Sensor (IoT) based Customer Intelligence and
Marketing Automation Solution for a
US-Based Marketing Automation Client


This technology vendor aims to use data collected via wi-fi sensors and beacons on customer traffic and dwell times at retail locations to drive and deliver automated intelligent marketing campaigns. These campaigns are designed to help companies (Retailers, Restaurants etc) retain ‘at risk’ customers, increase customers’ visit frequency and loyalty and allow companies to measure ROI on their investments.

Millions of interactions are captured per day across each company’s client base, therefore requiring a highly resilient and scalable solution to be in place to deliver a robust campaign management platform. The platform has to integrate with third party tools and social media platforms to offer a seamless delivery experience across multiple channels.

The Solution:

An end-to-end Campaign Management Platform that utilizes UniConnect for Data Management, campaign scheduling, trigger management and generation of specific triggers.

UniConnect enables the tech vendor to read the customer traffic data in JSON format natively and enables in-memory joins with a MySQL database holding customer demographics. Leveraging other open source tools such as Quartz, which were integrated with UniConnect, we built the back end of the campaign management platform to deliver near real time triggers, thereby helping companies to engage with their customers and deliver relevant and timely offers.

The solution was conceived and delivered within 6 weeks, demonstrating the power of UniConnect to harness and process data from disparate sources in a short period, rather than months or years. The companies using the platform can now identify and proactively engage with customers when they are entering or exiting the outlet as well as identify if the visit happens to be one corresponding to a significant milestone such as 5th in a week, 25th visit, visit on a birthday etc.

As part of the solution, some reports and online processes require unified access to data from Cassandra and MySQL. Given that Cassandra does not support standard SQL ( it only supports CQL – Cassandra Query Language), a typical brute force and legacy solution would have been to write ETLs to unify data.

Instead, UniConnect was used to seamlessly unify both datasources through in-memory processing with standard SQL. All the campaign and trigger definitions translate to highly parameterized SQL statements that execute via Uniconnect on multiple data sources, getting data efficiently in memory from these sources and doing the required processing to identify various trigger conditions and associated actions.

In summary, a typical solution for the above problem would have involved extracting information from multiple sources into a standalone database of the campaign management platform leading to data duplication with associated costs and risks as well as latency in generation of triggers. Using UniConnect, all of these campaigns leverages the underlying raw data without causing an overload on the platform, a very good use case of the paradigm shift from ETL to ELT.

We are now extending the capability to include an ability to identify dwell times and movement within a store based on a store’s location map. This will enable retailers to identify marketing opportunities associated with such location based data within the store and take action instantly on the same.

Percipient Partners Pte. Ltd. #09-04, LATTICE80, 80 Robinson Road, Singapore 068898

© 2017, Percipient Partners Pte. Ltd., All Rights Reserved. Reproduction Prohibited.