We recently sat down with Nader Anaizi, the Data Governance Strategy leader at Honeywell, to discuss his practical approach to data governance. As the leader of the data governance group at Honeywell, Nadar focuses on enabling the business to use data as a strategic asset for digital transformation. In his role, he does not do the data management or data governance himself, but rather Nadar helps the business use capabilities within Honeywell to get the most value out of their data.
Nader, first tell us about Honeywell. What is the data environment like?
Honeywell invents and commercializes technologies to serve several customer sectors such as, aerospace, energy, safety, security and productivity. As a diversified technology and manufacturing company, Honeywell blends physical and software products together to deliver the best experience for our customers.
As a large multinational conglomerate company, we have a lot of data throughout our organization such as, data surrounding our customers and products, material and vendor data, financial and reference data, and loT data. All of this data is dispersed across hundreds and thousands of systems like ERP, CRM, Marketing, HR, Finance, etc. and BI tools. These systems must process and view data consistently so that your data consumers can manage data as an enterprise asset. These data consumers or “customers” consist of the Data Analysts, Data Scientists, Data Stewards, Data Engineers and Data Architects.
That is a complex data environment. What are some of the challenges you face?
In this complex environment, there are several challenges that you may come across while implementing a data governance foundation within your enterprise, such as quality of data at source, reference data misalignment, inconsistent data governance practices, access to multiple data sources and limited self-service analysis, and lack of transparency of data lineage.
- Quality of data: Without data quality in place, you cannot guarantee that you are using the best data to make important business decisions.
- Reference data misalignment: Reference data will vary from system to system so you cannot manage your data effectively if everyone sees slight variations in the data.
- Inconsistent data governance practices: Most organizations do realize the importance of data governance so they try to create their own function within their organization. However, having inconsistent levels of governance between different organizations within a singular company means everyone sees and accesses data differently across the enterprise. This leads to misalignment and prevents digital transformation.
- Access to multiple data sources and limited self-service analysis: Self-service analytics has been a big buzzword for the past five to ten years, but it is not easy. It raises the question of, how do you give access to data across the company, while ensuring data is still used in a complaint manner?
- Lack of transparency of data lineage: If I am a data analyst and I am running a project to change something in the reference data, without data lineage I would not know where my data is coming from. This means that changes I make could affect a number of data processes and data customers down the line.
What is the vision of the data governance and management organization at Honeywell?
At Honeywell, the ultimate vision is to manage data as a strategic asset to enable digital transformation. We are not looking to do anything innovative with the data itself. We are really just looking to manage the data so it can be used for innovative projects and initiatives. There are four key things that help us achieve this goal:
- Single source of truth: One of the fundamental questions of data governance is, can we get this company to have a single source of truth where the whole company can see the data, understand what it means, see what reference data is being used, and who owns the data? When you try to do cross business initiatives, having a single source of truth is crucial to success.
- Governance: The next step is focused on establishing governance processes. First, we needed to assign decision rights; this is the crux of data governance. We must identify who owns the data and who has the rights to make decisions on that data. We also want to establish data policies or “rules of the game for data,” as well as the data quality rules.
- Operationalize: Third, it is crucial to establish structures, communities and domains that allow you to manage your data effectively. These structures essentially break up responsibilities based on who owns the data. You also need to establish governance workflows that allow people to define data. And, most importantly, you need to integrate data governance as part of the business and project process. You can’t just establish data governance; you need it to be part of everyone’s job and a crucial part of the culture.
- Simplification and automation: Step four is key to me. Just putting out tools and processes is not enough. You have to make it easy for data consumers to use the tools. This is called governance at scale. You also want to make sure you leverage the automation capabilities of your governance tool. You do not want to waste time doing everything manually.
Nader, you mentioned that data governance needs to be part of the culture and that it needs to make people’s lives easier. What role does Collibra play?
This is really important to me. It does not make sense to implement processes if they do not make people’s lives easier. Data analysts have full time jobs; they do not have time to spend countless hours learning how to use a new tool. That is why I love Collibra. Collibra is one of the more user-friendly tools on the market when it comes to data governance. Another strength of Collibra is the automated workflows. This allows users to spend more time using their data for innovation and less time manually creating rules and policies.
Thank you, Nader, for joining us and for your insightful responses.