Geo-targeting in Google Ads: configuring impressions by city district
'All of Tashkent' geo-targeting feels like the safe choice, but it's usually a way to pay for impressions from places nobody can physically reach your business from.
Radius versus polygons
Radius targeting around a location is quick and simple to set up, but it ignores the city's real geography: a three-kilometre circle might cover an industrial zone on one side and stop short of a boundary on the other where most of your customers actually live.
Polygon targeting, where you hand-draw a zone on the map, more accurately reflects real transport and foot-traffic routes, but takes more time to set up and needs manual updates when the service area changes.
For a business with a defined delivery zone — a food delivery café, say — a polygon almost always beats a radius, since delivery zones are rarely a perfect circle around the location.
Multiple campaigns for different districts
For a business with several locations across the city, split geo-targeting by campaign rather than setting one wide radius — this lets the ad copy name a specific address and hours for the nearest location instead of a generic line.
Competition can differ noticeably between Tashkent districts: central and business districts carry higher click prices due to competitor density, while the same services in residential districts often cost less at comparable demand.
Test new districts with a separate small budget before folding them into the main campaign — demand and purchasing power vary by district, and merging an unproven zone with a working one muddies the overall statistics.
Bid adjustments by location
Once click statistics accumulate by district, set bid adjustments: raise bids in higher-converting districts and lower or exclude districts that generate clicks but no leads.
Check the user locations report separately from the target location report — Google can show ads to people physically outside the set zone whom certain signals still attribute to the target region.
Setting 'people in your targeted locations' instead of 'people showing interest' is critical for local business — the second option by default extends reach to people merely interested in the city topic who could physically be anywhere.
Common geo-setup mistakes
Leaving targeting on the default 'Uzbekistan' for a business with one physical location in Tashkent is a typical mistake in new accounts, leading to impressions in regions where nobody can reach the venue.
Forgetting regional language differences in geo-targeting: in Tashkent it's sensible to emphasise Russian-language keywords, while in regional centres the Uzbek-language share of keywords deserves more weight, since audience language balance differs.
Ignoring exclusions for neighbouring countries that sit geographically close but where delivery or a visit is physically impossible — Samarkand borders Tajikistan closer than search logic assumes, and without an explicit exclusion, part of the budget goes to irrelevant impressions.
A practical setup plan
Start with a narrow radius or polygon around the location that matches the real service area, not an intuitive 'the whole city knows us' feeling — widening reach is always easier than narrowing it after budget's already been wasted.
After a month of running, check the click and conversion geolocation report and adjust the zone based on fact rather than the original plan — real data almost always differs from the initial guess.
Revisit geo-settings quarterly, especially if the city is growing or new residential complexes have opened in target districts — audience density in Tashkent's districts shifts noticeably faster than it seems.