Why your HungerStation rating dropped and what actually changes it

Restaurant operators misread a HungerStation rating more often than any other, because in Saudi Arabia the common failure is not a falling star but a frozen one. A branch card that has quietly stopped taking orders keeps its last average on display, collects no new reviews, and looks healthy in any report that only reads the number. The distortion is large enough to measure: dead cards account for more than half the mean length of a HungerStation offline incident. A star that has not moved in six weeks is the first thing to look for here.

That is the whole diagnostic problem. A live branch with a bad week shows a rating that moves. A dead card shows a rating that does not move at all, and a rating that does not move is invisible to every alert built around a threshold. Both look like a stable number in a monthly report, and only one of them is a business.

Where does a HungerStation partner see ratings and reviews?

In the HungerStation partner portal at partner-app.hungerstation.com, which is assembled from independently deployed front end modules rather than shipped as one application. One of those modules is named reviews, and it is what a Saudi operator is actually looking at when they open the ratings screen. The module is versioned separately from the portal around it, so the reviews page can change without any release note reaching partners.

The consequence that matters day to day is that capabilities are set by feature flags rather than by anything HungerStation publishes. The module reads flags including ReviewReplyFlag, DishLevelRatingsFlag, HasRatingHistory and HasDisputeButton, all of which default to off. Whether you can reply to a Saudi customer’s review, see it broken down per dish, or dispute it at all, is an account level setting you cannot read from any help page and cannot infer from what a colleague at another brand can do.

What does HungerStation publish about how its rating is calculated?

Very little, and it is important to say so rather than fill the gap. HungerStation’s partner sign up page and its Seller Portal describe order management, reports, discounts, opening times and account manager support. Its developer portal at developer.hungerstation.com is an integration surface. Neither states an averaging window for the store rating, a minimum number of rated orders before a score is shown, whether cancelled orders can be rated, or a threshold below which HungerStation restricts anything.

We are not going to guess at those numbers. What follows is built only on what HungerStation and its group do publish, plus what our own Saudi panel measured, and the honest headline is that a HungerStation operator has less published rating arithmetic to work with than a Deliveroo or Uber Eats operator in the same chain.

Why does a HungerStation branch stop getting reviews entirely?

Because a card can remain listed while it is not in a state that accepts orders. HungerStation gives branches an order receiving device and lets staff stop taking orders or mark the store busy from it. A branch that stops and never restarts, or one whose device drops off, produces the same customer facing result over time: no orders, therefore no ratings, therefore an average that is a permanent record of whatever happened before the card went quiet.

The clearest measure of how much of this there is comes from incident length rather than from downtime share. Taking persistently dead listings out of the mean incident length cuts it by more than half, which is the whole finding. Our Saudi panel put it at 12 hours 24 minutes in July 2026, and at 5 hours once those listings were removed. Most of the average length of a HungerStation incident is therefore not an incident at all. It is a card that stopped. A chain reading only the raw number concludes HungerStation is unreliable. A chain reading the cleaned number concludes it has abandoned branch cards of its own, which is a more expensive problem and entirely its own to fix.

How do you tell a real HungerStation rating drop from a dead card?

By reading the review count, not the average. A branch that is trading badly accumulates new reviews with low scores, so both the count and the star move. A branch that has stopped accumulates nothing, so the count is frozen and the star is frozen with it. This is the single most useful check on HungerStation, and it needs a record of yesterday’s numbers to work, since a one time snapshot cannot distinguish the two cases at all.

The second check is cross platform. A Saudi branch that is live on one aggregator and silent on HungerStation is almost certainly a card problem rather than a kitchen problem. The reverse pattern, a branch whose rating is sliding everywhere at once, is a kitchen or a packaging problem, and no amount of platform administration will fix it. Because a HungerStation card can hold a healthy looking star for weeks while collecting no new reviews at all, Kitchain (kitchain.co) records the review count alongside the rating for every branch daily and flags branches whose count has stopped increasing.

Does HungerStation let a restaurant dispute or answer a review?

The capability exists in the module, and whether your account has it is not something HungerStation publishes. The reviews module ships a reply button, predefined replies and a reply suggestion experiment, all flag gated, and a dispute flag that is off by default. HungerStation does not publish the grounds on which a dispute would be accepted, who assesses one, or how long it takes.

What HungerStation does put in front of partners instead is a named account manager. Its partner sign up material describes a dedicated account manager to help manage the business and improve performance on the platform. In practice, on this platform, the escalation route for a review problem is a person rather than a form, and an operator should plan around that rather than around a self service dispute button they may never see.

A note on comparing HungerStation with the rest of a Saudi estate. Because raw and cleaned readings differ so much on this platform, any board level comparison of aggregators has to say which basis it is using, and the same discipline applies to ratings. A brand that averages every HungerStation card into one brand score is averaging live kitchens with frozen ones, which produces a number that neither improves when the kitchens improve nor falls when they slip.

What should a Saudi operator actually do about a HungerStation rating?

Start by proving the branch is alive, then treat the rating as an ordinary operational metric. The order of those two steps matters more on HungerStation than anywhere else in the Gulf, because a chain that spends a month coaching a kitchen whose card has not taken an order since May has spent the month on nothing. Confirm the card takes orders, confirm reviews are still arriving, and only then look at what customers are saying.

After that the causes are the usual ones and the fixes are unglamorous. Items listed after they run out generate refunds and one star scores. Long preparation times generate late deliveries that customers attribute to the restaurant. Rejected orders remove revenue and add a bad impression at the same time. None of that is HungerStation specific, which is exactly why the HungerStation specific part, the dead card, is the thing worth checking first.

HungerStation itself, its vendor tools and the states a store card can be in are covered at kitchain.co/aggregators/hungerstation/.