal instead — it will tell you whether your next hire should be a lunch-only part-timer or a weekend closer.

One more decision input that operators skip: predictive-scheduling rules. Several cities and states require advance notice of schedules and pay penalties for late changes to a published shift. If you operate under one of those rules, the flex you gain from division has to be built *into* the published schedule, not applied as day-of cuts. That does not disqualify the method — it just means you commit to the numbers when you publish, and you stop treating "send someone home early" as your adjustment lever.

The numbers behind each option, worked all the way through

Here is the arithmetic, with real ranges you can check against your own P&L.

Setting the per-employee target. Sit down with whoever helps you run the shop and agree on one figure: the gross profit an average employee should produce working an average shift serving an average number of guests. For a fast-casual poke restaurant, $160 per shift is a workable starting floor. Size it to your actual margins — fresh fish is your single largest food cost line, rice yield varies more than people expect, and the protein scoop is where your gross profit is won or lost. If your food cost runs high because you are buying sashimi-grade tuna at a small-volume price, your target has to come down; if you run a tighter menu with more rice and vegetable volume per bowl, it can go up. Most operators land somewhere between $120 and $250 depending on ticket average and market.

Say it out loud to the crew, because the number does double duty: "If you show up, take care of an average number of guests, and give average service, you should produce no less than $160 a shift in gross profit." The employees who want real hours and real tips do not coast to $160 and clock out. They hit $160 on average work and then build one more bowl, upsell one more side, turn one more guest.

Getting gross profit per shift. Take the shift's average net sales and subtract cost of goods sold. Poke COGS typically runs in the low thirties as a percentage of sales, higher if your protein mix is heavy on premium fish. Worked example: a Friday dinner averaging $1,800 in net sales at 32% COGS gives you $1,224 in gross profit. Divide by a $160 target and you get 7.65 — call it 8 people. A Tuesday lunch averaging $940 at the same COGS gives $639 in gross profit, which divides to 4 people. That is the whole method. Two shifts, two very different answers, from the same one-line formula.

What the fixed crew costs by comparison. Suppose you run five people on every shift, fourteen shifts a week, at an average loaded wage of $19 an hour including payroll taxes, and each shift is eight hours. That is 560 labor hours a week, about $10,640. Now run the division method and assume it moves you to four on the seven weaker shifts and six on the seven stronger ones — same total bodies, redistributed. Your total hours do not change at all, but the hours are now sitting on the shifts that can pay for them. In practice the method usually also trims total hours, because at least two shifts a week turn out to support three people rather than four. Two shifts × one person × eight hours × $19 is $304 a week, roughly $15,800 a year, on a change that took you an afternoon to set up.

Where the ratio actually falls. Most poke shops do not split evenly between dayparts. Lunch commonly runs 35–45% of daily revenue and dinner 55–65%, though a shop next to an office park can invert that completely. Run a four-week average of your own hourly sales before assuming anything. If dinner consistently outruns lunch, your dinner crew should be meaningfully larger — not the same size, which is the default most schedules drift into.

Opening and closing load. Prep and cleanup are real hours that the service-count formula does not capture. Budget one to two people beyond the service number to cover the open and the close. Most poke prep — rice batches, protein portioning, sauce and topping mise — fits in a 90-minute window before the doors open. If your prep crew is overlapping the 11:00–1:00 rush for two hours because that is how the schedule was always built, you are paying for standing around. Stagger the starts: prep in at 9:00, service crew at 10:30. Fold those prep and close hours into your gross-profit-per-employee math so the target reflects total labor, not just the people on the line.

Starting with no history. A brand-new shop has nothing to average. Start with roughly one employee per $100–$150 of projected sales per shift, run four weeks, then switch to the real numbers. Start slightly lean and add rather than starting heavy and cutting — adding an hour is a text message; cutting one damages trust with the employees you need to keep.

Implementation and sequencing: what to do in what order

Do not try to convert the whole schedule in one week. Sequence it so each step produces a number you can check before the next step depends on it.

Week one — build the baseline. Export trailing 90 days of sales by day of week and by hour. Compute average net sales for each shift, apply your COGS percentage, and write the gross profit into a simple grid: rows are days, columns are dayparts. Do not schedule off it yet. Just look at it, because the first honest look at that grid usually contradicts something everyone in the building believes about which shifts are busy.

Week two — set and socialize the target. Agree on the per-employee gross profit number with your leadership, then divide the grid by it. You now have a headcount for every shift. Compare it against what you have actually been scheduling. Expect to find one or two shifts overstaffed by two people and at least one understaffed. Tell the crew the number and the logic before the schedule changes — a schedule cut that arrives with no explanation reads as punishment, and the same cut explained as arithmetic reads as management.

Week three — publish the new counts, keep the same faces. Change *how many* people are on each shift before you change *which* people. Move hours, not employees. This keeps the trust intact while the numbers prove themselves, and it isolates the variable so you can tell whether the new counts actually held service quality.

Week four — place bodies inside the shift. Now use your hourly data. If the lunch wave hits at 11:30 and dinner pickup at 6:30, stack builders and a runner into those windows and thin the 3:00 p.m. lull rather than parking everyone flat from open to close. On a fast-casual poke line the working configuration at peak is typically a checker or expo, two to three bowl builders, and a register; off-peak collapses to one on the line and one on the register.

Ongoing — recompute every four weeks, and watch the leading indicators. Ticket times over three minutes at peak, or a visible line that stops moving, means you cut too deep. Employees visibly idle for stretches at peak means you have room. Both signals beat the labor percentage report, which arrives a week late.

On tooling. Any scheduling app can publish a grid; the useful distinction is whether it connects to your sales data. 7shifts is built specifically for restaurants and ties scheduling to POS sales and labor targets. Homebase prices per location rather than per employee, which matters when your roster is a dozen part-timers. When I Work and Deputy both handle availability, swaps, and mobile clock-in well, with Deputy leaning harder into demand-based suggestions and compliance guardrails. Sling and Connecteam cover scheduling plus crew messaging cheaply for a small team. ZoomShift is a simple mobile-first option for a single unit. Check current pricing on each vendor's own site before you commit — plans and tiers change. None of these tools will tell you the Friday rush needs eight people. You bring the gross-profit headcount; the tool handles publishing, swaps, and cost tracking.

Related questions

What labor cost percentage should a poke restaurant target?

Fast-casual restaurants commonly aim for labor in the high twenties to low thirties as a percentage of sales, though it varies widely by market wage and menu price. Treat it as a scoreboard you check after the fact, not the input that picks your headcount.

How do I handle a shift that suddenly gets slammed?

Build a short on-call list of employees who want extra hours and can be at the shop in 30 minutes. Text one, not three. Check whether local predictive-scheduling rules apply to same-day additions before you make it a habit.

Should the manager count toward the shift headcount?

Only if they work the line. A manager expediting bowls at peak is a body on the schedule; a manager doing inventory in the back office is overhead. Count the hours where they actually produce, and be honest about which is which.

How often should I rebuild the schedule from scratch?

Recompute headcounts every four weeks and rebuild the underlying gross-profit grid quarterly. Seasonal shifts, a new neighbor tenant, or a menu price change all move the numbers enough that a stale grid quietly reintroduces the problem you solved.

What if two employees always want the same shift?

Set the headcount first, then allocate. The count is arithmetic and non-negotiable; who fills it is a management call. Rotating the desirable shifts on a fixed cycle removes the weekly argument and stops the schedule from becoming a favor economy.

FAQ

What gross profit per employee should I target at a poke restaurant?

Most operators land between $120 and $250 in gross profit per employee per shift. A higher-ticket urban shop can carry the top of that range; a smaller suburban location with lower prices should sit near the bottom. Set the number so it covers the loaded wage plus a real margin, then hold it steady for at least a quarter so you can tell whether the schedule or the target is the thing that needs adjusting.

How do I calculate gross profit for one specific shift?

Take that shift's average net sales over the trailing four to six weeks, then subtract cost of goods sold. Poke COGS commonly runs around 30–35% depending on your protein mix. A dinner shift averaging $1,800 at 32% COGS yields about $1,224 in gross profit. Divide that by your per-employee target to get the headcount for the shift.

Should lunch and dinner use the same per-employee target?

Not necessarily. Lunch tends toward lower tickets and faster turnover, so employees clear guests quickly and can carry a higher target. Dinner brings larger orders and more customization, which slows the line per guest. Many operators run a slightly lower target at dinner to protect service quality, and that is a legitimate adjustment as long as you decide it deliberately rather than drifting into it.

What if my sales swing wildly day to day?

Use a rolling four-to-six-week average per shift instead of any single day. That smooths one-off spikes and dead weather days. For genuinely unpredictable shifts, schedule a core crew at the low end of the range plus one on-call employee who wants the hours — but check your local predictive-scheduling rules before relying on same-day adjustments.

Do prep and closing employees count in the formula?

Include their hours in your labor cost, but schedule them separately from the service count. Prep and close typically add one to two people beyond what the service formula returns. Stagger prep to start about 90 minutes before doors open so those hours do not overlap the rush, where they add cost without adding throughput.

Can I use this method before I have any sales history?

Yes, with a placeholder. Plan roughly one employee per $100–$150 of projected sales per shift for the first month, track actuals closely, then replace the estimate with your real trailing average. Starting slightly lean is safer than starting heavy — adding hours is easy, and cutting hours from people you just hired damages the roster you need.

Sources

• https://www.bls.gov/iag/tgs/iag722.htm — U.S. Bureau of Labor Statistics, food services and drinking places industry data
• https://restaurant.org/research-and-media/research/ — National Restaurant Association research and industry operations reports
• https://pos.toasttab.com/blog — Toast restaurant blog, labor management and scheduling guidance
• https://www.7shifts.com/blog — 7shifts resource center, restaurant scheduling and labor forecasting
• https://squareup.com/us/en/townsquare/restaurants — Square restaurant resources on labor cost and staffing
• https://www.dol.gov/agencies/whd/flsa — U.S. Department of Labor, Fair Labor Standards Act wage and hour rules
• https://www.sba.gov/business-guide/manage-your-business/hire-manage-employees — U.S. Small Business Administration guidance on hiring and managing employees
• https://www.homebase.com/blog — Homebase blog, small-business scheduling and time-tracking practices

Related on PULSE

• How Many Employees Should I Schedule Each Shift at My Mediterranean Restaurant?
• How Many Employees Should I Schedule Each Shift at My Thai Restaurant?
• How Many Employees Should I Schedule Each Shift at My Vegan Restaurant?
• How Many Employees Should I Schedule Each Shift at My Indian Restaurant?
• How Many Employees Should I Schedule Each Shift at My Korean BBQ Restaurant?
• How Many Employees Should I Schedule Each Shift at My Hot Pot Restaurant?