Admit it. The analytic supply chain is broken. With limited resources, IT (the supply side) receive a massive demand from business while still maintaining the integrity of the data. Business analysts (the demand side) have lost their patience with IT bottlenecks but desire self-service, speed to insight and transparency.
It’s time to prove that IT and business analysts can work together to achieve true governed data discovery: the holy grail of self-service analytics. But how? Where do you start?
A Financial Services Case Study: Governed Data Discovery in Action
The information management team, from one of our large investment banking group customers, struggled to provide internal customers with quick turnaround on requests for data sets. Each request is an expensive custom programming job.
Their traditional process required a team of ETL developers to manually design an ETL model, implement and push data into a database, from where it was tested by a QA team and then placed into a sandbox for the original requester to access. The requester would need to understand how to access the data, which then is pulled into spreadsheets.
This process could take up to four weeks and require highly skilled resources from across various geographies.
When the banking group implemented Lavastorm, this time decreased to as little as four hours (even minutes) for some queries. With LAE, power users created a library of building blocks that went through the ETL process. Then an off-shore team consisting of less skilled “data clerks” would receive a request for data, drag and drop these building blocks, and build a model with limited configuration and no scripting on their part. The data clerks publish the results in a format the customer desires: in Tableau, Qlik, Tibco Spotfire, Microsoft Excel or SQL Server. The best part is that the model can be attached with any ticket (or data request) in case it needs to be re-run.
This method of agile data provision also spells cost efficiency, no longer costing the banking group $80-100 per hour to process complex data requests. Data clerks can do the work while still maintaining governance and control.
The original data requester (consumers of the data) can also see that visual model to best understand and validate where the data is coming from. All of this is possible without compromising the integrity of the enterprise data.
With a combination of technologies, there are many ways to enable governed data discovery and self-service analytics that keeps everyone happy and out of each other’s way.
What Effective Governance in a Self-Service Analytics Environment Looks Like:
- Producers create analytic models and “building blocks”: Tested, curated and governed
- Building blocks assembled and reused by consumers: Different roles with different access: View, Execute, Modify
- Enables true “Governed Data Discovery”: Provides more valuable, accurate data to guide results
In case you missed it, catch the full webinar we hosted along with Mindstream Analytics:
Investing in Agile Analytics Expansion
Learn more about how IT and Business users work together to drive ROI of their analytics data using Lavastorm.