Driving engagement, recommending the right product or pushing the best offers are some of the key challenges modern retail e-commerce platforms face. The typical concerns include — understanding the product mix; figuring out operational profitability; measuring performance for distributor, networks and operations; retaining customers; and improving the bottom line.

Vestige, a direct selling e-commerce site dealing in wellness products, is one such company looking to serve its distributors effectively, improve its customer engagement, and achieve the right level of impact on its revenue growth. The company was required to build a single version of the truth across its organisational data silos. To achieve this resilience, Vestige tied up with DAAS Labs (Data Science and Analytics Services Labs) — to integrate data and AI across all its business functions. 

To better understand the case, Analytics India Magazine spoke to Gaurav Shinh, the Founder and CEO of DAAS Labs. “As the fundamental problem was around data, we agreed to create an overall data strategy and roadmap which can help them to transform into a data-driven organisation,” said Gaurav.

DAAS Lab’s AI-powered holistic data platform has allowed the company to deliver data-driven decisions in a fraction of the time to enable business growth.

Also Read: Reimagining Interior Designing With Conversational AI: A Case Study

How DAAS Labs’ AI-powered Data Analytics Platform Helped?

Considering the demands of increasing customer engagement and recommending the best products to customers, DAAS Labs proposed a holistic AI-powered data platform — Scikiq. Scikiq has been designed to abstract all the data complexity from the business users via a no-code, drag and drop user interface for business to focus on driving value, “rather than worrying about how data is generated or curated,” added Gaurav. “This enabled them to grow, make faster, smarter and confident decisions.”

Built on Python-Django stack, Scikiq comes with pre-built data connectors and supports structured/unstructured data. Also, its advanced AI engine provides the firm with predictive analytics and enables it to build recommendation models, anomaly detection solutions and trend analytics across key business dimensions.

To initiate the process of data ingestion and discovery, DASS Labs decided to migrate Vestige’s data from operational databases and platforms to build a flexible and scalable data lake on AWS. “While migrating, we discovered the migrated dataset to enrich meta-data, identify quality rules, and build auditable and traceable data stores,” added Gaurav.

Additionally, the team implemented a data strategy to create a Vestige Data Platform on Scikiq, which now acts as a one-stop-shop for all data needs across the enterprise. “The Scikiq platform was deployed as a solution catering to all needs of the firm,” added Gaurav.

The solution deployment also involved the data capture mechanisms, curation, modelling, and visualisation, which are critically handled by the DAAS Labs’ team. To facilitate this, several BI and analytics cases were built on Scikiq’s inherent python framework and deployed within the product to be available for visualisation purposes as well.

Proposed architecture on Scikiq for Vestige

After deploying the data platform, the team delivered business value by creating reports and dashboards to provide end-to-end view and key insights on the site’s orders, sales, inventory, distributors, and promotions. This enabled Vestige to measure the performance of different functions and helped them in scenario planning and resource optimisation across functions. 

In addition to this, a number of business-specific proprietary models like sales forecasting, inventory optimisation, cohort analysis, customer segmentation, and product affinity, were built along with the product to serve the firm’s needs. These are further handled using proprietary models on Python fitted to the data available. Each model is subsequently retrained when more data points are available for minimising the error in predictions.

Scikiq also provides a natural language engine that assists in data storytelling, especially for the dashboards. A voice interface, built using Alexa, is embedded within the dashboards and trained on the metadata pertaining to each dashboard. This allows the voice interface to respond to specific queries while enabling navigation and filtering data via voice commands. “As a matter of fact, training of Alexa and other voice assistants is part of the capabilities within Scikiq,” added Gaurav.

“The entire product is cloud-based and can be deployed using significant DevOps expertise on any cloud platform, however, in this case, was AWS,” said Gaurav.

The Results

With Scikiq, Vestige has been able to reduce the Data-To-Action timeframe from months and years to merely days. Further, it has also helped them reduce their data migration effort by more than 50% through the no code UI based interface. The AI-based model, proposed by DAAS Labs, automatically discovers the relationships among the datasets, thereby saving at least 30% of the firm’s effort.

As per data, Vestige has also avoided nearly $8,000 per month costs in multiple tools that give point solutions for ETL, reporting, scheduling etc. plus integration cost of managing the complex data landscape. “The flexibility and the power to add new data sources and use cases in a fraction of the time helped Vestige’s business team to stay agile and ahead of its competitions,” said Gaurav.

“Unlike a large data programme that would take anywhere between 12-18 months to implement, the full setup with migration and reporting was business ready in just over a month,” concluded Gaurav.