Why Tracking #bikinis, #babies, and #versace Mentions Can Save Retail

Going beyond census data to inform retail location decisions.

At Spatial we analyze social media to reveal what types of people are in the surrounding area. This goes a step beyond census data to quantify the personality of the surrounding community. Something we all know is important but left neglected because it is hard to quantify.

For a city to thrive, it needs to have flourishing retail. For retailers, Spatial applies our 500 social segments to reveal sites where communities naturally gravitate to your brand.

idaho, feminine

urban, fashion

Sometimes, the results are surprising. Such as when we used social data to help a bank underwrite loans to business owners. We found places likely to rise in real estate prices had a volume of social media mentioning fermentation + fedora + upcycling + brisket  + cabbage.

Most recently, we did this same process using public data for a struggling retailer (Payless).

Quantifying human personality

By analyzing social media data alone we were able to predict 82% of Payless store closings. There were a few topics mathematically proven to spell life or death for Payless stores. Two of them were:

1. Presence of feminine interests (#cutestbabies, #prom dresses, #bikinis)

2. The absence of expensive tastes (#versace #laceup)

feminine

urban fashion


The takeaways are obvious, but until now we have never been able to quantify them. For Payless, areas that attract feminine interests are good. Areas where people express interest in status symbols, like Versace or Dolce & Gabbana, are bad.

Most important, the personalities of people surrounding your store have a bottom-line impact on success.

Ultimately, this is another line in the equation for optimizing retail success. And by extension, our cities. If you’d like to add this new layer of understanding to your toolbox, start by downloading the sample data package or simply contact us.

Lyden Foust
CEO / Spatial.ai

Want to see the data for yourself?

Download Sample Data