Price, Promo and Memory: Intent Against Fact

The price in your POS is what you meant. The price in the delivery app is what your customer pays, and the two do not always match: in August 2026 the same dish from the same UAE kitchen matched price between Careem and Talabat only 85% of the time. When that fact comes back, day by day, into the system where you set the price, a one-off audit becomes a standing control, and a dated record replaces memory in every dispute.

The price in your system is an intent. The price in the app is a fact.

Your finance director approved a new price list from the first of the month. In your POS there is one price per dish. In the delivery apps there can be several prices, and the POS does not tell you which one the customer sees.

This is common across the market, not an exception. Among restaurants monitored by Kitchain in the UAE, in August 2026, the price of the same dish from the same kitchen matched between Careem and Talabat 85% of the time. Between Noon and Talabat it matched 72%, between Careem and Deliveroo 73%. Where the price differed, Talabat was usually the cheaper side. We mapped that market picture in One Dish, Four Prices.

Even the best-aligned pair does not agree on every dish, and the others agree less often. The only way to know which price the customer sees is to check each storefront, platform by platform. Same for a discount: the plan says when it ends; the app shows whether it did.

Checking by hand does not scale. As an example: 40 branches × 5 platforms × 120 items is 24,000 prices. Nobody opens 24,000 prices, so the price in the POS is taken as the price the customer sees.

Kitchain monitoring already shows what customers actually see, with no integration and no platform login. This article is about the next step: what changes when that fact comes back into the system where you set the price. It is not about orders, and it is not about pushing menus to platforms. Kitchain changes nothing on a storefront. It shows what is there.

What this means for a restaurant: your POS holds your intent. Only the storefront tells you the price you actually sell at.

Your menu moves even when you don’t

In August, menus drifted between platforms within a month. A clean audit at the start of the month could already be stale by month end.

What happened by month end to dishes on restaurants’ delivery menus on 1 August 2026:

The last column is the one that matters. A listing problem means the dish kept selling on another platform of the same brand. The restaurant did not withdraw it; it dropped off one storefront. In the UAE that was four in five disappearances.

The drift is not the same everywhere. In the UAE in August, the share of items that vanished was 4.7% on Deliveroo, 3.8% on Talabat, 3.7% on Keeta, 2.9% on Careem and 0.7% on Noon. Noon was the only platform in our sample where nearly half the disappearances, 44%, were genuine withdrawals from sale.

Gaps also last. A vanished UAE dish stayed missing for a median 17 days, and one in ten for 25 days or more.

The menu you meant keeps changing too. In the UAE, new items that appeared in August came to 25% of the starting menu. Your intent moves, the storefront moves, and the two do not move together.

Scenario — the wrap that vanished from one app. Picture a delivery-only wrap brand on five platforms with a 70-item menu, which is 350 dish listings (5 × 70). At the UAE rate of 2.9% a month, that is about 10 dish listings gone by month end (350 × 2.9% ≈ 10). Wraps were the category that vanished most often in the UAE in August: 13% of the category’s items. Say the signature wrap drops off one storefront and keeps selling on the other four, which is how 79% of UAE disappearances looked. At the median of 17 days, the gap persists for almost two and a half weeks (17 ÷ 7 ≈ 2.4). A dated record shows when it began, without waiting for someone to infer it from sales, and on which dates the wrap was seen on the other platforms.

What this means for a restaurant: a one-off audit goes stale faster than a chain can repeat it. Control has to be continuous, or it is not control.

From a one-off audit to a standing control

The fix is a loop: you set the price, you see the fact, you see the date the two parted. Kitchain supplies the fact; your system holds the intent.

Here is what can come back into your own systems, per restaurant and per platform:

• The current menu on each platform: dishes, categories, the price the customer sees and, while a dish is discounted, the price before the discount.
• Price history with dates: a price change is visible from the day it was first seen, so every gap has a start date.
• Dishes that left the menu stay in the history, so a disappearance has a date too.
• A summary across every restaurant in the chain, not one branch at a time.
• A daily promo history on each platform: store-wide offers and dish discounts, with the mechanic and the conditions the platform shows, such as minimum order, cap, promo code and validity. “Checked, and no promotion was shown” is recorded differently from “not checked”.
• Storefront status history: when each listing was on and when it was off.

Matching a platform’s dishes to the items in your POS, and flagging where the fact departs from the intent, is the work of your own system or your integrator. Kitchain brings what the customer sees back into the system where you set the price.

Scenario — the price rise that half-happened. Picture a 12-branch burger chain in Dubai on Careem, Talabat, Noon and Deliveroo, with an 80-item delivery menu and one city price list. Even before any change, the fact already differs from the intent. At the 85% match rate between Careem and Talabat, about 12 dishes show a different price on the two apps (80 × 15% = 12). At the 72% match rate between Noon and Talabat, about 22 do (80 × 28% ≈ 22). Then the CFO raises prices on the first of the month. On some platforms the new price shows the next morning. On one, part of the menu is still at the old price at month end. Without a record, the CFO sees one price in the POS and assumes it is live. With the record in the chain’s own system, the CFO sees a list of gaps, each with the date it began, and a list of which ones closed.

The record has honest limits:

• Dishes and prices only. Modifiers, extras, photos and out-of-stock marks are not in the data.
• Renames can look like replacements. On some platforms, a renamed dish looks like one dish leaving the menu and another arriving.
• Same-day reversals leave no trace. A price that changes and changes back within the same day does not show up.
• Sizes differ by platform. On some platforms every size is a separate dish; on others one dish carries one price.

The price in the record is the price the platform shows, including a discounted price. The record does not say who set a discount, the restaurant or the platform.

What this means for a restaurant: the check stops being a project someone runs now and then. It becomes a list your system keeps current, with a date on every line.

Memory against memory: the dated record as an asset

In a dispute with a platform or a franchisee, the side with a dated record holds the conversation on facts. Without one, every dispute is memory against memory.

The questions repeat: what price was on the storefront, was the promotion live, was the dish on the menu, was the listing open. Three records answer them, and a chain can keep all three in its own systems:

1. Price by dish, with dates: what each dish cost on each platform, and from which day.
2. Daily promo history: which offers and dish discounts the customer could see, on which days, on which platform.
3. Storefront status history: when each listing was on and when it was off.

The numbers explain why memory loses. A vanished UAE dish stayed missing for a median 17 days. Over that span, what either side remembers of a specific screen is not a record. And in 79% of UAE disappearances the dish was still selling on another platform of the same brand at month end. In a “the dish disappeared” conversation, the record shows the date it dropped off and on which dates it was seen elsewhere.

There is a second reason to keep the record at home. History at any provider is limited in depth, and the data of a branch that closes drops out over time. Your own copy keeps it as long as you need, next to your own sales data. It makes year-over-year comparison possible and outlives the branch.

Be precise about what this is: a record for the conversation, dates and prices as the customer saw them. Whether a platform acts on it is the platform’s decision.

Marketing teams face the same gap from the other side. A campaign plan is an intent too. The daily promo history shows on which days the offer was actually visible on each platform, and which checked days showed no promotion.

What this means for a restaurant: the dated record is an asset you build every day you keep it. The day you need it, it is too late to start.

For the franchisor: the standard becomes measurable

A franchise standard on price, menu and promotions can be checked against a dated record, not against a store visit and a screenshot.

The record protects the franchisee as well. In four of five UAE disappearances, the dish dropped off one storefront while it kept selling on another. The conversation moves from blame for a broken standard to the question of from which day, and on which platform.

Scenario — the franchisee dispute. Picture a café franchise with franchisee-run branches in Kuwait. The network standard is one menu and a brand-wide discount for a fixed period. A screenshot shows that one franchisee’s signature salad is missing on one platform. The franchisor attributes the missing dish to a broken standard; the franchisee says they did not remove it. The record shows the date the salad stopped showing on that platform, and on which dates it was seen on the other two. This is an example of the broader listing-problem class, which accounted for 84% of Kuwait disappearances: the dish was still selling at month end on another platform of the same brand. The same record shows on which days the brand discount was visible on each platform. The dispute ends on dates, not on memory.

What this means for a franchisor: the standard stops being a matter of who saw what. It becomes something both sides can read.

How we measured

Prices. A group is one dish from one restaurant listed on two or more platforms. A match means the identical price on both platforms of a pair. The spread of a group is (max − min) / min. Several snapshots in August 2026, restaurants monitored by Kitchain in the UAE, methodology v1. The price is the one the platform shows the customer.

Disappearances. We counted dishes that were on a restaurant’s menu on a platform on 1 August 2026 and were not seen for 5 days or more by month end, while the listing itself stayed live. Listings that went dark entirely are excluded. A listing problem means the dish was still selling at month end on another platform of the same brand; withdrawn from sale means it was gone from all of them. New items are dishes that first appeared in August, as a share of the starting menu. Sample: restaurants monitored by Kitchain in the UAE, Kuwait and Saudi Arabia, August 2026, methodology v1.

What these numbers are not. They describe what customers saw on public storefronts. They do not say why a price differed or why a dish dropped off, and they are not a verdict on any platform. Menu data has been collected since July 2026 and the collection method changed in September, so there is no month-on-month comparison in this article.