There's a great example of location-based targeting in AdAge, showing how P&G used it to promote a special Star Wars edition of its CoverGirl cosmetics, only sold in Ultra stores in the US.
P&G created special filters for photos that could only be used near the stores, and was able to track sales by missing out some stores so that they could test (potentially) exposed and unexposed shoppers. No results are published, but it's a great illustration of what can be done.
"Snapchat offered something else that made it attractive to P&G: geo-targeting. P&G was able to set up geofilters around a majority of the 868 Ulta stores across the country stocking the Star Wars line. When people went to post a photo or video to Snapchat within a certain proximity to the stores, they could place a branded overlay, or filter, atop the post so that anyone they shared it with through Snapchat would see the cosmetics line as well as CoverGirl's and Ulta's branding.
For the CoverGirl campaign, P&G received three statistics from Snapchat: 1) how many views its ads received 2) how many times people used the branded filter and 3) how many times people swiped over to see its branded filter when posting to Snapchat vs. the number of times people actually used it, which Mr. Rose described as akin to the ad's engagement rate.
In order to correlate those numbers with in-store sales, P&G had to isolate a few things. It limited its campaign to Snapchat only and didn't run those ads around every Ulta store in order to create a control group. And the campaign only featured two versions of the creative, one for each of the collection's two product lines. As a result, P&G was able to cross-reference sales between the stores it advertised and the ones it didn't and could drill down to see how many more sales of which of the collection's two product lines the geo-filtered stores generated, which could then be attributed to the Snapchat campaign."