Yandex.Eda: How to Get Your Restaurant to the Top Without Paid Ads

Your kitchen sends out food that regulars praise in person, and your listing still sits on the third screen of Yandex Eda, below a pizzeria with a 4.3 rating and a 50-minute delivery promise. Every order already costs you a commission. Now a platform manager suggests paid placement on top of it, and the math stops working before you even sign.

There is a cheaper path. Yandex Eda ranks restaurants on signals you control from your own kitchen and your own admin panel: rating, real cooking time, cancellation rate, menu quality, and how often people who open your card actually order. Fix those and your position moves without a promotion budget. We have watched partner restaurants climb from the depths of a category into the first two screens in 4 to 8 weeks doing exactly this kind of work, and nothing else.

This playbook walks through how the ranking works, a step-by-step optimization sequence, review handling, menu and photo setup, and the unit economics that decide whether delivery platforms make you money at all. Numbers in the examples are illustrative. Commission terms and interface details change, so verify current conditions in your partner account before you build a budget on them.

One note for readers comparing platforms: Delivery Club, once the main rival, was absorbed into the Yandex ecosystem after the 2022 asset swap with VK and later folded into Yandex Eda. For most Russian cities today, Yandex Eda is the delivery aggregator, which makes your position inside it worth real money.

How Yandex Eda Ranking Works for Restaurants

Yandex does not publish its ranking formula. What follows comes from partner documentation, platform statements, and patterns we see across restaurant accounts. Treat it as a working model, and expect weights to shift over time.

The feed a customer sees is personalized first: their location, their order history, their favorite cuisines. You cannot control that layer. Underneath it sits a quality score built from signals you control almost entirely.

Rating

Your average score from customer ratings is the loudest signal. It appears on the card, it filters who taps at all, and it feeds the algorithm. The difference between 4.5 and 4.8 sounds small. In practice it separates restaurants that appear in "top rated" style selections from those that never do. Recent ratings weigh more than old ones, so a rough month drags you down faster than a good year props you up.

Actual cooking time versus promised time

The platform tracks how long you take to confirm an order and hand it to a courier, then compares it against the time you promised. Consistently beating your promise builds trust with the system. Consistently missing it gets your delivery estimate inflated on the card, and long estimates kill conversion. A customer choosing between two similar sushi places at 25 and 55 minutes rarely picks the second.

Cancellation rate

Every order you cancel, whether the dish ran out or the kitchen got slammed, is a broken promise the platform paid for in customer goodwill. Cancellation rate is punished harshly. Partners with chronic cancellations report losing visibility long before any human manager contacts them. Keep this metric near zero. It matters that much.

Menu quality and card conversion

The algorithm watches behavior: of the people who opened your card, how many ordered? Photos, dish descriptions, category structure, and price presentation all feed that conversion number. A card with dark photos and one-word dish names converts poorly, the platform notices, and your impressions shrink. This is the lever most restaurants underuse, and it costs nothing beyond a photographer and an afternoon.

Order volume and repeat orders round out the picture. The platform earns commission, so it favors restaurants that turn impressions into revenue. That creates a flywheel: better card, more orders, more visibility, more orders. Your job is to start the wheel.

Step-by-Step Optimization: The First Three Weeks

Here is the sequence we run with new restaurant clients. It front-loads the fixes that move ranking fastest.

Week 1: measure and repair the basics.

Pull your numbers from the partner dashboard: rating trend, average cooking time, cancellation rate, card-to-order conversion. Write them down. You will compare against these in a month.
Remove every dish you cannot reliably cook. A shorter menu with 100% availability beats a long menu that generates "item unavailable" cancellations.
Rewrite your promised cooking time to something you hit 95% of the time. Honest 30 minutes beats fantasy 20.
Answer every unanswered review from the last 90 days, negatives first.

Week 2: rebuild the storefront.

Shoot new photos of your top 20 dishes (details in the menu section below). Upload them the same week.
Rewrite names and descriptions for your bestsellers: ingredients, portion weight, what makes it worth ordering.
Restructure categories so a first-time visitor finds a main dish in two taps.

Week 3: turn on demand mechanics.

Launch one platform promotion: a gift item over a threshold, a first-order discount, or free delivery above a set amount. These are standard partner tools, available without an ad budget.
Build combo sets around your highest-margin dishes to lift average check.
Set a daily 10-minute routine: check reviews, check cancellations, check availability. Small, boring, decisive.

Expect early movement in 2 to 4 weeks if your starting point was neglected. A card that was already decent moves slower, because the remaining gains sit in operations, and kitchens change slower than photos.

Menu and Photo Setup That Converts

Photos carry the card. On a screen full of thumbnails, the customer's eye lands on food that looks like food. Phone snapshots under kitchen fluorescents lose to a competent photographer with natural light every time. You do not need art. You need bright, sharp, honest images of the actual dish, shot from a consistent angle, on a clean background. One illustrative reference point: a mid-size sushi cafe we worked with reshot its full menu for roughly 15,000 rubles with a student photographer and saw card conversion climb double digits within a month. Your mileage will vary, and that is the point of measuring first.

Names and descriptions do quiet selling. "Borscht" tells a customer nothing. "Beef borscht with sour cream and garlic croutons, 350 g" answers the three questions every delivery customer asks: what is in it, how much do I get, will it fill me up. Put weight in grams on everything. Russian delivery customers check portion sizes obsessively, and a missing weight reads as a small portion hiding.

Structure matters more than most owners think. Categories should mirror how people decide: Soups, Hot dishes, Salads, Combo sets, Drinks, Desserts. Put combos and bestsellers at the top of the menu. Bury nothing important below the fold.

Two pricing notes. First, account for commission when you set delivery prices; many restaurants run a separate delivery price list, slightly above dine-in, and pair it with free delivery thresholds. Second, build combos deliberately. A "dinner for two" set lifts average check, and average check feeds both your margin and, indirectly, your attractiveness to a commission-based platform.

Reviews: A System, Not a Mood

Handle reviews like a process with an owner and a deadline, because on aggregators your rating is your shelf placement.

Respond to every review within 24 hours. Positive reviews get a short, human thank-you, ideally referencing the dish. Negative reviews get a three-part answer: acknowledge the specific problem, say what you changed, offer to make it right. Never argue in public, even when the customer is wrong. Future customers read your answer as a preview of how you will treat them, and the same logic that governs online reputation management for any business applies here at full strength.

Volume matters alongside score, so ask for reviews. A card insert with a QR code in every bag, a courier script, a small bonus for staff when tagged reviews come in. Keep incentives on your side of the counter: reward your team for service that earns reviews. Paying customers for positive reviews violates platform rules and, when detected, hurts more than it helped.

Fake or mistaken negative reviews happen. If a review describes an order you never received, collect evidence from your order log and file a support ticket. The platform removes reviews it cannot match to real orders, though response time varies. This is an ongoing maintenance task, ten minutes a week, worth every minute.

One number to internalize, as an illustrative model: if you sit at 4.4 with 50 ratings, you need roughly 25 five-star ratings to reach 4.6, and one new 1-star wipes out about ten of them. Ratings are asymmetric. Preventing one bad experience does more for your score than soliciting three good reviews.

Operations: Cooking Time and Cancellations

This section is short because the advice is blunt. Ranking follows operations.

Track the interval from order receipt to courier handoff, daily, per daypart. Most kitchens discover one bottleneck: packaging during rush, a single fryer, a printer nobody watches. Fix the bottleneck and your real cooking time drops without hiring anyone. An illustrative case from our practice: a pizzeria cut average handoff time from 55 to 38 minutes purely by moving packaging out of the cook line and assigning one person to assembly during peaks. Repeat orders rose within weeks, and position followed.

Cancellations deserve a hard rule: the kitchen never cancels for stock reasons. Sync menu availability in real time. If the salmon runs out at 19:40, the dish goes dark on the platform at 19:41. Every avoided cancellation protects both your rating and your standing with the algorithm.

Unit Economics: Does the Platform Even Pay?

Ranking work only matters if each order earns you money, so run the math before you chase volume. If you have never modeled this, our guide to unit economics covers the general method; here is the delivery-platform version.

Aggregator commission depends on your model. Delivery by platform couriers costs substantially more than delivery with your own couriers or pickup, and exact rates vary by contract, city, and category. Check your current terms; do not budget from a blog post, including this one. The table below shows an illustrative model for one order at a 30% commission with platform couriers.

Line item	Illustrative value	Share of check
Average check	1,200 ₽	100%
Platform commission (30%)	-360 ₽	30%
Food cost	-360 ₽	30%
Packaging	-60 ₽	5%
Promo discounts (averaged)	-60 ₽	5%
Labor and overhead share	-240 ₽	20%
Contribution per order	120 ₽	10%

All numbers are illustrative, and the fragility is the lesson. At a 10% contribution margin, a 5-point commission change or a careless sitewide discount flips you to breakeven. Three moves defend the margin. Raise average check with combos, since commission is proportional and your fixed costs per order are covered faster. Push repeat orders, because a returning customer costs you nothing to acquire; the platform even lets you nudge them with targeted promos, the delivery-world cousin of customer reactivation. And route regulars toward direct channels where contract terms allow it: your own site, pickup, a phone number on the packaging insert. Aggregators are a discovery engine. Loyalty should live with you.

A note on "free" promotions: a gift dish or first-order discount comes out of your margin, so treat each mechanic as a customer acquisition cost and measure whether those customers return. A promo that brings one-time bargain hunters at negative margin is paid advertising wearing a disguise.

Common Mistakes That Keep Restaurants Buried

Chasing paid placement before fixing the card. Promotion multiplies your conversion rate. Multiplying a bad rate wastes money.
A promised cooking time you miss daily. The system sees through optimism and inflates your visible estimate anyway.
Ignoring negative reviews. An unanswered complaint tells hundreds of readers you do not care, and tells the algorithm your engagement is low.
Stale menu data. Unavailable dishes still listed, old prices, a summer menu in November. Each mismatch generates cancellations or disappointed one-star ratings.
One-word dish names with no weights. "Salad, 320 ₽" loses to a competitor who says what is in it and how much it weighs.
No competitor monitoring. Your prices, photos, and promos exist in a lineup. Check the top five cards in your category monthly and note what they do that you do not.
Measuring nothing. Without baseline numbers you cannot tell whether last month's changes worked, and you will quit the practices that were quietly paying off.

FAQ

Do I still need Delivery Club?

No. Delivery Club was absorbed into the Yandex ecosystem after 2022 and later merged into Yandex Eda. Focus your aggregator effort there, plus your own direct channels.

How long before my position improves?

First movement typically shows in 2 to 4 weeks if you started from a neglected card: better photos and cleared review backlogs act fast. Reaching a stable top-10 spot in a competitive category usually takes 2 to 3 months, because rating and repeat-order signals accumulate slowly. These timelines are estimates from partner experience, and results depend on your category's competition.

What rating do I need to rank well?

There is no published threshold. In practice, cards below roughly 4.5 struggle in competitive categories, and every tenth of a point above that helps. Watch your three-month trend; a single snapshot misleads.

Is it against the rules to ask customers for reviews?

Asking is fine: QR codes, courier scripts, packaging inserts. Paying customers for positive reviews or fabricating them violates platform rules and risks penalties. Reward your staff for service quality instead.

Should I raise delivery prices to cover commission?

Most restaurants do, within reason. A markup of 10 to 15% over dine-in prices (illustrative range) is common and rarely hurts conversion if photos and descriptions justify value. Test it on part of your menu and watch conversion in your dashboard before rolling it out.

Do platform promotions hurt my margin too much to use?

They cost margin, so use them the way you would any acquisition channel: measure cost per new customer and repeat rate. A first-order discount that brings customers who order again next month pays for itself. A permanent sitewide discount usually does not. Run each mechanic for two weeks, read the numbers, keep what earns.

Conclusion and Checklist

Position on Yandex Eda is built in your kitchen and your admin panel, week after week. Before you spend a ruble on promotion, work through this list:

Baseline metrics recorded: rating, cooking time, cancellations, conversion
Menu trimmed to dishes you can always deliver, availability synced in real time
Professional photos on all bestsellers, weights and ingredients in every description
Promised cooking time you hit 95% of the time
Every review answered within 24 hours, negatives first
One promotion live, measured as acquisition cost
Combos lifting average check
Unit economics per order calculated with your real commission

Restaurants and delivery are one instance of a wider pattern we work with across service businesses, from hospitality marketing to lead generation for B2B firms: platforms reward operators who measure, fix the funnel, and compound small advantages. If you want a second pair of eyes on your delivery-platform economics or your listing, ask Lead The Way for a short audit of your current setup. We will show you which two or three levers are worth pulling first, and which spending you can skip.