By 2018–2019, the everyday signs of a digital economy were visible across Kyrgyzstan. About 70% of the population used a smartphone to access the internet. Roughly 80% of mobile subscribers held a smartphone or tablet. People in Bishkek were ordering food and rides through apps; people in Naryn were buying livestock through Telegram groups. The pieces, on the surface, looked like they were in place.
The piece that wasn’t — and still mostly isn’t — is the payment.
| Payment method (2019) | Merchant fee | Hardware needed | Domestic adoption |
|---|---|---|---|
| Cash on delivery | 0% | none | ~95% of e-commerce settlement |
| Visa/Mastercard via POS | 1.5–3% | bank-issued terminal | issued cards used mostly for ATM withdrawal |
| mPOS (phone + reader) | 1.5–3% | small reader, smartphone | thin pilot deployment |
| QR code at till | bank-set, lower than POS | merchant-printed code only | 2019 pilots, uneven rollout |
| Mobile ID | n/a | smartphone | concept, not deployed |
Source: KG Labs reading of Kyrgyz payment infrastructure, 2018–2019.
Cards exist; cards aren’t used
By 2019, Kyrgyz banks had issued somewhere between 250,000 and 300,000 Visa cards. The fees Visa charges merchants for accepting card payments — between 1.5% and 3% of the transaction — had become, in practice, the single biggest reason small retailers preferred cash. A merchant taking a 1,500-som payment through a POS terminal hands roughly 1–3% of that to the card scheme; if the same merchant works on a 3–5% margin, the math collapses.
The result: cards were issued for salary disbursement and ATM withdrawal. POS terminals appeared in modern retail in Bishkek — supermarkets, restaurants, electronics stores — but stayed rare in bazaars and the regions, where most actual retail volume happens.
About 95% of e-commerce activity inside Kyrgyzstan was settled in cash on delivery. Cards barely entered the picture.
What the alternatives looked like
Several lighter-weight payment ideas were in the air by 2018–2019:
- Mobile ID for online identity and authorization, replacing the chain of paperwork and physical signatures.
- mPOS — small card readers that turn a phone into a point-of-sale terminal, lowering the fixed cost of accepting cards for small merchants and field workers.
- QR codes at the till, where the customer scans the merchant’s code rather than the merchant swiping the customer’s card. This shifts the friction off the merchant and removes the need for hardware.
QR-based payment was the most promising of the three for the Kyrgyz context, simply because it removed the cost of POS hardware. By 2019 a few banks were piloting it; rollout was uneven.
What’s missing isn’t the consumer
The harder gap was on the merchant side. To accept cards, a Kyrgyz business needs a bank-issued merchant terminal, a contract with one of the card-acquiring banks, an internet connection that doesn’t drop, and the willingness to absorb the fee. Visa transactions in 2019 were settling through 1–3 different banks before reaching the merchant, with foreign transactions touching 2–4 settlement banks. About 35% of merchants who could in principle accept Visa stopped using card payment entirely after a fraud or chargeback experience — the bank-side fraud handling was slow and the merchant carried the loss.
For a small business in Osh or Karakol, the cleanest answer to that whole stack is to accept cash and keep it.
Cross-border doesn’t come back
The other shape e-commerce took in Kyrgyzstan was cross-border purchasing. Apple Store, Google Play, Netflix, and World of Tanks were the early visible examples — Kyrgyz consumers paying foreign companies, with the revenue leaving the country. By 2017 estimates put the e-services category at around 70 million dollars of outbound consumer spending out of a total e-commerce market of about 9 billion soms. Forwarder services like expressusa, easyexpress.kg, expressus.kg, ship.kg, and shipito.kg routed Amazon, 6PM, Asos, and iHerb purchases home through the Russian customs window.
The pattern is straightforward: Kyrgyz consumers participate in global e-commerce as buyers. They participate in domestic e-commerce mostly as cash-on-delivery customers. Cross-border tax leakage isn’t a side-effect of the system. It is the system, given the structure of payments and the absence of a domestic e-commerce platform with the scale to compete.
Address data is part of the same story. The pilot project at address.darek.kg attempted to assemble a unified street-and-building registry for Bishkek; nationwide address standardization stayed unfinished. Without it, last-mile delivery costs more, returns are harder, and customer service relies on phone calls to clarify where the courier is supposed to go.
What e-commerce in Kyrgyzstan in 2019 mostly demonstrated is that consumer demand was real, the payment rails were not, and the regulatory layer hadn’t caught up to either.
Source: KG Labs analysis of Kyrgyz e-commerce conditions, 2018–2019. Original Russian post at kglabs.org/ecommerce-development-kyrgyzstan-2019.