Selling eight coffees does not tell you how much milk is left.

Separate three questions before the demo

Ask how many menu items were sold, how much ingredient the configured recipes would use, and how much ingredient is physically present. These are three different questions. A report showing eight coffees answers the first; it does not establish the other two.

Decide whether you need a dish-sales report, finished-portion availability, ingredient tracking or all three. A sold-out flag for a menu item is not evidence that raw ingredients were deducted. Likewise, a lower ingredient balance does not prove that the correct recipe was used for each sale.

Agree on units and portion links

Write one base unit for the ingredient and convert all quantities into it for the rehearsal. In the sample below, milk is measured in millilitres: one litre is 1,000 ml. A purchase pack, an open container and a recipe portion must not be added as if their numbers describe the same unit.

Use the exact pack sizes and portion definitions from your own records. Do not assume that one bag, carton, ladle or cup has a universal size. Ask the demonstrator to show where the conversion and the menu-to-recipe link are configured, and which staff role may change them.

Oracle’s restaurant inventory documentation describes menu-to-recipe links and portion ratios as inputs to stock depletion. Its recipe documentation also separates batch yield from portion size. These are useful concepts for a supplier demonstration, not evidence that DeccanPro has the same implementation.

Calculate a small ingredient example by hand

Use fictional quantities for a quiet training exercise, not as a recipe or purchasing recommendation. Start with 2,000 ml of milk and record a further 1,000 ml received. Define a regular coffee as using 100 ml and a large coffee as using 150 ml for this example only. Six regular and two large coffees give eight drinks, with theoretical milk usage of 900 ml: 6 × 100 + 2 × 150.

Record a separate 100 ml discard. With no other movements, the expected closing quantity is 2,000 ml: opening 2,000 + received 1,000 − theoretical usage 900 − recorded discard 100. If the physical count is 1,900 ml, the difference is minus 100 ml, calculated as counted minus expected. The difference is an observation to investigate, not a reason to alter the sales record until it agrees.

Catch a wrong recipe even when dish counts agree

Repeat the same eight-drink example with one deliberate configuration error: map the large coffee to the regular coffee’s 100 ml quantity. The dish count still shows eight. The calculated milk usage becomes 800 ml instead of 900 ml, and expected closing stock becomes 2,100 ml instead of 2,000 ml after the same receipt and discard.

This is why an agreeing sales count cannot validate ingredient deduction. Check regular and large portions individually before trying the mixed order. Record the configured quantities and the resulting stock movements. Correct the test configuration with the authorised demonstrator and repeat the same exercise; preserve the first result so the reason for the difference is clear.

Test cancellation before and after preparation separately

In a separate practice case, cancel an order before any ingredients are used. Ask which event created the theoretical deduction and what the demonstrated cancellation does to it. Then use a fresh example where a drink has already been prepared before the order is cancelled. The billing change does not physically put its ingredients back into storage.

Have the supplier explain how the installed workflow represents these cases. Depending on the system and restaurant process, an authorised separate stock or waste record may be needed. Do not assume that cancellation always restores stock correctly, or that retaining the deduction is correct for an order never prepared. Keep the kitchen outcome, bill status and ingredient movement together in the test notes.

Investigate the difference without counting it twice

Use the same cutoff for the opening balance, receipts, sales, discards and physical count. Check whether a delivery was entered after the count, whether an open container was measured in the agreed unit, and whether another approved movement has already been recorded. A difference can come from timing, units, configuration or missing records; it does not by itself establish theft or identify a staff member.

Keep transfers distinct from consumption. Moving milk from one recorded storage location to another should be investigated as a transfer between those locations, not automatically as another drink sold. Ask whether your intended software setup handles that distinction and demonstrate it if transfers matter to your restaurant.

Do not record the same discard once in a recipe allowance and again as a separate movement without understanding the model. Preserve the original observations and have an authorised person document any adjustment. A zero difference produced by an unexplained adjustment is not evidence that the original workflow was correct.

Bring an inventory requirement sheet to the Chennai demo

Bring one ingredient, two portion sizes, their approved quantities, a purchase pack size and an anonymised stock sheet. Ask the supplier to demonstrate the full path from menu selection to the resulting ingredient record. Record unsupported steps explicitly, including any separate module, manual entry or external system needed.

DeccanPro’s restaurant software offer is Rs.6,000 per year with unlimited use and no setup or hidden software charges. That price does not establish an untested recipe or stock workflow. Confirm your inventory requirements alongside billing, table orders and kitchen tickets, and keep equipment and third-party quotations separate.