Google Ads or Yandex Direct in Uzbekistan: choose by audience and economics

Choosing an advertising platform in Uzbekistan starts with a specific sale: who needs the product, where the customer is served and how much the company can spend acquiring them. A search engine’s overall popularity does not answer those questions. Without your business data, there is no defensible universal winner. The following matrix and bounded test plan support campaign preparation. They are not a published experiment or an estimate of Google and Yandex market shares.

1. Define a suitable enquiry

For services, qualification might mean the correct region, an actual task, an acceptable budget and a reachable contact. For a store, it might mean a paid order with sufficient margin and no cancellation. Agree criteria with sales before launch. An enquiry, a qualified lead and a sale are different stages; Google documents qualified and converted lead categories at https://support.google.com/google-ads/answer/11459091. Record source and rejection reason in your CRM. A phone click or form opening should not silently replace the business outcome used for optimisation.

2. Calculate an affordable acquisition cost

Start with the contribution from a sale after variable costs and agree the portion available for acquisition. Illustrative example: if acquisition can cost UZS 600,000 per customer and one in five qualified leads buys, the reference is UZS 120,000 per qualified lead. This is arithmetic, not a market forecast. The real limit depends on margin, repeat purchases, refunds and data quality. Include media, management, creative production and analytics when comparing sources. Convert costs to one currency with a dated exchange rate; do not compare only click prices.

3. Choose a testable audience segment

Use customer enquiries, available search terms and the languages you actually support. For B2B, include buying roles and approval time. If the team cannot respond in Uzbek, an Uzbek ad without corresponding service may create unsuitable enquiries. Results for one segment do not establish the platform’s overall suitability.

4. Separate search from network placements

A search ad responds to an expressed query; network advertising appears in a different context and needs its own hypothesis. Google distinguishes Search and Display at https://support.google.com/google-ads/answer/2567043. Yandex Direct supports placement selection for search and its advertising network: https://yandex.com/support/direct/en/unified-performance-campaign/create-campaign. Check the settings currently available in your account. Keep placements and reports separate for an initial comparison. A low blended CPL can hide poor quality from one source; the same goal label does not make audiences equivalent.

5. Prepare the landing page and measurement

Match the destination to the ad’s language, service and conditions. Test the mobile form, enquiry delivery, confirmation and manager response. Separate test submissions and duplicates, and agree the source of truth for sales. Keep personal information out of public URLs and analytics parameters. For long sales cycles, decide which intermediate stage can be reviewed. Otherwise a new campaign may seem weaker simply because its sales have not matured while the alternative is judged through a different attribution period.

6. Write a bounded test agreement

1. Choose one offer, region and language segment; agree separate placement hypotheses.
2. Fix the spending cap, qualification goal, observation window and early-stop conditions.
3. Check payment, advertiser requirements and moderation before promising a launch date.
4. Compare sources using one lead-quality log and account for delayed sales.

With a small budget, sequential testing may be more practical than splitting spend between platforms. Different weeks also have different demand, so record that limitation. An expensive enquiry that becomes a sale and a cheap unsuitable enquiry need different decisions. If evidence is thin, continue bounded collection or revise the hypothesis instead of declaring a winner from a few clicks.

7. Let evidence choose the next action

Where enquiries qualify, expand gradually and recheck economics. For irrelevant queries, refine targeting; for an unclear offer, improve the destination; for lost sales, repair handling. Advertising management: https://samsa.agency/en/services/paid-media. Landing pages and analytics: https://samsa.agency/en/services/performance. Before commissioning a website, review https://samsa.agency/en/blog/stoimost-razrabotki-sayta-uzbekistan; for selecting a team, use https://samsa.agency/en/blog/kak-vybrat-veb-studiyu-tashkent. The outcome is an agreed budget and next testable hypothesis, rather than faith in a universally superior platform.