There's a second image of the menu partition: how each dish takes over the back.

The customer menus, which are arranged in the form of cold dishes, staples or signboards, deal only with “how the customer finds the dish”; the rear cook needs another production chart, which answers the preparation, cold storage, cooking, assembly and meal nodes of each course, and which people, equipment and space are used at the same time. The two diagrams are linked to the unique dish code, but to verify separately the search experience and the production load, the back-up load cannot be assumed directly by using the front-office classification.

WebstaurantStore's menu design articles discuss the organization, layout and customer reading of vegetables. [1] It supports the front desk information architecture without providing the local order portfolio, process sequencing or equipment capacity. The changes in purchases after the menu has been reordered can also be observed only with actual orders on the line, and historical order simulations are used only to check if the hypothetical combination is carried.

Creates a dish master data instead of changing only the theme of the column

Each dish saves the only dish code, the name of the customer, the front section, the specifications, the options, the conditions for the suspension, the reserve card, the allergy or food information interface, the name of the kitchen bill, the version of the process and the date of entry into force. The same version is cited for paper menus, online portals, cash collections, feed cards and training materials.

The front-office zoning can be changed and the number of vegetables and the process relationship should not be changed quietly. The channel may display multiple entrances if the “branch” also belongs to the “principal food”, but the late cook still receives the same dish code and avoids replicating or two sets of recipes.

Full demonstration map of a dish

The following is a demonstration, not the real process or capacity of the door: the dish D101 is in the “hot food” section of the customer's menu, the order is entered to obtain pre-treatment material from Refrigerated Node R1, the ration and assembly preparation at the P1 level, then the approval process at the Cooking Node C1, and the final check at the E1 point. The purchase orders, feed cards and training are shown in D101 and the same version.

Food code	Front desk partition	Preparation and refrigeration	Cooking/processing	Organise dinner	Load to verify
D101	Hot food.	R1, P1	C1	E1	Time taken, hair taken, toggle
D202	Snack.	P1	C1	E1	Sharing P1, C1, E1 with D101
D303	Vapour Point	R2	C3	E2	Vapour Batch and Wait

When on-site implementation takes place, nodal code binds the actual asset number, slot and program version; only equipment that is actually involved in the process and has a mission record enters the map and does not use product names or web-page parameters as evidence of capacity.

Historical baseline describes the facts first

Selecting a comparable time period by which a combination of historical orders, arrival points, cancellations, stoppages, kitchen bills, completion points, back-to-work and equipment or space occupation records are measured. The baseline is to retain dates, shifts, people, menu versions and anomalies, and not just a daily average.

Historical data indicate what happened under the old menu and do not prove how the new layout will be bought by customers. If the historical records lack a process or asset number, they can only be marked as gaps and cannot be converted to a flat power of the equipment.

Check the carrying conditions with three scenarios

The baseline situation is re-ordered with historical orders and actual arrival sequences, and the validation model is now largely known for queue and nodal occupancy. Conservative scenario assumes that the new division has not changed the purchase combination, but only the presentation; it is used to check the system map and does not claim a change in sales. The peak scenario is clearly assumed by the operator, e.g. D101 increases in the same window at the same time as D202, while D303 remains historical; the source, range and rationale are stated at the head.

Simulations are carried out on a case-by-case basis through the preparation of materials, cooking and meal nodes, recording the queuing of shared nodes, the co-distribution of equipment, batch waiting, cross-section of personnel and stoppages. It can only be exposed to “possibly overload under these assumptions” and does not prove that the new layout necessarily creates the order combination.

A fully recalculated hypothetical window is given: C1 has only one equivalent space, D101 has four minutes, D202 has three minutes, and the field has been cleared for each of its own time periods; 12 noon to D101, 12:01 to D202, 12:02 to D101, starting at 12:04, 12:07 on a first-come first-served basis and waiting for 0, 3 and 5 minutes. If the menu shows the changed situation, the 12:01 is changed to D101, while the remaining arrivals, personnel and nodes remain unchanged, the three will begin at 12:00, 12:04, 12:08, waiting for 0, 3, 6 minutes, and C1 will be occupied from 12:00 to 12:12. The change only indicates that this assumption combines the third order of one minute; it is not a sales projection, nor does it prove that a real door store needs additional opportunities.

If a more radical combination is to be made, then the situation will be increased rather than the baseline will be changed. Each time only the specified variable is changed, the previous result is retained to avoid being attributable to simultaneous changes in menus, personnel, equipment and processes.

Customer missions and production missions are tested separately

The customer side tested for a typical mission to find the dish, to understand the specifications and options, to find a stop-sale or a restriction, and to be able to cook. The back-cooking side verifys whether the simulators, empty containers or approved cold-forming missions match the dish code, feed cards and processes and do not treat staff memory as a systematic map.

The actual purchase orders after they are online can be used to observe changes in the search and purchase. The date, activity, price, inventory and channel need to be taken into account in comparison, while retaining conditions that are not controlled and that do not refer directly to sales as causes and consequences.

The load must come from the record.

The equipment web-based power indicates only rated electrical fields and cannot be converted to single time, bulk capacity or a combined maximum. The load sheet requires on-site task records, current procedures, batches, loading and unloading, cleaning switching, and a sample of personnel and failure. Node unrecorded “to be verified” does not add a theoretical yield because of the association of a model.

In particular, shared workspaces are considered portfolio occupancy: D101 and D202 can be completed, respectively, and do not mean that P1 can still be carried at the same time. Statistical queue distribution and extremes, safety, food and equipment restrictions take precedence over shorter waiting periods.

Count nodes from the order combination

The matrix is not just about a "hundred-thirty-thousand-one-hundred " , but lists the number of singles, the number of nodes per single or batch in each time window, the number of parallels, and the number of actions required to switch. The same total unit may create a completely different bottleneck: the D101 concentration will press C1; by this demonstration, the D101 and D202 rise together will occupy P1, C1 and E1, and which first creates the bottleneck will be judged by the length of the field at each node and by the rules of the parallel.

Mission duration uses on-site observation distribution rather than taking the fastest. Loading, waiting, procedures, offloading, cleaning, patching and abnormal stops are recorded separately; multi-person collaboration is also separated from natural life. If a process is run in batches, the model needs to indicate batch rules and unauthorised waiting periods and cannot simply add up each time.

The presentation results can be written as “Assuming that the P1 to be processed under the peak group continues to grow”, but not as “menu reordering leads to P1 overload”, as the purchase combination remains a hypothesis. The verification path is a small-scale go-live, observation of the real combination, and then an update of the situation and the personnel equipment arrangements.

The failure is more indicative of the boundary than the average.

The average wait remains unchanged, and there may still be some type of long-term retention of orders, misalignment of bills, failure of switchover of shared equipment or staff circumventing food control to catch up. Simulation and commissioning are maintained as the largest waiting, back-to-work, stop sales, equipment alerts, cross-district and uncompleted orders and return to specific dish code and node.

An anomaly cannot be automatically attributed to the menu, but cannot be averaged by a large number of successful orders. The team differentiates data, maps, processes, personnel, equipment and external events before deciding on the re-division of menus, limiting options, changing materials or adjusting production lines.

Food variants require independent mapping.

The additions, drops, specifications, spicyness or package replacements may change the process and nodes and cannot always inherit standard dishes. Each option that changes the equipment, batch, food information or assembly creates a variable code or clear rules to test whether the instrument conveys the difference to the back cook.

If the front desk allows an option and the cook does not approve the process, the option is closed to prevent the staff from creating a path on site. At the time of the closure, the customer entrances, the cash register and the materials were synchronized so as not to be known only to the cook.

Upline observation and cessation conditions

The observation window, the person responsible and the conditions for the stop, such as the error of the paper map, the continued growth of the shared node queue, the triggering of equipment or food restrictions, and the need for staff to bypass standard actions, are pre-set before they go online. The suspension of the affected vegetables or their return to the verified version on condition that the conditions are met does not have to wait until the end of the day.

Observation records are maintained by order and node, indicating unusual activity, shortage of goods, personnel and equipment. A few days of data can reveal a clue to the problem and cannot automatically extrapolate the whole year; seasons, promotions or channels change and need to be reassessed.

Publish with a Change Matrix

The matrix lists the old subdivisions, the new subdivisions, the dish code, the name of the POS, the kitchen bill, the preparation card, the process version, the training materials, the channel and those responsible. Arrangements are not effective until all affected persons have passed; the uncompleted channels are temporarily shut down or the old editions are maintained, and the front desk is not allowed to sell and then cook.

The reporting line sets back conditions, but before returning, it must be confirmed that old vegetables, prices, formulations, stocks and food conditions are still valid and that projects that have been discontinued cannot be voluntarily reinstated. The failure record goes to the next version of the change without an oral statement to cover the difference.

The final delivery should include both customer partitioning, a menu-by-show, historical baselines, situational results of explicit assumptions and a full-source version matrix. This way the menu design will not be clear only on the page, but will leave the congestion behind.

References

• WebstaurantStore: menu design guide