Skip to main content
Geo-Location Targeting Options

How does Nosto's geo-location targeting works and what to take into account

Lari Lehtonen avatar
Written by Lari Lehtonen
Updated over a week ago

Nosto's Segmentation supports targeting based on geo-locations. All Nosto products and features leveraging geo-location through segments, such as Onsite Content Personalization or directly such as weather targeting and geo-based best sellers in Onsite Recommendations make use of the same technical solution for resolving shopper's approximate location.

Geographical location is based on resolving shopper's IP-address by comparing this to known locations of IP addresses, assigned by Internet Service Providers (ISPs) as an approximate radius of the area. As a basis, Nosto uses commercially available geo-IP database. A shopper using a proxy or a VPN service is resolved to the location of the proxy or VPN, and respecting user's decision of masking their actual location using a privacy & security technology.

IP based geolocation is inherently and intentionally imprecise. It is not built to be used to locate shoppers or specific households, which would be the practical use case for mobile map and taxi apps making use of also GPS signal and WiFi-networks. Instead, the aim of Nosto's geo-location option is to provide an approximate, yet useful location of a shopper, which is at best a city location or a district in major cities such as Paddington in London or Staten Island in New York City.

Geo-targeting for on-site campaigns is based on segments, which means that shoppers from a specific are included or excluded from a group. A practical example could be including all US shoppers, but excluding shoppers in Hawaii, Alaska and Puerto Rico.

Weather targeting uses the same IP resolution and technical methods, but by automatically comparing shopper's approximate location to the nearest possible weather station and its weather data, which is commonly at least within 50km / 30mi radius and often closer.

In the segment creation locations are represented hierarchically:

Country / region, province or state / city, municipality or commune / district

Area higher in the hierarchy has higher likeliness to include all shoppers from the area, whereas smaller area is least accurate.

Due to the nature of how IP addresses work, how these are assigned by ISPs, and how the geo-resolving process works, some IP addresses (as shopper locations) are not necessarily precisely resolved to the given location. By default, the resolution accuracy proportionally depends on geographical location's size and Internet population in the area.

In practice, accuracy of geo-targeting is very good and closer to 100% at a country level such as United States or Sweden, and still very good at a state/province/region level such as California, Bavaria, Normandy or Skåne, and less accurate at a city or district level where possibly just about two thirds of the traffic can be resolved accordingly. City level accuracy is also dependent on the city and its size, or technically its Internet population. Shoppers in Los Angeles or New York City can be targeted at a higher degree of accuracy and resolve rates closer to a country or state level, while in small towns and areas where Internet population is smaller, targeting is less accurate. As an example, a shopper actually located in Palm Springs (CA) could be positioned in shouldering municipality of Palm Desert.

Keeping in mind these limitations, a good guideline is to create geo-segments for different purposes. Accurate small-area segments and campaigns are good, when the intent is to include a very valuable micro-audience, while exclusion rules for campaigns should be fairly broad. For example, at a country or region-level this could mean something such as: don't show a same day delivery campaign if shopper's location is at a geo-location where this is not possible: Hawaii (US example) or Northern Ireland & Jersey (UK examples)

Did this answer your question?