Capture Inventory Pars: 2 Tools 2026 [Workflow Recipe]

The restaurants category decision is which system actually owns on-hand quantity when a cook needs to reorder, not which vendor has the prettier tablet. Tracking inventory par levels for reordering is the process of comparing counted or depleted stock to a written minimum and creating a purchase action only when the gap is real. Toast is a restaurant POS and operations suite that can hold items, modifiers, and, on inventory-enabled sites, stock. OpenTable is a guest-reservation network. Neither product is a warehouse management system, and neither is a substitute for a named par list.

Toast vs OpenTable for par-level reordering is a comparison of a POS that can see item depletion against a reservation book that can see covers, judged on whether those signals can become a purchase order without a clipboard. US Tech Automations binds only when par math must cross POS depletion, vendor files, and a human hold. no restaurants vendor paid for inclusion.

TL;DR: Choose Toast when the item, the check, and the depletion event should live in one POS catalog and you will actually run inventory counts there. Choose OpenTable when the gap is guest-count planning, not SKU stock. Orchestrate across both only after unique item IDs, vendor packs, and a reviewer exist.

Labor is the reason this workflow dies on paper. Independent labor cost: 32-36% of revenue according to Toast (2024), 32-36% of revenue for independent restaurants, with the range varying by service model. A kitchen that already spends that band on people cannot also staff a perfect nightly count unless the count is short and the reorder is written. The manual par comparison is the control case: same par list, no pipe.

Key Takeaways

• Toast is the closer restaurants system of record for SKU depletion; OpenTable is the closer system for covers, not cases.

• List prices (checked 2026-09-04): Toast restaurant POS and inventory add-ons, contact vendor; OpenTable guest-center plans, contact vendor.

• Independent restaurant labor in the cited Toast band is 32-36% of revenue, which is why par lists fail when they depend on unpaid closing time.

• Native Toast inventory or a standing OpenTable cover report can be enough when only one signal is required.

• Orchestrate POS, vendor, and reservation data only after unique item IDs, pack sizes, and a reviewer exist.

What a par-level reorder actually is

A par is the minimum on-hand quantity that should be on the shelf, in the walk-in, or in the prep bay before the next delivery. Reordering at par is not “buy whatever we used yesterday.” It is “buy the difference between counted or depleted stock and the written minimum, in vendor pack sizes, with a human able to stop a bad line.” The failure mode is a 86’d item that still shows as in stock, or a standing auto-order that buys a case the kitchen already prepped.

Restaurant sales sit in a large U.S. foodservice market according to National Restaurant Association (2025), $1.5 trillion forecast for 2025, which is why a missed par is not a boutique problem. Food-away-from-home spending is now a majority of the food dollar according to USDA ERS (2024), more than 50% of the food dollar in recent Food Expenditure Series years. Those two facts explain why depletion is fast and why a weekly spreadsheet lags.

Food services and drinking places employ about 12 million people according to BLS (2024), about 12 million jobs in that CES sector, which is the labor pool that would have to do the count if software does not. The Census Bureau counts more than 700,000 food services and drinking places establishments according to U.S. Census Bureau (County Business Patterns, recent year), more than 700,000 establishments. Most of those sites are not running a warehouse WMS. They are running a POS, a reservation book, and a vendor portal.

Food-away-from-home share: over 50% of food dollar according to USDA ERS (2024), more than 50% of the food dollar. That is the demand side of par math: guests eat out enough that on-hand stock turns faster than a weekly order cycle.

EPA’s food-waste range is the other side of the same coin. A par that is set from fear, not from depletion plus covers, buys waste. The food-waste tracking workflow is the sibling motion when the gap is spoilage, not stockout. The supplier ordering workflow is the sibling when the purchase order already has a vendor file.

Par-level automation is therefore three objects, not one app: an item ID that matches the recipe and the vendor pack, a quantity that is either counted or depleted from sales, and a cover or event forecast that changes the par on a Saturday. Toast can hold the first two if inventory is actually enabled. OpenTable can hold the third as reservation covers. The reorder lives in a vendor portal or email unless someone writes the handoff.

Weighted evaluation criteria

Weights assume an independent or small-group restaurant that already runs a POS and may run a reservation book. A commissary with a real warehouse system should raise “vendor pack and receiving” and lower “reservation covers.”

Item-level par is weighted highest because a cover count without an item ID cannot buy romaine. API or export access matters for the same reason: if the POS will not emit Order.guid and item quantities, the kitchen is back on a clipboard. Confirm inventory is on the quoted Toast edition. Confirm OpenTable is not being asked to store cases.

Normalized capability matrix

Scores from public product restaurants pages checked 2026-09-04: 2 = first-party restaurants description of this restaurant job; 1 = adjacent, confirm in the restaurants contract; 0 = not found for par-level reordering. The USTA row is a first-party publishing-velocity figure, not a POS benchmark. It does not mean Toast indexes inventory faster than OpenTable.

Pricing and TCO, dated

Toast does not publish a single national list price for POS plus inventory on Toast restaurant pricing that can be quoted as a universal monthly number; hardware, software edition, and inventory features are quoted. Write contact vendor. OpenTable does not publish a single national list price for all guest-center packages on OpenTable for restaurants that covers every market; per-cover and subscription mixes vary. Write contact vendor.

A two-system book is two invoices plus the person who reconciles item IDs on Friday. Those invoices are not TCO until you add count time. If independent labor already sits at 32-36% of revenue, unpaid closing counts are not free. Put the closer’s hours on the sheet before you call either option cheaper.

U.S. food supply wasted: 30-40% according to EPA (food-waste overview), 30-40% of the food supply. That range is why a padded par is not a free insurance policy: the extra case still dies in the walk-in. Price the spoilage next to the software line, not as a surprise on the food-cost report.

Vendor profiles

Toast: POS depletion, inventory when enabled

Toast is the POS to keep when checks, modifiers, and item IDs belong in one restaurant catalog and closers will actually count stock there. Evidence: Toast restaurant POS. You are paying for this job only after inventory, purchasing, or a documented item export is on the quote. A terminal that takes a card is not inventory.

Limits: payments and inventory are different SKUs. Multi-location recipes, commissary transfers, and vendor EDI are easy to over-promise. Keep Toast when the model is “the POS holds the item.” Walk away when the only pain is reservation pacing, or when next quarter’s inventory module is missing from the quote.

OpenTable: covers, not cases

OpenTable is the reservation book to keep when the demand signal is booked covers, waitlist, or a guest-center event, and SKUs already live somewhere else. Evidence: OpenTable restaurant solutions. It wins guest-count visibility. It is not an inventory ledger.

Limits: there is no honest par object here. Keep OpenTable when Saturday night covers should change prep, not when you need to buy a case of oil. Walk away when a buyer is shopping it as a stock replacement for a walk-in count.

Operators who skip this split pay twice. They buy Toast for payments, then discover inventory was never on the edition, then buy a separate inventory app, then keep OpenTable for reservations and still email the vendor from a phone photo of the walk-in. Write the restaurants system of record in one sentence: “Toast holds item quantity” or “the vendor portal holds quantity.” Every other tool is a signal. If you cannot write that sentence, pause the purchase.

A second common miss is edition math. A Toast payments go-live looks complete until the first par list needs an item ID that matches the vendor pack. An OpenTable go-live looks complete until someone asks where on-hand cases live. Put the edition you will actually run—the one that stores quantity, the one that emits covers, the export that a reviewer can read—on the 12-month sheet before you compare logos.

Implementation hours belong on that sheet too. Contact-vendor POS plus contact-vendor reservations plus a closer who counts 40 SKUs at close is a real cost even when software is “included.” Count that person as a line item. If you will not staff them, do not buy the more flexible stack.

Par lists fail in predictable places. Produce and dairy move in days; dry goods move in weeks; liquor may sit behind a lock that the closer does not open. If one par worksheet treats all three the same, the produce par is always wrong and the dry par is always padded. Split the list by shelf life before you connect a pipe. A depletion event from Order.guid is a good trigger for the fast list and a noisy trigger for the slow list.

Unit conversion is the next silent break. Recipes speak in ounces and each; distributors speak in cases and catches; the POS may speak in modifiers. If romaine is a salad modifier in Toast and a 12-count case in the vendor file, a naive quantity write will order 12 salads of lettuce or 1 leaf of case. Store the pack size next to the item ID. Round up only after a human sees the raw gap.

Location matters as soon as you have a prep bay and a walk-in. A par of 12 cases that lives “at the restaurant” will hide 4 cases already prepped and 2 cases on a speed rack. Count points should match how the kitchen actually stages food. Transfers between a commissary and a unit need the same item ID on both sides or the unit will reorder what the commissary already sent.

Receiving is the last mile people skip. A PO that never gets a receive leaves on-hand low forever, so the next night’s job double-orders. If you will not tap in a receive, do not auto-send the PO. A hold that waits for a receive photo or a vendor invoice quantity is slower and honest. The point of automation here is fewer surprise 86s, not a faster way to buy the wrong case.

Covers from OpenTable should change prep pars, not dry pars. A booked jump from 80 to 140 covers can justify another case of fish and another pan of dessert, and it should not by itself buy fryer oil. Tag which SKUs are cover-sensitive. Weekend events, patio weather, and large-party notes are demand signals; they are not on-hand quantities. If your only reservation export is a PDF, plan a nightly file drop rather than pretending the book is an API.

Vendors have cutoffs. A par gap that appears at 10 p.m. cannot ship on a truck that closed at 4 p.m. Encode the cutoff as a clock, not as hope. The workflow should either hold until the next order day or page a local runner. Silent overnight emails that the distributor will not read are not a process.

Multi-unit groups should not start with a company-wide par. Pilot one location, one vendor, and the 12 SKUs that cause 86s. Copy pars only after item IDs match. The multi-location inventory tools list is useful when you are choosing a stock product; it does not replace a written par owner at each unit.

Count cadence is a design choice, not a vibe. Fast lists (produce, dairy, fresh fish) need a count that matches the delivery calendar, often daily or every other day. Slow lists (dry, paper, chemicals) can live on a weekly cycle. If you force one cadence, the closer will fake the slow list and skip the fast list. Encode two jobs. A POS depletion feed can fill the gap between counts on the fast list; it should not replace the weekly walk of the dry storage.

86, void, and waste are three different quantity events. An 86 in Toast means do not sell; it does not always decrement on-hand. A void means the check lied. Waste means the item left the building as trash. If your par job treats all three as depletion, you will reorder ghosts. Map each event to a quantity rule and a reviewer. The food-waste tracking workflow is the place to park true waste; this page should only consume a waste quantity when you trust that feed.

Holiday and weather weeks need a par overlay, not a new system. A written Saturday par that is 1.4 times Tuesday is better than a model nobody can explain. Store the overlay as a dated exception the chef can turn off. If OpenTable covers are the overlay input, round to whole cases after the chef sees the raw cover delta. Do not let a patio-weather guess buy a second case of oil.

Distributor substitutions are why a human still belongs on the PO. If romaine is short, the vendor will offer a different pack or a different green. An auto-send that cannot accept a substitution will either fail or buy the wrong item under the same ID. Hold the line, show the substitution, and let purchasing pick. That hold is slower than a night email and faster than an 86 at 6 p.m.

Par-gap walkthrough (configurable)

An illustrative two-unit kitchen tracking 42 produce SKUs, a par of 12 cases of romaine, and 8 hours between counts can treat each Toast Order.guid as a depletion event: when checks post, a configurable workflow sums the romaine modifier quantity, subtracts it from last count, and opens a purchase task only if on-hand would fall below 12 cases before the next delivery window. Prerequisites: Toast API credentials, a uniqueness key on item ID plus location, vendor pack size in the same unit, and a reviewer who can stop a holiday par. Outputs: a task, a G11169 pass/fail reason, and a restaurants exception list—not a promised food-cost percentage. This is a proposed, configurable capability, not a live customer result.

When the Saturday book moves, OpenTable covers can be a second trigger, not a stock number. US Tech Automations can read a reservation export, compare booked covers to the Tuesday baseline, and hold the romaine PO until a chef confirms the par should rise by a whole case. The finance and accounting agent path is the matching product route for that purchase handoff. Nothing here is a measured conversion rate.

For groups that already compared inventory products, the multi-location inventory tools list is the catalog view; this page is the par-to-PO decision.

Who this restaurants page is for

This comparison is for a chef, GM, or operator who already sells food on a POS, may book covers in a reservation product, and needs a named owner for the par list. It assumes you already buy from vendors somewhere else. Foodservice jobs: about 12 million according to BLS (2024), about 12 million jobs in food services and drinking places, which is why a par process that depends on unpaid heroics does not scale across a closing crew.

Red flags: skip a restaurants orchestration layer for automate track inventory par when Toast inventory already decrements the only items that matter, when OpenTable is being asked to store cases it cannot store, or when nobody will own item IDs. Do not buy OpenTable to replace a walk-in count. Do not buy a second POS to fix a missing inventory SKU on the first quote.

When NOT to use US Tech Automations: leave it out when Toast inventory already is the par process, when a vendor portal already suggests the only order, or when a restaurants no-code scenario with error branches already notifies the GM. honest restaurants self-selection beats a second automate track inventory fee.

Common par-list mistakes

Setting par from the last stockout instead of from depletion plus covers. Ignoring vendor pack size so the “smart” order is a quantity the distributor cannot pick. Counting in recipes while buying in cases. Letting an 86 happen in the POS without a matching inventory adjust. Treating OpenTable covers as on-hand stock. Skipping a reviewer on holiday weeks. Buying inventory software on a POS quote that does not include inventory.

Par-level FAQ

Should a restaurant pick Toast or OpenTable for pars?

Toast wins when item depletion and on-hand quantity must live in the POS. OpenTable wins when the only gap is guest-count planning.

Does Toast inventory come with every POS quote?

No. Confirm inventory, purchasing, and item export on the quoted edition before you treat Toast as the par restaurants system of record.

Is OpenTable an inventory system?

No. OpenTable runs reservations and guest experience. It does not replace a stock count.

When NOT to use the workflow team?

Do not add the workflow team when native POS inventory already covers the motion, when a vendor portal already writes the only PO you need, or when there is no second system to sync.

How should we pilot par-level reordering?

Trial 30 restaurants days: 12 SKUs, 8 count mismatches, 10 cover spikes, and 6 vendor-pack rounds. Scale on unique item IDs and reviewer holds, not on chrome. Put a human hold in front of every vendor send.

Choose the stock ledger, then the cover signal

Choose Toast for item-level depletion, OpenTable for covers, and a written par list that uses vendor packs. Then prove unique item IDs from check to PO.

The team at US Tech Automations can map a configurable depletion-to-PO trail. Review US Tech Automations after you have named the automate track inventory POS edition, the reservation export, and the reviewer.