1. Design partnership
A three-to-six-month engagement, scoped and quoted in writing before you start. This is separate from the ongoing subscription.
The economics.
Know what you pay. Then measure what comes back. Software is $149/month per site after your design partnership. Managed procurement is included with a 0.9% fee on purchasing routed through MagicMenu.
Your ongoing software cost
$149/month per site
Plus 0.9% of routed purchasing, including managed procurement.
See how fees work ↓Projected · 140-person assisted living scenario
$73,841 / yearNet annual value after software and procurement fees; before the separately quoted design partnership. Includes the value of staff time released, not just cash savings.
23.8×Projected annual value ÷ ongoing fees. Not first-year ROI or a payback estimate.
Calculation and assumptions
Mid-market assisted living model scaled from 62 to 140 diners: $229,950 annual food spend; 70% routed purchasing; $77,077 gross annual value less $3,237 annual fees. This illustrative scenario is not an industry average or measured result.
What you pay.
Three costs to understand up front.
2. Software
$149/month per site after the design partnership. The 0.9% routed-purchasing fee is additional.
3. Transactions & managed procurement
Managed procurement is included in the 0.9% fee on purchasing routed through MagicMenu. Any separate settlement service is scoped in writing.
How procurement fees work
Managed procurement is part of the 0.9% routed-purchasing fee. The fee applies to purchases routed through MagicMenu, not your entire food budget. Your separately quoted design partnership covers implementation.
Purchasing & settlement
Transaction revenue is tied to purchasing and settlement routed through MagicMenu.
Fees, made clear
Every fee is disclosed before the first routed order. Product rankings never see those fees.
Your value, checked
You should be better off after fees, measured against your kitchen’s own baseline.
Start with a design partnership. Three to six months in your kitchen, scoped and quoted in writing. Then $149/month per site.
Beyond the FairShare settlement test, transaction revenue is modeled, not earned.
Where the hours come from.
Time back, cycle by cycle.
Compare seven residential-kitchen workflows, today and with MagicMenu. Most time comes from daily meals and diet-order changes. The model projects 16 to 40 hours a week for a 140-resident kitchen; plan conservatively around 11.9 to 25.4 staff hours a week. Select a workflow for the steps.
Daily meal cycle6–14 h
The biggest weekly pool. The tray cards, station cards and production sheets are generated from one source of truth instead of printed, sorted and corrected by hand.
| Step | Today | Time | With MagicMenu | Time | Time saved |
|---|---|---|---|---|---|
| Order taking | Selections collected tableside against a printed diet list. | 25–45 min | Selections captured against each resident's live profile; invalid choices caught at the table. | 15–25 min | 10–20 min |
| Tray cards and production sheets | Printed from the dietary system, hand-sorted by room, corrected in pen. | 15–30 min | Generated from the locked plan, sorted by station and route, versioned. | 2–5 min | 13–25 min |
| Prep and mise en place | Recipes scaled by hand to the batch size for each variant. | 60–120 min | Station sheets show the base and its controlled variants, already scaled. | 45–90 min | 15–30 min |
| Cooking | Allergen isolation between variants juggled on the line. | 60–120 min | The branch model drives the sequence; clean-down steps inserted where shared equipment would carry contact forward. | 50–100 min | 10–20 min |
| Plating | Portion, texture and allergen matched by reading the card and the sheet. | 30–60 min | The card states the exact plated spec, texture level included. | 25–50 min | 5–10 min |
| Verification | Plates spot-checked against the diet roster before they leave. | 15–30 min | Pre-verified at the artifact: the card itself is the check; the chef sees exceptions only. | 5–10 min | 10–20 min |
| Serving | Returns and re-fires when the wrong plate reaches the table. | 60–90 min | Fewer returns because the variants are deterministic and the card travels with the plate. | 55–80 min | 5–10 min |
Modeled time saved per step: lower estimate minus lower estimate; upper minus upper. “Not timed” marks outcomes without a duration estimate.
Diet order change cycle4.7–9.3 h
The clinical order stays with the provider. What changes is that the order propagates to the profile, the tray card and the production sheet without a paper slip.
| Step | Today | Time | With MagicMenu | Time | Time saved |
|---|---|---|---|---|---|
| Order issued | The provider updates the order in the clinical record. | 5–10 min | Unchanged — clinical authority stays with the provider. | 5–10 min | 0 min |
| Clinical review | Nursing verifies the order. | 5–10 min | Unchanged, and no longer a blocker for what follows. | 5–10 min | 0 min |
| Communicate to dining | A printed slip, a verbal note, a paper hand-off. | 5–15 min | The order flows from the clinical record to the resident's profile. | 0–2 min | 5–13 min |
| Dietitian verification | The change interpreted into a therapeutic-diet spec by hand. | 10–15 min | An impact report: which upcoming items, variants and services change. | 3–7 min | 7–8 min |
| Profile and tray card | Card updated by hand; the old one removed from the sorter. | 10–20 min | A new card issued at the next snapshot. | 0–2 min | 10–18 min |
| Production sheets | Variant counts incremented by hand for upcoming services. | 5–15 min | Recomputed automatically at the next snapshot. | 0–2 min | 5–13 min |
| Staff briefing | Verbal, especially critical for high-severity changes. | 3–8 min | The briefing stays; the delta is highlighted for it. | 2–4 min | 1–4 min |
| First service under the new order | The gap between order and tray card is where errors live. | error exposure | The first service already carries the change. | error eliminated | Not timed |
Modeled time saved per step: lower estimate minus lower estimate; upper minus upper. “Not timed” marks outcomes without a duration estimate.
Substitution cycle0.75–4.8 h
The truck brings something different. Today that is a hallway decision and a phone call; here it is a twenty-second check with a logged approval.
| Step | Today | Time | With MagicMenu | Time | Time saved |
|---|---|---|---|---|---|
| Shortage identified | Noticed at delivery or, worse, at prep. | 5–10 min | Flagged at receipt and linked to every affected dish. | 1–3 min | 4–7 min |
| Impact assessment | The chef traces which items, services and residents are affected from memory. | 10–20 min | Computed from the graph: ingredient node → every downstream recipe and variant. | 2–5 min | 8–15 min |
| Candidate substitution | A hallway decision from memory of the catalog. | 10–20 min | Pre-approved alternatives ranked by availability and fit. | 3–7 min | 7–13 min |
| Dietary validation | Checked by hand against each affected resident's restrictions. | 10–20 min | Re-evaluated against the full roster with the same deterministic allergen logic. | 1–2 min | 9–18 min |
| Dietitian approval | A phone call or an email, often the same day at best. | 10–20 min | A routed approval with the validated comparison attached; the decision is recorded. | 5–10 min | 5–10 min |
| Update sheets and cards | Handwritten corrections on printed sheets and cards. | 10–25 min | A new locked snapshot; every artifact regenerated. | 3–8 min | 7–17 min |
| Communicate to kitchen and floor | A verbal huddle; information decays across shifts. | 5–10 min | The huddle stays; the artifact carries the change and the delta is highlighted. | 3–5 min | 2–5 min |
| Log for compliance | A paper log after service, if at all. | 5–15 min | Logged automatically with timestamp, reason and approving authority. | 0–2 min | 5–13 min |
Modeled time saved per step: lower estimate minus lower estimate; upper minus upper. “Not timed” marks outcomes without a duration estimate.
Weekly order-build cycle1–4 h
The largest recurring per-week wedge: the shopping list is generated from the plan, the counts and what is on hand.
| Step | Today | Time | With MagicMenu | Time | Time saved |
|---|---|---|---|---|---|
| Inventory walk | Cooler, freezer and dry storage walked with a clipboard. | 20–40 min | Exception-oriented: par-level alerts and usage history; only exceptions need a look. | 10–20 min | 10–20 min |
| Reconcile menu, on-hand and usage | Mental math across what the menu needs and what is on the shelf. | 30–60 min | Draft need computed across every variant for the window. | 10–20 min | 20–40 min |
| Real-world adjustments | “We always need extra of this.” | 15–30 min | Overrides captured once as reusable per-item rules. | 5–15 min | 10–15 min |
| Build the order | Need translated into cases across the distributor's item numbers by hand. | 20–45 min | The order generated vendor-ready from the approved menu, the variant set, the census and inventory. | 5–10 min | 15–35 min |
| Send and handle the reply | Sent through the distributor's system or by email; shortages handled later. | 10–25 min | One click for supported vendors; likely shortages surfaced before the send. | 5–10 min | 5–15 min |
Modeled time saved per step: lower estimate minus lower estimate; upper minus upper. “Not timed” marks outcomes without a duration estimate.
New resident admission2.1–3.8 h
The first meal after admission is the riskiest meal a kitchen serves. Here it runs through the same plan as every other.
| Step | Today | Time | With MagicMenu | Time | Time saved |
|---|---|---|---|---|---|
| Intake received | Care plan and diet order arrive from admissions on paper. | 15–30 min | The resident record flows from the clinical system; dining verifies. | 2–5 min | 13–25 min |
| Preferences captured | Interviewed and written on a card. | 15–30 min | The same interview, captured into a structured profile. | 10–20 min | 5–10 min |
| Allergen profile | Written on the diet card, communicated verbally. | 10–15 min | Coded to the nine-allergen schema with severity. | 3–7 min | 7–8 min |
| Meal plan configured | The chef and dietitian work out which items fit, by hand. | 20–40 min | Every upcoming item evaluated against the new profile; the dietitian reviews. | 5–10 min | 15–30 min |
| Tray card generated | Printed and inserted into the sorter. | 10–15 min | Generated at the next snapshot. | 1–3 min | 9–12 min |
| Production sheets | Variant counts incremented by hand. | 10–20 min | Recomputed for every upcoming service. | 0–2 min | 10–18 min |
| Census updated | Flows to the next order only if someone remembers. | 5–10 min | Propagates to the next draft order automatically. | 0 min | 5–10 min |
| First meal | Highest-risk meal of the stay. | error exposure | As reliable as any other because it came through the same plan. | error eliminated | Not timed |
Modeled time saved per step: lower estimate minus lower estimate; upper minus upper. “Not timed” marks outcomes without a duration estimate.
Receiving and inventory feedback1.25–2.7 h
Modest on its own; it compounds, because what arrived and what was used feed next week's order.
| Step | Today | Time | With MagicMenu | Time | Time saved |
|---|---|---|---|---|---|
| Delivery receipt | Checked against the purchase order line by line. | 15–25 min | The unload is unchanged; the checklist is pre-populated and only exceptions are flagged. | 15–20 min | 0–5 min |
| Invoice scan and match | Filed and reconciled against the order by hand. | 10–15 min | Captured and auto-matched; only mismatches need a person. | 3–8 min | 7 min |
| Inventory update | Received quantities keyed in item by item. | 5–10 min | Inventory increments from the matched invoice; unit conversions handled. | 1–3 min | 4–7 min |
| Post-service depletion | Reduced by theoretical usage; leftovers rarely logged. | 10–20 min | Depletion computed from what was actually produced, variant by variant. | 2–5 min | 8–15 min |
| Drift detection | Spotted ad hoc, usually after a shortage. | 15–30 min | Repeated mismatch patterns surfaced automatically for review. | 5–10 min | 10–20 min |
Modeled time saved per step: lower estimate minus lower estimate; upper minus upper. “Not timed” marks outcomes without a duration estimate.
Selected total across these workflows: 16.5 h / week.
Projected savings model.
Annual value for a reference kitchen, using fixed midpoint labor assumptions. Choose an operation to see its benchmark.
Assisted living — premium · 88 daily diners · 24 special diets a service · $208,780 food spend a year
Per site, before the design-partnership fee and $149 per month software cost.
From gross value to what you keep
- Labor+ $46K
- Food waste+ $14K
- Compliance+ $12K
- Verified procurement+ $4K
- MagicMenu's share−$1K
- You keep= $74K
The arithmetic
| Menu-planning + procurement hours a week today | 21 |
|---|---|
| Compliance-documentation hours a week today | 14 |
| Special diets per service | 24 |
| Hours recovered a week | 20 |
| Blended loaded rate | $43.87/h |
| Labor savings a year | $45,646 |
| Food-waste pool a year (spend × waste %) | $91,863 |
| Food-waste savings a year (15% of the pool) | $13,779 |
| Compliance hours recovered a week | 4.6 |
| Compliance savings a year | $11,996 |
| Food spend routed (70%) | $146,146 |
| Verified procurement savings (3.0% of routed) | $4,384 |
| MagicMenu's share (30%) | −$1,315 |
| You keep of the procurement savings | $3,069 |
Fixed cycles (menu planning, procurement) and variable cycles (diners × diets) are indexed to the reference profile (140 residents, 24 diets, 21 planning hours a week); loaded rates are hours-weighted across dining roles at BLS wages with a 30 percent benefit burden; the routed share and the verified-savings rate are the model's base case. Very large kitchens run past the model's reference; read them as direction, not a number.
Quick estimate · including seasonal camps
12–25 hkitchen staff hours per operating weekabout 0.3–0.6 of a full-time position
$52K–$85Ka year, before feeslabor recovered + waste avoided + compliance hours
Mostly from the Daily meal cycle and the Diet order change cycle.
Anchored to 11.9 to 25.4 staff hours a week at a 140-resident assisted-living community; residents scale the hours sublinearly and stop at 500. Modeled, not measured.
Illustrative camp scenario, not measured results. Extrapolates the care time model using the camp archetype’s planning workload, wages and food budget; assumes a 20-week reference season. Set your operating weeks. No RD savings assumed. Your baseline replaces these assumptions.
Managing procurement.
Match products
Review ingredients, pack sizes and approved substitutions against the menu.
Plan the order
Connect meal counts and quantities to the products your kitchen needs.
Check what arrives
Review substitutions and invoice differences; flag issues for follow-up and credit.
Connect the menu to what you buy.
Nine levers that save and three that cost, on the spend you route — contract utilization, invoice recovery, pack-size and substitution checks. Net, 3.75 percent of routed spend at full maturity, before MagicMenu's share.
One real catch: a distributor substitution on Annunciation Heights's invoice, 12 percent over the item it replaced — caught, blocked before the pull list, and flagged for credit.
The twelve levers
- Contract utilization +0.60%
Activating and using the contracts you already have.
- Invoice and contract-compliance recovery +0.50%
Cash recovery and prevented overcharge.
- Distributor pricing and service terms +0.70%
Volume, prompt-pay and delivery-size mechanics distributors already publish.
- Cooperative or GPO switch (tier-adjusted) +0.30%
Only where eligibility and lost tiers allow it.
- Multi-supplier optimization +0.40%
Broadline, regional and specialty competition on the same specification.
- Retail and same-day fill-in +0.10%
Small-quantity and minimum-order exceptions.
- Order timing, delivery size and payment behavior +0.40%
Fewer, fuller drops; paying the way the terms reward.
- Product standardization and approved equivalents +0.50%
Requires culinary and safety approval; ranked compensation-blind.
- Aggregated network contracts +0.70%
Scales with routed volume; a later rung.
- Friction: freight, minimums, split cases −0.15%
Added back to the optimized cost.
- Friction: lost rebates, tiers and credit value −0.20%
Added back to the optimized cost.
- Friction: service, switching and working capital −0.10%
Added back to the optimized cost.
For illustration: every $100,000 of routed purchasing would yield about $3,750 in net purchasing savings before MagicMenu's share — a modeled percentage at full maturity, not a guaranteed return.
How we know.
- Verified A public source you can open — identified beside the figure.
- Company-reported Counted by us in a real kitchen; with source details attached to the figure.
- Modeled from a 140-resident reference; being validated in live kitchens; replaced by your own baseline in the first phase.
Source details remain attached to the figures. Projections are not measured results.
What "your own baseline" means
Homepage role splits and seasonal camps
The six homepage bars divide existing modeled pools: by default, kitchen labor is 40% chef, 20% front of house and 40% back of house. The separate documentation pool is split equally between RD and admin. These are allocation assumptions at blended rates, not measured savings for each role. Camps assume no RD recovery.
The camp scenario extrapolates the care time model using the existing 220-guest seasonal camp archetype: 13 planning/purchasing hours versus 21 in the care reference, $29–$31 loaded labor rates, and a $180,000 food-budget scenario. A 20-week reference season is an explicit planning assumption; changing operating weeks scales seasonal value. Population scales hours sublinearly (exponent 0.8). It is not camp outcome evidence.
In the first phase we digitize your invoices, rebuild your recipes from your menus, count a week of service and weigh what comes back. Food cost per served meal, waste, labor minutes, substitutions and rework — each labeled by how it was measured. The projections on this page are then replaced by your numbers, and the after is measured the same way.
Get your own baseline.
Thirty minutes with Dusty. Bring your dietitian to the first call if diets are clinical.