Mistake 1: a menu that is too broad
"Let's run 4 cuisines — Italian, Asian, burgers, and healthy" — that is the road to a dark kitchen's death.
The reality: each virtual concept needs its own brand (name, logo, in-app photos), its own procurement, its own kitchen processes. With 4 concepts you do not have one dark kitchen — you have 4 micro restaurants with all the overhead.
What works: 1–2 concepts maximum at launch. Expand only after 6 months of operational stability.
Mistake 2: choosing the location on "cheap"
A dark kitchen "does not need" a dining room — so the temptation is to hunt for the cheapest industrial-zone warehouse on the outskirts.
The reality: aggregator couriers take 12–18 minutes on average to reach the customer. If your site is in Kaskelen while your core customer is in the Almaly district, every order adds +25 minutes of travel. That means **returns** (cold food), bad ratings, and getting dropped from the aggregator's top listings.
What works: a 5–7 km radius covering 80% of your target audience. In Almaty that means the Almaly / Bostandyk / Medeu districts or their borders.
Mistake 3: betting on a single aggregator
"We'll launch on Choco first and figure it out later" — a typical story. Then Choco raises its commission from 18% to 25%, and there is nothing you can do — it is your entire channel.
What works: at least 2 aggregators from day one plus your own website/bot (a Telegram ordering bot costs 50–150 thousand ₸ and brings in 5–15% of turnover — at 0% commission).
A realistic 2026 mix: Yandex Eats 35%, Wolt 25%, Glovo 20%, own channel 20%.
Mistake 4: not doing the unit economics
AOV (average order value) in Almaty in 2026: 4,500 – 6,500 ₸.
Of that, the aggregator takes 18–25% (commission + customer delivery).
Tax: 12% VAT, or 3% for an IP (sole proprietorship) on the simplified regime.
COGS (food cost) must stay ≤ 30%, otherwise you are working at break-even.
Rent + utilities + staff + depreciation ~ 30–40% of revenue.
A target margin of 12–18% is normal for a dark kitchen in Kazakhstan. If yours is 5%, the mistake is somewhere in menu pricing or procurement.
Mistake 5: manual processes instead of automation
A KDS (Kitchen Display System), automatic order intake, inventory, cost accounting — at launch it all seems "expensive", and it feels like "we can manage with Excel".
The reality: at 80–120 orders a day without a KDS, the cook spends more time "reading the order on a phone" than cooking. Order errors = returns = lower ratings = falling out of the top listings.
The minimum stack for a dark kitchen in 2026:
— iiko / R-Keeper / Poster — POS and inventory
— TastyIgniter / Restaurant Connect — aggregator integration
— A KDS on 1–2 kitchen screens
Altogether ~ 200–400 thousand ₸/month — it pays for itself through fewer errors and returns within 4–6 months.
- 60% of dark kitchens in Almaty closed within the first year — almost always because of these 5 mistakes.
- Launch with 1–2 concepts, not 4.
- Location beats rent: a 5–7 km radius to your target audience.
- Do not bet everything on a single aggregator.
- Do the unit economics before launch, not after.