Industry
CPG
- Enabled state of the art recommendation engines and a new customer engagement platform, without hindering customer shopping experience
- Responsive design helps in better understanding of user traffic (mobile v/s desktop)
- Speeding up the website increases organic traffic by an estimated 20%
- AI personalization saves time for both merchandisers and customers
- Scalability and reliability during flash sales
- Customer focused features to enhance brand loyalty
- Eliminated third party dependency on multiple branches, divisions that fuel the website content and framework and created a one-stop solution for development, integration & production of web enhancements
- Improved, cost-effective & remotely maintained enterprise data warehouse enabling usage of advanced analytics
- Revamped ETL architecture enables decommissioning of older code base and efficient rewriting for new codes
The client’s existing e-commerce platform resided on multiple vendor software and the Datawarehouse (DW) infrastructure was fuelled by de-centralized e-commerce systems made up of complex architecture, which was far from being cost-effective. The absence of a centralized framework resulted in multiple divisions and setups that powered the website. Without a robust pipeline in place for the instream data, turnaround time was high, maintenance costs were increasing, and there was heavy dependency on existing applications, external vendors and data sources. The ongoing ETL processes had to revamped to decommission older code base, enable efficient rewriting for new codes, and thereby align to latest marketing, personalization & user experience initiatives.
Given that data was spread over various data marts and that there was no central DW that could fuel a new framework, testing and deploying a website enhancement model was tedious. To ensure feasible synchronization between DW & the website, MathCo. helped migrate existing and historical data to Sales Force Commerce on Cloud, an e-commerce solution that had all the necessary sophistication for a contemporary, nouveau e-commerce portal. An Automated framework was built to transfer all the data from the existing DW to the SFCC framework on cloud, and create a centralized, one-stop e-commerce platform.
Unifying the ecommerce site for seamless operations and interactions speeds up the website and increases organic traffic by an estimated 20% on an average. Furthermore, switching to a responsive website increases likelihood of repeat visitors by a potential 74%. Here’s how the solution deployed by TheMathCompany enabled the customer to unlock their true potential:
Move to the latest ecommerce platform with no disruption to the customer shopping experience.
Our work accelerated initiatives that centralized customer touchpoints with one source of truth for multiple business functions, and accelerated time to decision making.
These initiatives triggered multiple new horizons in redesigning data science models that are vital for business, and have direct impact for in-store & online shopping experience and various marketing initiatives.
Here’s a detailed overview of the solutioning process:
Existing Web Framework:
The lack of a centralized sourced hindered the process of working towards various analytical use cases and migrating to a large data warehouse was challenging, as data was spread across many marts. Therefore, MathCo. enabled migration to SFCC, to create a backend warehouse optimized to store needed information for sales force commerce on cloud integration.
Migrating to SFCC:
Data was transmitted to SFCC and retrieved from SFCC via XML (Extensible Markup Language) files stored on an AWS S3 bucket using an agreed data transfer cadence for the newly built digital data warehouse. This enabled the centralization at DW, as well as the web layer, and the SFCC had the necessary sophistication to mobilize a new age e-commerce portal. The website traffic too was scaled gradually to the new platform ensuring zero downtime and outages.
Once the Central E-Commerce cloud was up and running, all the historical data had to be plugged in to the new platform, so that the platforms could sync with each other. Sub-streams such as third-party feeds of external data sources had to be converted into a format that can be read by SFCC. This step was a one-time integration effort, which once transferred to the SFCC cloud, would completely remove third-party dependency. Similarly, historical data concerning order, product, and inventory details, and data migration, had to be integrated to the new cloud. Once the existing content was integrated to the new system, these processes could cease to operate, given that the data was now available on the new platform. Clickstream data on the other hand, was an effort that not only required back up of historical data, but also required data input concerning new streams of live data. This reporting pipeline was and is continuously evolving and requires constant updation when being transferred onto the SFCC platform.
Therefore, the data migration took place through two methods – Full Load and Incremental Load.
The full load is a temporary, one-time loading process that is not a part of the automated workflow. It is necessary at the start of the migration process as it supports the incremental load process. In this step, Oracle databases (Staging and Destination) were emptied, Destination tables were loaded directly from source (application data). The process was performed using shell scripts & pl/SQL block
On the other hand, the incremental load is an automated ETL process, wherein the data was loaded from source to destination using a MERGE pl/SQL block, and the staging data was transformed using shell scripts. The data was then compared with existing data in the destination tables, and if the record was already present (as a part of full load/previous incremental load), the timestamp was updated. If the record was new, the needed content was inserted. The process was performed using AUTOMIC scheduler which triggers the ETL scripts based on a pre-defined cadence.
Advantages of the Newly Deployed Automated Framework:
Once the Automated Framework was deployed, the client was empowered with:
- A software-as-a-service (SaaS)-based, fully managed multi-tenant platform.
- Robust native merchandising features & Page builder solutions that give merchants more control over the design templates and content.
- 1:1 level personalization and machine learning applied across various areas via Commerce Insights/Einstein.
- Ability to manage multiple products, catalogs, inventory, orders, customers, online marketing, search settings, SEO, and complex product data.
- Efficient omnichannel features: order management, endless aisle, etc.
- Native features around internationalization (multi-currency, multi-language, and multi-store capabilities)
- Built-in A/B testing capabilities for merchandising that can efficiently take care of crucial aspects like merchandising logic, promotions, base sorting logic, etc.
