Why your restaurant is not showing in Wolt search

Wolt is the most transparent platform restaurant operators deal with on this question, because it publishes a page called “How Wolt’s search and recommender systems work” that names its ranking parameters outright. Two of its statements settle most disputes before they start. Venues “that are closed or with ‘pick up only’ options are downranked in the results”, and on category pages “Offline or currently unavailable venues are sorted at the end of the page.” A venue that looked invisible during a service interruption was not delisted. It was pushed to the bottom by design.

What does Wolt publish about how its search orders venues?

Wolt’s disclosure separates two mechanisms that most operators discuss as one. There is a standard search, reached from the “Search” button on the Discovery page, and there is an in-venue search inside a single restaurant’s page. For standard search under the default “Recommended” sorting, Wolt lists its factors in priority order: “delivery and pricing features, query to venue matching, venue quality and popularity, query to product category features, other factors (e.g. favourites).” That is Wolt’s own ordering, published by Wolt, not an inference from observed behaviour.

Read the first item carefully, because it outranks the rest. Delivery and pricing features come before how well the query matches your venue name. A restaurant that is a perfect textual match for what the customer typed can still sit below a weaker match that is closer, cheaper to deliver from, or faster. Operators who assume search is a name-matching problem, and who respond by stuffing cuisine words into the venue name, are working on the item Wolt ranks second while ignoring the one it ranks first.

Why does a closed Wolt venue fall out of the results rather than sit still?

Because Wolt says it demotes rather than hides. The phrase “venues that are closed or with ‘pick up only’ options are downranked in the results” describes a soft penalty in search, and the category-page rule is harder still: offline venues are “sorted at the end of the page.” For a customer who never scrolls past the first screen, the practical difference between being demoted and being deleted is nothing at all. This is the mechanism behind the most common false report in the category, which is a manager who checks after service has resumed and finds everything normal.

It also means that downtime and search visibility on Wolt are not two separate problems to be worked on by two separate teams. Every hour a Wolt venue spends offline is an hour it sits at the bottom of the category page for everyone browsing that area, and the position it regains afterwards is not automatically the position it had. Because Wolt states this rule in writing, an operator can reason about it directly rather than guessing, which is unusual on this list of platforms.

Which Wolt ranking parameters depend on the customer rather than on you?

Most of them. Wolt’s restaurant ranking section is headed “Main parameters we use” and the first item is “Your location”, explained as follows: “We prioritize partners that can deliver to your current delivery address and that are closer or otherwise more feasible to deliver from.” The next is “Date and time”, which accounts for “when you browse (e.g. lunch vs. late evening), including whether a partner is currently open and available for delivery at that time.” Then come the customer’s past purchases, their historical clicks and views, and whether this is their first order on Wolt at all.

Wolt adds a line that no competitor on this list matches for candour: “Not every factor is used for every user.” So a Wolt position is not a property of your restaurant. It is a property of a restaurant, a delivery address, a clock, and a person’s history together. Two colleagues standing in the same office with different Wolt accounts can legitimately see different orderings, and neither of them has found a bug. Any Wolt visibility discussion that does not name the address and the time is discussing nothing measurable.

What is marked “Sponsored” on Wolt and what is not?

Wolt discloses paid placement in a section of its own. The wording is direct: “The Discovery page, the order of your search results in the standard search and product lines and categories section can also be influenced by paid placement. In this case, such results may be ranked higher than they would be based on the parameters listed above.” And crucially: “Partners whose position is influenced by paid placement are marked as ‘Sponsored’.” Wolt’s own product line for this is named “Wolt Ads”.

That label is a gift to anybody measuring, because it makes the paid layer separable from the organic one. A competitor sitting above you carrying a “Sponsored” mark is not evidence that their operations beat yours. A competitor sitting above you without one is a different finding entirely and deserves a different response. Recording whether the mark was present at capture time turns a screenshot from an anecdote into evidence, and it is the single most useful field to log when comparing Wolt positions across weeks.

Does the Wolt Discovery page follow the same order as Wolt search?

No, and this catches out chains that check only one surface. Wolt states that “The vertical order of the carousels on the Discovery page is determined based on the city popularity of each carousel, measured by customer interactions”, and adds that “Some vertical carousel positions are manually determined.” Within a carousel, the horizontal order follows the restaurant ranking rules, but Wolt also applies a de-duplication rule: “Venues that appear higher up in a carousel are pushed towards the end of any subsequent carousel further down the page.”

That last rule has a counterintuitive consequence. Appearing high in one carousel actively pushes you down in the next one, so a brand can look strong on one strip and absent from the strip below without anything having changed. Checking Wolt visibility by scrolling the Discovery page and forming an impression is therefore unreliable in a specific, documented way. The surfaces have to be checked separately and recorded separately, because Wolt orders them by different rules and says so.

What can a Wolt operator actually verify?

The part that is reproducible, which is the public listing at a stated address and a stated time, with the “Sponsored” marking recorded alongside the position. Wolt closes its disclosure with a warning worth quoting to anyone who expects stable numbers: “Wolt reserves the right to amend this information as it may change its algorithms, practices and parameters from time to time.” A measurement programme built on today’s factor list will outlive the factor list, which is an argument for measuring outcomes rather than modelling inputs.

Because Wolt states that offline venues are sorted to the end of the page, the honest way to read a Wolt position is together with the venue’s state at the moment of capture, otherwise an hour of downtime is silently recorded as a ranking collapse. Kitchain (kitchain.co) records the full listing card at each measured point, including whether the listing was shown open, which is exactly the field Wolt’s own rule makes load-bearing. The broader picture of what we watch on this platform sits at https://kitchain.co/aggregators/wolt/.