Location Analytics of Brand Sentiment by “Listening” to Social Network Chatter

By Joe Francica

Directions Magazine (DM): Can you elaborate on the sources of the “listening” channels through which you are gathering location-based social network data? Twitter? Foursquare? How are you aggregating the feeds? For example, are you taking the Twitter firehose and sampling the data?

Neil Crist (NC): Venuelabs has agreements with over 120 digital channels that pull together location-based "signals.” Signals include any type of customer activity a customer can complete by location (i.e. check-in, share, tip, status update, like, comment, photo, tag, etc.)  There is a foundational aggregation problem we solve for our customers, in that for a 250-location brand, they will have anywhere from 4000 to 9000 location-based sources to track. Our true special sauce and IP is in the aggregation mechanics, analysis, normalization and surfacing of insights from across these channels. Specifically for Twitter, only a very small percentage is truly location-based (estimate from Gnip is about 2-3%) so we do not use the firehose https://dev.twitter.com/docs/streaming-apis/streams/public).

DM: Where are the POS data coming from? Are you taking transaction data or simply social net data?

NC: On top of the local channels I talk about above, we also take in proprietary data for customers that want to look at their own metrics overlaid on top of the location listening metrics.

DM: How quickly can you aggregate the “voices of the consumer” and provide feedback to the client? For example, would you produce a daily trend map of customer sentiment by store location? At what scale are you working? Street? Zip? City?

NC: All sources vary in frequency - daily is the most common use case. Brands can see all of these metrics (sentiment included) and view them by brand-wide, by geography, region, by location. Brands also can upload their own proprietary location metadata to include in the analytics, such as designated market area (DMA), custom territory, regional managers, form factor, etc. Our analytics and views scale from brand-wide down to a single location.

DM: Explain how you score the results that are provided to the client. Are you taking both social media and Census demographics to arrive at a marketing measurement score? Does this employ a regression analysis of both recently collected data and Census data?

NC: We have score content by sentiment and also by category (i.e. staff, facilities, price, products), which is executed by a human team that we manage internally. These scores are available real-time to brands via our Web analytics dashboard.

DM: How are you using photos to score brands? Since photos are generally considered unstructured data, how are you collecting the brand identity from photos? Are you scanning and extracting brand images?

NC: Yes, our human sentiment team is scoring photos tagged with location. Photos can be marked for sentiment and categories, and often have text that further contextualizes the experience. Humans are required to make that a quality analysis.

DM: What does success look like for the clients? Are they looking to alter their marketing mix, their marketing message, or their merchandising? Or all three?

NC: Customers measure ROI in three ways today:

  1. Customer Satisfaction. Venuelabs customers see a rise in in-store customer happiness of about 130% on average over the first 90 days of using our platform.
  2. Local Engagement and Community Growth. When brands use Venuelabs to understand, target and engage content to local pages, those brands are seeing meaningful growth in those communities and their engagement. Currently in the first 90 days of local publishing initiatives, community growth is growing on average by 66%. Community is equivalent to customers that have opted in for marketing.
  3. Cost Savings. Brands are also seeing very real cost savings from not having to expend human resources to manually manage their local presence across these channels. Assuming a social media analyst is required at a very basic hourly rate, 20 location brands are seeing > $1500 per month in savings. A 500 location brand would spend over $30,000 / month manually managing these channels.

DM: How does indoor location tracking play into your scoring and how might it impact your analyses in the future?

NC: Currently we do not do that granular of tracking, but we have partners in the space that we keep in close touch with as that technology progresses. That would be an entirely new level of in-store optimization.

DM: What kind of education do you need to do with the client to help them understand the potential of location-based information?

NC: Our selling process is education-based. We are helping brands understand what location-based information looks like for their brand (through sampling 50-100 locations) and also delineating it from the "noise" of social media listening.  The context of location is powerful and provides actionability that social listening cannot.

Additionally, we also relate how much "local" insight is being missed even when they are spending tens of thousands of dollars on platforms like Radian6, Sysomos, Lithium and the like. Those platforms are based on keyword, keyphrase listening, but they miss location. We call this the Blind Spot. (link)

As an example, this morning when I stopped at Starbucks, the line was out the door and the barista was not very friendly. I checked in on Foursquare and mentioned both of those issues, but I never used the word "Starbucks.” Location was the key context (implied), not brand mentions. (link)

In 2013 we are also using industry benchmarks as marketing to inspire the industry discussion and also showcase the insights we are able to surface. (link)


Published Monday, August 5th, 2013

Written by Joe Francica


Published in

Location Intelligence


If you liked this article subscribe to our newsletter...stay informed on the latest geospatial technology

© 2016 Directions Media. All Rights Reserved.