3 Steps to Start Governing Data Successfully

In the past three years, we at Cognopia have seen a dramatic increase in the interest in Data Governance across the core Asian markets we work in. What was previously a discipline focused purely on regulatory compliance in Banks (e.g. BCBS 239) has evolved to be seen as an essential function across industries and public entities alike.

Whilst it’s great to see the increase in interest and adoption of Data Governance practices in our core markets, it’s also quite common that the prospective clients we meet are a little unsure how to go about adopting Data Governance “best practices” and getting engagement from their management stakeholders to make this a success. It’s not uncommon for us to be asked “How do we start with Data Governance?”, and this article is designed to help.

How do We Start Data Governance? – Start with the Basics

Start with the basics for Data Governance

Given we represent a range of tools that help increase productivity of Data Governance teams it’s not surprising that a lot of our conversations revolve around how to automate the tasks and activities these newly minted Data Stewards will need to engage in. In my opinion it is a mistake to focus on the tools you need before your team are clear on the use-cases, outcomes and objectives the business has for adopting Data Governance practices in the first place.

It’s usually worthwhile documenting the reasons why you need to Govern your data and prioritising those according to the value this data will bring to your business. “Boil the Ocean” or “Big Bang” projects have never succeeded and there’s no reason to believe they will now. From the conversations and clients Cognopia has, it’s clear the best-placed Data Governance teams have usually settled on some basic use-cases that will deliver a quick win to their company.

If you’re asking “How do we start with Data Governance”, maybe look at some great examples that we typically see:

  • Regulatory pressure – typically when a company needs to report data externally and to prove where the data came from and how it was calculated and transformed into the external report.
    • Many organisations have costly internal or external teams manually tracing this and want to cut these operational costs.
    • Across Asia we are increasingly seeing fines for some industry sectors (e.g. telcos, banking) if they cannot verify and validate the Identity of their customer base.
Personally Identifiable Information (PII) needs to be governed with more care.
  • Data Privacy Legislation – over 80 different jurisdictions have enacted data privacy legislations, and fines are increasing for Data misuse – GDPR fines can be up to 4% of an organisation’s revenue or 20 million euros. Whilst the regulators in Asia are not yet issuing such large fines, the bad press and corporate embarrassment of being caught in breach are enough to focus the minds of the executives we speak with.
    • Organisations need to profile their data sources and repositories to identify all data elements that contain Personally Identifiable Information (PII), then build out Data Lineages to see how this data flows and changes through the IT landscape
    • It’s no use having a strict internal policy on the use of PII data unless and until you have a map showing where this data resides and who has access to it currently.
  • Technical Debt – organisations that have lots of legacy IT or who have grown by acquisition have often inadvertently ballooned their IT maintenance costs by housing multiple versions of the same data.
    • These organisations want Data Lineage to help understand which data is really required to switch off redundant IT systems and make it easier for teams to find relevant data
  • Report Data Anomalies – different divisions operate in siloes, each reporting up to the CEO based on their own data or from their own skew on Enterprise data. Typically this results in “data brawls” between management all trying to present their own version of the truth.
    • Building a common Data Glossary and Catalogue helps these teams ensure they’re all speaking the same language, and it further benefits the Data Scientists creating the reports as they waste less valuable time finding trustworthy data in the first place
  • Data Quality Issues – our clients are normally heavily focused on customer related data in their first project phases, as this drives corporate revenue and better-quality customer data helps you understand and interact with your customers in more meaningful ways.
    • Up to 23% of revenue is wasted with poor quality customer contact data alone – and it leads to other operational inefficiencies such as poor customer segmentation, an inability to deliver orders, and time wasted by call centres chasing customers on mobiles that are no longer active

Whatever the reason you are considering adopting a Data Governance framework, be sure to document this thoroughly, get senior management engagement, and determine what the success criteria will be when it comes to measuring your Return on Investment.

How do We Start Data Governance? – Start Small

Start Data Governance in a small, measurable manner

It may seem counter intuitive for a company like Cognopia, as a reseller of multiple leading Data Governance and Quality solutions, to recommend starting a Data Governance initiative without purchasing technology. The advanced technological solutions we work with are going to accelerate and automate your Data Governance programme, but they’re not a silver bullet by themselves. When starting out you need to establish the value of the work quickly and to do so it is often easier to use existing tools like Excel, email, wiki pages etc to establish the first phases.

Starting small is also applicable to the use case. Let’s say you’ve decided you need Big Data Governance – your team are about to adopt a Data Lake and you want to ensure the Data in the lake is well governed rather than degenerating into a Data Swamp. To do so on all the data within the lake would be wasteful, time consuming, and is unlikely to deliver value before the project is cancelled – instead you should focus on the core data within the lake that will drive better business outcomes. Some approaches to defining which data to start with:

Which data is used to drive major business decisions?

1.) Often known as Critical Data Elements, or CDE’s, these data elements are typically those used as key metrics in your internal and external reports.

2.) Trace these back from report to source and govern the meaning, quality, access and usage rules to drive trust in the reports issued

Which data drives the greatest value?

1.) This could be customer data – where better definitions and understanding of customers will lead to revenue growth or higher customer retention rates

2.) HR data is another core data area for the larger enterprises – 39% of employees would work harder and increase their productivity if they are happy

Which business units are most progressive in working with Data?

1.) Data Governance should not be a task solely for the IT function. Engaging the business users that create, work with, and rely upon the data involved in a governance initiative will substantially increase the results you see.

2.) We often see marketing teams under the CMO heavily invested in Data Governance. They know the modern marketing landscape is driven by data and the ability to segment their prospects and customers will drive engagement. Delivering a Single Customer View is often their Holy Grail.

3.) The Head of Legal or Compliance will often sponsor an engagement to maintain and protect the PII data your organisation captures.

Which Data is of the lowest quality?

1.) Some businesses are acutely aware of the impact low quality data has – either because they cannot create the reports they need, or because the reports they have created are so clearly misleading

2.) Other organisations suspect they have dirty data hiding in their landscape and elect to run a Data Profiling exercise to surface the issue and understand the impact.

3.) Even basic Data Quality interrogation using SQL scripts can highlight major issues that will deliver a substantial ROI if they’re resolved successfully

How do We Start Data Governance? – Start with the Right Team

Get the right Data Governance Team or you will fail.

As mentioned above this shouldn’t be an exercise purely for the IT team, unless you’re only concerned with technical debt or reducing the number of irate queries about the source/quality of your report data. To embed Data Governance across the Enterprise you need to succeed in your first foray – having identified the business case and scoped the work to an achievable outcome, you now need to bring on your “A Team” within the organisation to ensure you succeed.

Find those that already value their data.

One reason to find a use-case within a Business Unit that values data is that you’ll be more likely to find a team who are responsible, empowered, enthusiastic and able to enact the changes you need to succeed. Top-level management engagement from the head of department (or a C-level stakeholder) can remove obstacles and ensure you get the time commitment required to adequately document and catalogue the data you need to govern.

Find your internal expert.

After identifying the stakeholders that will drive your programme through, you then need to find Data Stewards (to corral and facilitate the process), Subject Matter Experts (to help build and democratise the understanding of the data in their domain), and Data Owners (to sign off on policies, definitions and rules – at the top of the escalation chain). Typically there’s pushback when first asking business users to engage, as this is “more work on top of their day job”.

When resources are tight you need to get creative with your Data Stewardship and Ownership models, and the only way you’ll succeed is to designate the Subject Matter Expert as the nominated person to take on this work. Typically the SME is keen, will not procrastinate on decisions, are usually known as the data owner anyway, and has been with the company for a while.  On the other hand, when the steward allocated has none of the attributes above, they typically fail or do nothing.

Empowered Teams Achieve More

A small, agile, empowered team can achieve dramatic results in a short period of time. Clearly communicate the goals and objectives and set a timeline for the activities to complete. Your aim here isn’t to govern the data into oblivion – you need to show the value in this work, iron out any kinks in the model, and fail fast (if you’ve picked the wrong team).

Assuming your team was well selected and has delivered useful work back to the business you can expand the scope or pick a new use case. This is usually the right time to talk about automation and tool selection, as a reward for the success and to help expedite the roll out to the rest of the business. You’ll have a solid business case for investment and can use the results achieved to get other senior stakeholders engaged when it comes to governing the data in their domain.

How to start Data Governance – Talk to an Expert!

If you need guidance or advice on the right tools to help your Data Governance initiative, please let us know. We’d be happy to share our experience and provide templates, resources or workshops that will help you achieve your goals quickly and with the minimum amount of fuss.