Retailers are in a tough situation. With a lot of consumers shifting preference towards online shopping, retails can’t be complacent. We think investing in these 4 trends for big data in retail will help these retailers thrive in tomorrow’s marketplace.
Increasing Integration of IoT and Mobile Device Data
According to the 2017 Retail Vision Study, 70% of retailers are ready to adopt IoT technology. While some companies like IKEA have introduced Augmented Reality in their catalogs, some stores have gone even bigger. Smart mirrors, self-checkout scanners and tracking beacons can be used to understand how customers feel about store inventory. For example, frequently price-checked items that were not checked out could point to a price being mislabeled.
Some stores have even started introducing interactive fitting rooms with smart mirrors. These can automatically recognize products through tags or barcodes. When the customer wishes to try on a different style or size, they can request it from the mirror by pushing a button. This is especially helpful for ‘introverts’, who would rather click a button than talk to a (potentially pushy) store employee.
Some of these mirrors can also send 360-degree views to friends, so that the customer can make an informed, ‘crowd-sourced’ decision.
Granted, integrating these sensors to your retail analytics platform sounds daunting specifically because of the sheer amount of data that can be collected from these devices. However, if you have a modern big data database, you should have no trouble loading many terabytes of data per hour.
Big Data in Retail Shift from Predictive Analytics to Prescriptive Analytics
While some retailers still haven’t adopted predictive analytics, some are advancing further into what’s called explanatory or prescriptive analytics. While predictive analytics lets the business make pretty good future predictions based on large data sets of existing data, it is only a reliable metric if nothing changes. Because we know that everything changes all the time, that can be very partial indeed.
Prescriptive analytics is a way of saying “How can we make it happen?”, instead of just “What is likely to happen?”.
By adopting a more explanatory analytic modeling, retailers can correlate weather patterns with news and customer buying patterns to help predict which products will be popular during the rainy season, and how the prices of good will change. This allows stores to stock up on relevant supplies and change prices accordingly – thus “making it happen”. Proactive instead of reactive.
Outside of retail, a good use of prescriptive analytics and explanatory modelling is in healthcare. It is able to identify the relationship between the price of cigarettes in a specific region and the relation to cancer. Contrast this with predictive analytics, which will only be able to tell you which hospital sees more cancer patients in specific regions.
Integrate More Location Data
This isn’t actually a “new” trend, it’s been going on for years. If you haven’t gotten location analytics yet, you’re missing out on very valuable information. It’s not just for retail either. Last year, SQream helped a telecom integrate location data analytics to get a holistic view of their network. SQream let’s the telecom analyze the way customers behave. This opened up a variety of new income streams. The telecom operator can now offer products to customers based on their “daily grinder” locations.
Already, 65% of companies across all industries have adopted location analytics. In retail, the adoption rate is even higher. Just like with the IoT revolution, companies spend millions outfitting stores with beacons and smart sensors. A single customer may generate tens of thousands of data points per visit in a large store like Macy’s, Nordstrom or Kroger. This information could be used by the retailer to optimize product and promotion placement.
GPU Databases are especially well-suited to location data analytics. The high rate of data and relatively simple data structure created by these are much better suited for the GPU.
Well-established companies like ESRI, Foursquare, Cisco are joined by (relatively) newer companies like Euclid Analytics, in making the most out of location based data, to track everything from how many users enter stores, how long they stay, and the number of times they return to each store. It’s the web analytics revolution brought down to brick-and-mortar stores.
Converging AI with data analytics
While AI in retail isn’t as enticing as, say “AI in the autonomous car”, there’s still a lot going on. Retail is actually a field that has already seen quite a bit of AI integration. “Smart home” AI helpers like Google Home, Amazon Echo and Apple’s Siri are already affecting the way people shop both online and physically.
With computer vision, companies like eBay have already started integrating AI to help identify miscategorized products. For example, after uploading a photo of a skirt, eBay’s AI algorithms understand what cut the skirt is. The AI algorithm also understands if it’s sleeveless, formal, or even really just a miscategorized pair of jeans.
On the chatbot-front, both traditional and online shops Starbucks, Staples, Pizza Hut and others have implement chatbots to help customers purchase. Virtual agents can now start building a personal relationship, even using different “accents” to match their customer’s style. Understanding cultural aspects of the costumer leads to integrating recommended products or services based on the customer profile.
We started by saying retailers are in a tough spot. And it’s true. The sector is incredibly innovative. To continue being innovative, retails must keep investing in new technologies, like machine learning, GPU databases and prescriptive analytics. The number of companies offering services to retail are increasing exponentially every 10-12 months. In 2018, we’re likely to see more competition and a more dynamic retail environment. More vendors find new and clever ways to customize and improve the personalized experience.
SQream is here to help with your next big data investment.