Proximity Marketing – An IoT Based Approach for Improved Results
IoT has revolutionized the industry altogether, and this article will reveal the capabilities of IoT for Proximity Marketing.
Join the DZone community and get the full member experience.Join For Free
IoT has penetrated many domains and areas and has been successful. It has revolutionized the industry altogether, and this article will reveal the capabilities of IoT for Proximity Marketing. The reader would be able to connect and understand the problem statement and realize the importance of this innovation.
The Problem Statement
With the advent of technologies and smartphones, brick-and-mortar store owners have increasingly seen consumers migrate away from brick-and-mortar retail stores in favor of convenient digital outlets over the last few years. With the ease of smartphones and “one-click” shopping, many shoppers feel that browsing for products in a physical store is almost obsolete. On researching to go over the issue, it is found that BLE Beacons are a great way to guide customers their way through a large store and find the intended product. Also, for owners, it helps in targeted product recommendations. The proposed solution consists of Proximity sensing and a Product recommender system.
Proximity marketing targets valuable customers with targeted advertisements, backed by the proximity of consumers (or devices) to a specific section. This kindles them into making a purchase decision in the future. It powers communication with consumers at the correct place, at the proper time, and with targeted, personalized. Relevant messages on their mobile phones — by welcoming at the entry areas, special in-store offers at the store sections or receiving reviews for a newly released product.
When a customer comes into the proximity of a particular section in a store, the recommender system generates total and personal recommendations for the customer. Matching entries between recommended products and the specific section products is the final optimized recommendation delivered to the customer via mobile app.
The entire system built is based on BLE Beacons and has a Mobile app with HMS Capabilities and a Web application. The Mobile app will sense customers’ proximity using beacon data and deliver customer-specific recommendations. The web app will enable store owners to add/delete/modify product/section information.
- The use-cases that will be covered are,
- Inform Shoppers
- Boost in-store sales
- Interact and engage with shoppers
- Help visitors navigate through the store
Using BLE Beacons and Mobile App/Web technologies, product details and offers can be seamlessly integrated into the context of a customer’s proximity. The offer/product mappings are further triggered based on the consumer’s previous purchase history; the offer/product mappings are further triggered. This avoids an over-invasive shopping experience while correctly engaging consumers in proximity.
Food for the brain: Sensor selection is essential. What are the parameters one should look into while selecting sensors?
a. Proximity Sensing
As location service and Wi-Fi are used for the first level of fencing, the user’s proximity. Once the user enters a store, BLE Beacons are used to power product updates and overall customer experience within the store.
b. Product Recommendation
Product recommendation is achieved via :
- RFM Analysis — Signed-up users get personalized offers via their purchase preferences and history.
- Predictive Analysis — New users fit into the existing model using ML and store and offer details. The HMS kits would undoubtedly be a value add for the product’s functioning. However, the ML and awareness kits shall be the best pick.
Food for brain: What is the Cyber-Physical System? How is it different from IoT?
c. In-Store Mapping
In-store mapping opens room for,
- Map customers in-store currently — The triangulation sensing method is used to detect the exact spot of Bluetooth-enabled devices.
- Feedback to store about customer interests — What routes do customers take within a store. Helps in product placement and display.
The product has store-side, server-side, and user-side applications to be developed. The below section detail the same.
The shopkeeper uses a web application to manage the offers assigned to the shop's products. The shopkeeper will be able to modify the recommendations and descriptions of the products via this web application.
These changes will be reflected in real-time to the back-end database maintained for the shop.
When location services are not enabled but are connected to the internet through a mobile network or Wi-Fi, offers can still be provided to the user on a more general level.
Later, RFM analysis will still be done based on the user's purchase history and individual preferences. However, the offer list generated would be about the region of the user instead of a pinpoint accurate result.
Bluetooth beacons presented within each shop are used to generate the discounts being offered on the products within its vicinity. The offers are manually assigned to each Bluetooth beacon with the shopkeeper's help.
AWS EC2 Ubuntu Instance acts as the server with a Cassandra databaIn addition, flaskIn addition flask development server is used for APIs to interface with the database and provide recommendations.
The user flow is presented below as Fig. 1.1. Also, the store owner workflow is shown below in Fig. 1.2.
The technical architecture is presented below as Fig. 1.3. One can have a clear view of the entire system through the figure.
Results and Analysis
Once the buyer enters the mall, the Mobile application developed shall provide the details of the beacons available in the Mall. The names of the beacon can be the names of the stores. Here, in figure 1.4, the name of the beacon is revealed as mint. Since this is an experimental setup and proof of concept, we tested the same with one beacon. The gird view for a single shop is presented below as Fig. 1.4.
The offers recommended products shall be presented immediately to the user by clicking the Beacon icon. Fig 1.5 shows the recommendation list for a particular customer of a specific shop in-app. This includes recommendations and suggestions based on the previous purchases as well. The request is made through RFM analysis, and the same can be visualized from the software aspect in Fig. 1.5.
Post-2008, shops and stores started sending location-powered SMS texts to mobile phones that were in proximity to a Bluetooth network and had a discoverable mode for Bluetooth in their phones turned on. This system uses RFM analysis, a marketing technique that determines the customer's frequency and monetary details, which increases the sales of the store.
This strategy's future scope and challenges lie in tackling a few questions first, how First, how can Fitting be done for a completely new customer into the current model? Demographics-based prediction and recommendation can be a start to this. How to avoid invasive experiences? Churn Analysis is an answer directed at such problems. Finally, on the app-side quick querying and real-time expertise are expected. A combination of robust server-side and an advanced front-end is necessary.
There is a lot of room for growth and development in the same product. Adding machine learning concepts will make this application more versatile and more accurate.
Published at DZone with permission of Tushar Srivastava. See the original article here.
Opinions expressed by DZone contributors are their own.