Restaurant customization is becoming a production-policy question, not just a hospitality debate. In Eater’s October 22 reporting on substitution culture, Chicago’s Taan describes limiting modifications after requests slowed service in a small kitchen. Phoenix’s Lom Wong similarly explains why some changes compromise the identity of its dishes. Neither example justifies treating every request as unreasonable; both expose the cost of promising something the operation cannot reliably deliver.
A separate development points toward the same management responsibility. Fast Casual reports that Sweetgreen appointed John Karangis executive chef, with responsibilities including translating culinary ideas into consistent restaurant execution. This is not evidence that Sweetgreen is changing its substitution policy. It illustrates the broader challenge: someone must own the translation between a food promise and repeatable production.
Our earlier analysis of menu growth and kitchen zoning addressed permanent additions. This week’s narrower issue is the unofficial menu created by modifications: which requests can the existing station execute, which require approval, and which should never be promised?
The unofficial menu needs an owner
A printed menu may contain twelve dishes while the ordering system permits hundreds of combinations. Every substitution can introduce a different ingredient container, preparation sequence, utensil or verification step. The guest sees an omission or addition; the kitchen may see an entirely different workflow. Complexity becomes expensive when that workflow is undocumented and a server must renegotiate it during every rush.
USA-RS’s view: accommodation should be designed, not improvised. Establish an approved modification matrix jointly with culinary, operations and the person responsible for food safety. Front-of-house employees need an explanation they can deliver respectfully. Kitchen employees need a reproducible sequence. Neither should be expected to invent the policy while tickets accumulate.
USA-RS take: A bigger refrigerator cannot fix an undefined promise. First define the accommodations your team can execute consistently; then purchase capacity for that documented workload.
Separate preference, feasibility and allergy communication
These are different conversations. Removing a garnish before assembly may be straightforward. Removing shrimp paste already incorporated into a sauce is not the same operation. An allergy-related request requires accurate ingredient information and an assessment of cross-contact controls—not merely a modifier button labeled “no shrimp.” Declining a particular preparation is different from dismissing a guest’s concern.
The FDA Food Code provides a model framework for retail food safety; actual requirements depend on the locally adopted code. Train employees on the applicable procedures, escalation and communication. Do not describe shared equipment as allergen-free or guarantee safety because a utensil was swapped. A manager must evaluate the complete preparation pathway.
Eater’s reporting offers a useful hospitality distinction: Taan maintains a substantial alternative rather than treating accommodation as an afterthought. The transferable lesson is to develop an intentional option where feasible, while explaining honestly which existing dishes cannot be reconstructed. Recipe identity and guest care do not have to be opposing goals.
Build a modification matrix before changing the station
Use a small number of clear classifications. The following is an operational planning template, not an allergy protocol or a description of the restaurants in the reporting. Each restaurant should approve its own ingredients, procedures and guest-facing language.
| Request classification | Production decision | Required communication |
|---|---|---|
| Approved assembly omission | Omit an ingredient before it touches the dish; retain the established sequence. | Identify the exact omission on the ticket and confirm it at handoff. |
| Approved alternative preparation | Use the documented recipe, designated supplies and approved preparation pathway. | Explain any material difference in preparation time, ingredients or price before acceptance. |
| Ingredient already incorporated | Do not pretend that removing visible pieces changes the underlying recipe. | Explain the limitation and offer an appropriate existing alternative if available. |
| Unresolved allergy-related request | Escalate before preparation; assess ingredients and actual cross-contact controls. | State verified capabilities and limitations without promising a risk-free preparation. |
Review the matrix whenever suppliers, recipes or equipment change. A substitution approved last month can become inappropriate after a sauce reformulation. Employees should see one current reference, not conflicting handwritten notes at the register, prep station and expo counter.
Measure refrigerated positions, not imagined flexibility
Approved alternatives consume physical inventory positions. A second dressing needs a container, a lid, replenishment stock and a consistent location. Several low-volume alternatives can crowd out the ingredients that sell all afternoon. Review actual pan configurations in sandwich and salad prep tables against the approved recipe matrix, including backup storage and cleaning access.
For a compact example, Atosa offers the MSF8301GR one-section prep refrigerator, listed with eight included one-sixth pans and 7.2 cubic feet. Eight openings are not eight unlimited menu platforms. Assign every position, determine replenishment responsibilities, and verify the exact electrical configuration, installation clearances and operating instructions before ordering.
This is a capacity example, not a recommendation that a particular refrigerator solves allergy accommodation. Segregation, ingredient verification, cleaning and employee practices remain essential. If backup containers repeatedly migrate onto the assembly surface, the problem may be the number of approved alternatives rather than insufficient cabinet width.
Price the reset, not only the extra ingredient
Modification labor often hides after handoff. Count container replacement, utensil handling, surface cleaning, replenishment and ticket verification. An ingredient with modest purchase cost can still create expensive interruptions. Conversely, an approved omission that removes an assembly step may add little burden. Do not impose one operational assumption on every modifier.
Evaluate commercial preparation work tables by usable workspace and separation from routine traffic, not nominal width alone. A surface already occupied by packaging cannot simultaneously become an exception-preparation station. Any designated preparation area also needs a workable route for soiled tools and clean replacements; otherwise the reset simply moves congestion downstream.
Record the exception’s labor and waste separately from the standard dish. The point is not to punish customization. It is to stop a popular accommodation from quietly borrowing minutes, refrigeration and attention from every other guest.
Test an exception queue during real production
Run a controlled rehearsal with the employees who normally work the station. Include standard tickets, approved modifications, an unavailable ingredient and a request requiring manager review. Observe whether employees identify the right pathway without abandoning ordinary orders. Do not introduce unapproved allergen challenges as a casual experiment.
- Track acceptance: Does the order taker identify the request before committing to a preparation?
- Track execution: Can the cook locate approved ingredients and tools without interrupting another station?
- Track verification: Does the handoff identify the modified dish without confusing it with standard orders?
- Track recovery: Are cleaning, replenishment and communication complete before the next request arrives?
Compare standard-order delays, modification remakes and manager interruptions across comparable service periods. Set your own acceptance thresholds before testing. An additional workstation is justified only when a defined, repeatable workflow needs capacity; it is not the first response to conflicting instructions.
Make the ordering channels tell the same truth
A policy fails if the server explains one restriction while an ordering app permits the opposite combination. Reconcile menu descriptions, modifier choices, kitchen-ticket language and packaging identification. When a capability changes, update every channel before the next service. Otherwise the cook becomes the last-minute messenger for a promise already sold.
Give employees a short escalation script and a realistic alternative. “That ingredient is incorporated into our sauce; let me check which existing preparation fits your needs” is more useful than an unexplained refusal. Record requests the matrix cannot address and review them outside service. Repeated demand may justify recipe development, but one difficult ticket should not redesign the restaurant.
Give one person authority to release the revised promise
A policy document is not ready simply because culinary approved the recipe. Assign a release owner who checks the ingredient reference, ordering channels, preparation sequence and shift training together. The owner should confirm that the opening crew and closing crew understand the same limits. Otherwise a technically correct recipe still becomes an inconsistent guest experience.
Build a short change record for each new accommodation: what changed, why it changed, which ingredients are affected, where supplies live, and who approved the workflow. Include the effective service date. Remove obsolete instructions at the same time rather than leaving employees to guess which version applies. For multi-unit businesses, distinguish a pilot restaurant’s tested capability from a permission that applies across the system.
During the first week, review exceptions with both a server and a cook present. Ask which requests were misunderstood, which alternatives guests accepted and which resets took longer than expected. Those observations may support a better menu description or an additional approved recipe. They may also show that the right answer is a clear limitation, not another piece of equipment. Only expand the promise after the complete operating pathway is ready.
Check replenishment before calling the pilot complete
A station can pass a quiet rehearsal and fail when the first ingredient container empties. Include a replenishment cycle in the test: where the backup is stored, who retrieves it, how the current ingredient identity is confirmed, and what happens to the empty container. Verify that a replacement does not silently introduce a different recipe or supplier formulation. This is also where a small cabinet can reveal its true limit. If every approved alternative requires a remote restocking trip, revise the preparation quantities or the approved assortment before assuming that more menu flexibility is sustainable.
Why we’re watching this
The news is not that personalization has suddenly become universal or that every restaurant should restrict it. Eater documents specific operators setting boundaries; Sweetgreen’s leadership announcement emphasizes consistent execution at scale. Together, these distinct developments make culinary ownership worth watching: who decides what can be promised, and who verifies the promise on the line?
We would watch repeated modification requests, remakes and reset interruptions before watching equipment size. A well-supported alternative may create better hospitality than endless improvisation. The strongest operation makes its capabilities clear and develops new ones deliberately.
If you’re planning around this
Bring your approved modification matrix, ingredient positions and reset observations to the equipment discussion. Contact USA Restaurant Suppliers to evaluate the physical workflow, or browse the full equipment catalog. Buy for a process your team can explain and execute.