Industry:
E-Commerce
Service Provided:
Data team as a service
Problem Statement:
Providing near real time visibility regarding sales ROI
A prominent e-commerce company, engaged in end-to-end manufacturing of its online products through a dedicated application, faces a multitude of measurement challenges intrinsic to its diverse operations.
On the sales front, the company contends with substantial website traffic, necessitating the implementation of robust big data architecture and tools for user traffic monitoring. Internally, the company utilizes a billing system tailored to its specific needs.
From an expenditure standpoint, billing cycles with manufacturers are monthly, lacking mid-month tracking, leading the company to rely on estimates. The challenge further extends to working with various shipping vendors, each providing data through disparate methods and timelines. Additionally, the company is a significant advertiser across multiple marketing channels.
Implemented Solutions:
The architecture was implemented in an agile manner, while maintaining existing legacy solutions until we were able to transition working with the new data model completely.
The complete solution includes an entire data warehouse with processed and modeled data that is being refreshed once an hour. The data warehouse is being built and the batch data pipeline systematically collects data from diverse sources ranging from marketing channels, billing systems, production databases, factories, to shipping CRM and channels it into a data lake system. Within this pipeline, intricate business logic is applied to extensive datasets, involving predictions for operational expenditure (Opex) data, attribution, billing logic, and more. To accomplish this, the pipeline leverages a range of technologies, including Fivetran, Bigquery, Airflow, DBT, and Spark, orchestrating a seamless integration to deliver a sophisticated data model.
Business Impact:
Given the substantial daily investment in advertising, providing decision-makers with precise, near-real-time data is imperative. This data encompasses segmented sales figures across marketing channels and campaign levels. Moreover, it incorporates marketing expenses, and operational expenditure (Opex) data derived from predictive models, as actual data is only available on a weekly or monthly basis. The Opex prediction boasts an impressive accuracy rate of 99.5%, thus enabling decision makers to make informed decisions regarding marketing campaigns and operations. The data is used for financial reporting, ad hoc analysis on company performance and strategic planning as well.
Sales Growth:
While it is impossible to measure the direct impact of such a system on the sales figures, there is a direct correlation between the use of these reporting solutions and the ability to increase marketing spend in an effective manner, which results in additional sales. Since using the system the marketing budget has grew 50% and the overall ROI remained at the same level.