Initiatives like BCBS239 are a great starting point for making data governance operational and getting some core processes 'baked in' to the organisation.
In this post, I interview Taher Borsadwala, to learn more about his experience of implementing and operationalising BCBS239.
Dylan Jones: What is your background, how did you get started in the profession?
Taher Borsadwala: I started in the IT industry in 99 as a programmer and though I was dealing with data all the time, data was limited to a “back-end” task.
I have always been open to new challenges, whether it is disciplines, domains and/or technologies and that’s how I got my first break in the data management space, more than a decade back.
It was a Basel II compliance initiative comprising of customer data integration (master data management). Since I always make it a point to understand the ask/deliverable from functional/business/conceptual and technology standpoints, I started exploring data management from a discipline perspective. Positive factors in all of this is that data is omnipresent. We are always governing or managing data in some form-or-other in any and all projects. Given that overlap and considering my readings/exposure to the data governance/management space, I became end-to-end “data-aware”.
And that’s how & when the epiphany occurred: it’s all about data, everything else is peripheral.
And since then, my viewpoint remains data-centric: data is the least common denominator (LCD), the base building block.
I am thankful for having worked with, and having been mentored by, very experienced and practical people in this data space.
Dylan Jones: Let’s explore your experience of data governance ‘operationalisation’, an area I understand you have been focusing on.
What are some of the initiation tasks that the organisation must undertake?
Taher Borsadwala: I categorize Data Governance as a cultural change, a transformation initiative, if you will and so, awareness is primarily essential.
Data Governance programs most definitely need the vision, mission and strategy part chalked out clearly and that too, upfront.
Folks all across a company need to understand the reasoning behind Data Governance through the benefits it is going to bring. Besides defining the governance bodies, rolling out processes and finalizing the technology stack, teams need to be educated about the Data Governance roles that they are going to have to play.
Mainly from an operationalization perspective, they need to be made aware that Data Governance will make their lives simpler, their day to day efforts meaningful rather than becoming just another unwanted bureaucratic process.
As mentioned earlier, data is omnipresent and we govern and/or manage data in some capacity all through. Operationalisation provides the necessary scaffolding to get it to BAU mode.
In order to do this, once the organization has its Data Governance strategy defined, it should spend time and money on making its teams ‘Data Governance aware’, preferably through town-halls or luncheons.
Additionally, the organization should start with a couple of LOB’s or Subject Areas to roll out the operationalization that sustains success.
Second focus area has to be the measurement, especially of the data quality levels and that maturity that the firm has achieved through this initiative. Anything done in the absence of a target becomes just that – an “absent-minded” task. Nowadays, there are multiple maturity assessments available that a firm can leverage or create its own custom one to ensure that it is making progress.
Such measurements are done quarterly initially and then the frequency is revisited as the firm progresses. One point to note here is that an understanding of the “AS-IS” state has to be achieved before defining the targets/objectives.
Again, such “AS-IS” and “TO-BE” states need to be conveyed to the organization through the awareness sessions.
Dylan Jones: Can you an example of a recent Data Governance ‘town-hall’. Who did you invite? What was the format and agenda for the event? What do people learn and leave with etc.?
Taher Borsadwala: As highlighted earlier, awareness is of paramount importance. Town-halls have been in incremental order in regards to data governance/management content.
Recent ones have been specifically geared towards wins, highlighting how data governance has enabled regulatory compliance. More of a reverse-engineering flow from target report to sources and all people, processes and technologies involved.
Such events have been made available to all LOBs so that folks across all team and management layers are able to understand and appreciate the efforts, thereby applying their learnings in BAU mode.
Dylan Jones: Many companies are leveraging the need for compliance against regulations such as BCBS239 in an effort to kickstart their data governance efforts.
What does a high-level process look like for BCBS239 in terms of data governance?
Taher Borsadwala: I view BCBS239 more so as a data governance “enabler” as it has successfully and smartly put the focus on data governance/management rather than on the data itself in comparison to its Basel predecessors.
Simply put, Basel III allows working with any plain vanilla data governance/management process and then enhance the same per the regulatory requirements. Each requirement fits perfectly well into the various categories such as data quality, master data management, metadata management, data architecture, etc.
Since BCBS239 is tailored for data governance/management and since it enables measurable objectives/results, it’s a no-brainer towards kick-starting data governance efforts.
Dylan Jones: How should companies assemble the delivery team for something like a BCBS239 data governance implementation?
Taher Borsadwala: This question merits the classic consultant reply – “it depends”.
Answer to this question depends on multiple factors, mainly on the model that the organization has selected to move ahead with. Models available are centralized, decentralized and federated.
Since it’s not a one-size-fits-all, delivery team formation becomes very subjective.
From a generic viewpoint, below teams would be ideal:
- Data Governance Office
- Data Governance PMO Team
- Data Governance Product Team
- Data Governance Customer Team
- Data Governance Helpline (Support) Team
Data Governance Program ideally starts off in “change-the-business” mode and eventually transitions to “run-the-business” for BAU mode.
Dylan Jones: What kind of technology stack should companies consider for data governance implementations?
Taher Borsadwala: Given the plethora of tools and technologies that exist as of today, especially in the data space, finalizing a technology stack for governance feels overwhelming. Remember one thing, technology is simply a means to the end. An organization does not need the latest and the greatest for a successful data governance implementation. Spreadsheet programs can surprisingly work out well, if and only if, the people and processes part has been comprehensively defined and fairly but strictly rolled out.
Remaining vendor agnostic, the following would be needed in any data governance implementation stack:
- Taxonomy/Glossary/Data-Dictionary supporting tools
- Data Stores based on “polyglot persistence” principles that can cater to:
- Structured / Relational / Dimensional data
- Non-relational and non-dimensional data
- Cater to the 3 V’s of big data
- In-memory / Cache based storage
- Data Quality tools that can double up for cleansing, match n merge, validation functionality
- Visualization and Reporting tools
- Workflow engines
- Rule engines
- ETL / Data movement tools
- Integration tools
- Storage in terms of graph-model that facilitates linking as well as elegant visualization
- Exploring Lambda architecture for certain datasets
- Data integration and re-use, mainly through APIs
Dylan Jones: Finally, what is one parting piece of advice you can share with our audience?
Data Governance initiatives are not just projects but rather fall in the ‘disruptive’ program space. Be prepared with this mindset and expect challenges at all levels of your enterprise before delving into the same so that you can plan and build accordingly.
- Establish individual and collective belief that “data is an asset”
- Put data in the front-seat: apply the big data principle of making code work for data, rather than the outdated reverse version
About Taher Borsadwala
Taher has been responsible for delivery/implementation of BCBS 239 (Basel) com, SR-14 & other regulatory compliance programs through an exhaustive and integrated Data Governance and Data Management stack.