Hello 😀
When it comes to Data Governance, the one million dollar question is still “what’s the value? what do you deliver?”
Imagine having a store : you want people to come in !
Let’s see how to stop the bullshit with awesome services👇
Agenda
Data Assessment
Data Skills Accelerator
Data Domains Builder
Data Assessment
All eyes are on AI at the moment. It is going to bring more productivity, more creativity, more revenues, more revenues.
👉 These AI projects need good data. Even GenAI, don’t get me started on this !
To get this good data, data scientists are spending 80% of their time on data cleaning and data preparation. Sometimes after 3 months the team realizes that they won’t be able to deliver because it is simply impossible to build a model with crappy data. What a waste !
Plus, junior profiles tend to believe that machine learning can solve everything.
😈 They will wait until the last second to reveal that they can’t build a suitable model.
Here’s your first customer as Data Governance : you will serve the Data Team.
What you need
A Data Governance Manager
A Data Quality Analyst
What you’ll offer
The Data Assessment is a framing phase to estimate the Value of each AI project before launch. To do it, you will evaluate both Business scope and Data scope :
Business scope :
Clear understanding of business scope : Which question is at stake? What alternatives exist to answer it? What are the drivers : costs reduction, productivity, new revenues?
Are there business experts available to bring their knowledge and validate model results?
Data scope :
Is the data available? At the required granularity and depth?
Is the data retrievable? Manually or automated? In the right databases?
Is the data clean, consistent and ready to manipulate?
Setup a presentation showing the consolidated Data Scorecard and Business Scope, leading to conclusions on the potential Value.
💎 If the value is high enough, great ! The project can start.
❌ If not, you’ll support the Data Team towards data-readiness by crafting solutions to close the gap on data : it can be through data collection or remediation plan on quality.
With this service, you’ll become a prerequisite for AI projects.
You’ll also keep an eye on data protection for each project.
🔍 Tip #1 : You can extend this service to the lifecycle of AI projects by providing long-term solutions to monitor data quality.
Data Skills Accelerator
Usually HR & Communication departments have the objective to diffuse a data-driven culture. They will organize awareness sessions and workshops.
👉 But we all know that’s far from being enough to truly change how people use data.
Most of Business team members are lost when it comes to data. They end up doing everything in Excel.
Or there are Business Analysts doing all the job of refining data and preparing dashboards for everyone. If this person leaves it’s a nightmare.
This is sad. We have many sources of data but we just don’t use them.
Here’s your second customer as Data Governance : you will serve the Business Teams.
What you need
A Data Governance Manager
A Data Governance Expert
What you’ll offer
You have to increase autonomy when it comes to data.
This service will help design the right roles, responsibilities, and skills for successful data governance within your organization.
You will provide :
Guides defining roles, objectives & councils :
Roles : Who should be doing what? How to identify Data Stewards? How much time should be allocated to these roles for their data governance tasks?
Objectives : Which tasks should each role do? Which target for each task (# of validated definitions, resolution time of quality issues…)?
Councils : Which purpose and outputs? Which frequency? Which attendees?
🧐 Needless to say that to fill these guides you need buy-in from Sponsors and Business executives. And you have to prepare objectives of each role with team managers.
In the long term, meetings with HR will be necessary to define incentives and career path for these roles.
Regular trainings “Data 101” including :
Data governance roles : we are all data producers and consumers ! Who is doing what? Who can help you with data quality issues?
But also specific trainings for identified Data Owners and Data Stewards.
Data basic tools : Where can you find available datasets? How to manipulate data with PowerBI / Tableau / etc? How to prompt ChatGPT properly?
🔍 Tip #2 : Don’t hesitate to make these trainings appealing, with catchy titles like “learn how to use AI in your work” or “leverage data for business insights”.
Data Domains Builder
Disclaimer : this service is for Federated data governance model only !
Most of Business teams won't be “doing AI”. And that’s fine.
They just need to use data properly. Which means : having a correct customer database to contact them when they need to, knowing how to name new products, being able to look for last months sales metrics, etc.
👉 They might have understood the roles, but they are still missing standards, processes and tools to execute tasks.
There is a need for a clear methodology to structure “Data Domains” to help Business teams to better leverage data.
Here’s your third customer as Data Governance : you will serve the Business Executives who know that data is a growth driver for their domain.
What you need
A Data Domain Sponsor, aka Business Executive
A Data Domain team member, who represents the “data pain points”
A Data Domain Manager (can be the Data Governance Manager from central team)
A Data Analyst
What you’ll offer
A “Data Domain” is a cross functional team that delivers best in class data to meet the requirements of priority business questions.
This service provides a structured approach to defining and implementing data domains, ensuring alignment with business objectives.
You will provide :
A list of critical data elements for each domain : a data element is a basic unit of information akin to a building block of data, such as a person's name, a product price, or a date.
The critical elements are essential for business operations and taking decisions, they are linked to the use cases identified by the Data Domain Sponsor.
You can define criticality factors according to other aspects such as regulation, security risks, etc.
Data elements are sorted in categories called data families and at the top level, you have the overarching domain, which represents a broad category of data.
All critical data elements must be governed properly.
This means that each data element should have a Data Steward attached.
This means that each data element should have clear standards attached : data quality levels, format, unit, refresh frequency, linked use cases, etc.
A Data Quality toolbox :
Data quality assessment checklist for assessing the quality of data systematically
Data quality dashboard templates to monitor and report on the quality of each domain data, such as data completeness, accuracy, and timeliness.
Data validation tools - usually custom scripts - for validating data against defined standards, criteria and business rules.
Data enrichment tools that provide additional context or information to existing data, such as demographic data or geospatial data from providers like Acxiom or Experian.
Issue tracking system to log, track, and manage data quality issues and their resolutions like JIRA or just a Sharepoint List.
Your value here is to offer pragmatic tools to start governing data of each domain.
At first, your team will be the one doing the job, logging tickets, setting up dashboards.
🚀 But in the long term, it should be the job of the Data Stewards identified in each domain, thanks to your Data Skills Accelerator service !
Why it’s great
Clarifying your value proposition as Data Governance team will have a huge impact.
Once people know what you deliver, they'll come to you and you won't have to keep looking for “Data Governance use cases”.
See you soon,
Charlotte
I'm Charlotte Ledoux, freelance in Data & AI Governance.
You can follow me on Linkedin !