Data & AI Governance Expert 🔎 | Data Governance coaching | Women in AI | Speaker | LinkedIn™️ Top Voice in AI 🇫🇷
I don’t understand why companies hire Data Stewards. It doesn’t make sense. ➡ Data Stewards are already here as non-official roles. Do you think business people are using dashboards with incorrect numbers and ignoring the issue? ❌ No! They find ways to fix it, whether it’s cleaning data in Excel or asking someone they know from tech to correct it. In doing so, they establish data quality rules from a business perspective. Isn't this exactly what you expect from a Data Steward? #datagovernance #organization #data
Charlotte Ledoux while this is true, I often see the problem that business people simply don't have the time to take up a "second job" as a data steward again. Also they often need support in fixing the data, because they don't have the tools to do it on a greater scale. What is your approach to this?
True, real Data Stewards are already there 💪. 👉 But in modern workplaces, teams are stretched thin. Adding new tasks means prioritizing them over BaU. If Data Governance isn't a priority, hiring specialized resources can help, allowing SMEs to focus on business-critical tasks. 🤞 Alternatively, we know that the better solution would be to convince the business that prioritizing internal data governance and quality activities adds more value than hiring externally. #awareness #datastrategy
Data Stewards are not hired. In the NIDG approach, people are “recognized” as stewards based on their relationship to the data (as definers, producers, and users) and are held formally accountable based on those relationships. This is not something that people can opt into or out of. Potentially stewards can be everybody. Imho
I usually find Technical Data Stewards to be people in the business who have been involved in defining business and quality rules for their own systems such as ERP or CRM, DWH or single source of truth (SSOT). They know what they need for basic daily operations, and business data stewards for defined business metrics and performance KPIs that they verify daily and monthly from data residing in SSOT. The challenge is to convince them to document everything, maintain the metrics and the data catalog, that's a data culture thing and we need to guide them in the beginning and hopefully continue on their own. I have been working on DG initiatives in fintechs and in the banking industry, it is not easy because business teams are stressed every day and need support.
Correct : you should not hire new Stewards, you should identify your current (even if not official) data stewards, train them on the overall framework, and hire people needed to give them back the time to do their job properly.
Agree but how do you do when you don't have the desire to do it in the team and not enough data awareness?
I'm going to disagree with you on some of the points you've made. While I completely embrace NIDG (if it's not broken, don't fix it), there are many limitations to having business-centric people make corrections in excel. First is the idea of efficiency, reuse and discoverability - individuals working only on their desktops can't easily share or maintain data cleaning rules, transformations and categorization with others. You also may be hurting productivity by having multiple people performing the same validations (and then having to review their discrepancies together). Last point is that your model may work well in a stable, established environment - but if you're faced with onboarding new data sources (with the associated ingest validations) on a regular basis, having a technically oriented steward can help smooth that transition by leveraging rules across tools and platforms, to reduce redundancy and improve consistency. As an old school DBA, my suggestion is to take advantage of the capabilities built into higher-order data management tools. That will streamline your processes and improve your overall maturity. And you'll sleep easier.
Data stewards are often needed to support data owners, stakeholders and data trust, depending upon organization size, data maturity level, and data governance framework. Karen as the Accounting Manager may be a data owner but due to her endless list of financial tasks every day/week/month, she may need support and assistance to keep up with her data owner responsibilities. Data champions should never be voluntold they have these new responsibilities, especially without some type of shift or reduction in their current duties. And establishing data quality rules from a business perspective is great, but where are they documented, prioritized, observed and enforced? A data observability tool, data contracts, test automation? Karen doesn't have capacity to manage any of those. Most business people that see dashboards with incorrect numbers may not use them, but they may also not speak up about it either as they find alternative means to come up with the data and analytics they need, creating multiple versions of the truth. It's a double edged sword for sure.
🌱💡 Data Governance & Data Management | DAMA Norway | Podcast host MetaDAMA | CDMP Master
3moVery much agreed! The Data Steward role is the connector between Data and Domain knowledge. And that is what should be embraced. Yet there are some problems I see when trying to enable your data Stewards: 1. The role is often too specifically focused on singular data issues, which leads to a service center mentality for the specific domain. 2. Data Stewards are implemented as part time roles, on top of existing responsibilities. 3. the role is given to the "Data Nerds" in the domain, where "hard skills" and data expertise are valued more than communication skills, storytelling, etc. Btw: I don't really like the name Data Steward.. i started calling them Data Ambassadors. But maybe there is an even better name?