Better decision making via data pipeline automation
Blog | September 23, 2022 | By Anish Bhola
Evolving customer preferences, ever changing fashion trends, rapid advancement in digitization, retail, e-commerce marketplaces, so on… have changed the ground rules of business. Companies can no longer bank on their heritage, brand reputation or customer loyalty to stay in the reckoning.
With the current pace of digitization, added pressure staying relevant and getting a share of millennial customers’ wallets is a challenge few can fathom, let alone win without rehashing operational efficiencies holistically. While managements are willing to walk the talk and invest top dollars to make this change happen, in most cases the bottlenecks are procedural and deeply ingrained within an organization’s DNA.
Decision making, or the lack of it is always a reflection of poor data management practices. While commonsense dictates automation is the panacea, the analytics team at one of our clients realized this is a bigger challenge to solve. Consultants from USEReady were sought out for their expertise in helping with such knotted issues.
The analyst community at our client was fluent with excel and the leadership didn’t want to migrate them over to the new tech stack without their buy in, however they were also cognizant of the issue at hand which couldn’t be solved with simple band-aid solutions.
Analysts work with files from multiple sources, each with vital information from a line of business or marketing channel. Data cleansing, transformations & file merge – the range of manual operations was simply mind boggling to grasp. This process was not just laborious, but also prone to errors and difficult to scale. Every time a new line of business or marketing channel had to be incorporated, the analysts would have to replicate the entire process with the new set of files. The process to update KPIs or redefine business logic was even more cumbersome and laborious.
With rapid progression of digitization, the file sizes only increased, and it was no longer possible for excel to handle this volume. Manual workarounds slowed the process, in turn the decision making suffered.
USEReady consultants were tasked with finding an efficient, scalable solution which is both forward looking but not disruptive that would put off the analyst team from adopting it.
With this one-line mandate, our consultants evolved a simple solution that automated significant parts of this process without altering the end user experience. All data sources, whether legacy/on-prem or API based were ingested into Snowflake, with this scalable storage was no longer an issue. Additional perks – it supports multiple workloads like data warehousing, data science, data engineering, data shares; is cloud native & near zero administrative overhead.
Matillion automated data sourcing, application of business rules for data cleansing or transformation. Data thus cleansed & transformed is exposed via ODBC connections within MS Excel.
Power Query & Power Pivot functionality within excel eliminated commonly prevalent issues surfacing in excel based reporting applications like data duplication or incorrect data (due to wrong clicks or mouse/track pad drag).
What we achieved:
- Process efficiency without sacrificing ease of implementation.
- reduction in time consumed, what would take almost 3 days to complete takes less than a few hours even on the busiest day with high volumes of data
- Effective management strategies – data quality, governance issues
- Quicker decision making
- Data analysts readily adapted to new processes.
- Forward looking scalable approach.