A day in the life of a…..data governance consultant
Here’s another of our conversations in our ‘A day i the life of…’ series with the specialist associate consultants working with us here at Start with Data.
This interview discusses the importance of data governance as an ‘umbrella’ for covering company data and looks at the key components of a successful data governance framework.
Let’s kick off with a big question. What exactly is ‘data governance’?
There are a lot of definitions on the market, but in a nutshell, it’s an overarching process or function for managing data across the organisation to make sure they are usable, available, understood properly and ultimately, consistent. Data governance is like a big toolbox containing several compartments. Data quality, for instance, is a key part.
Yes, I assume data quality is at the heart of data governance…
It is one of the cornerstones. Reference and master data management are others. The whole data architecture, data modelling and of course the security of the data are also important elements. Recently, metadata management has also become very, let’s say, ‘popular’, as companies understand the importance of proper understanding of metadata from various sources…
Can you just explain exactly what metadata is?
Sure. Basically, it’s ‘data about the data’, which might sound like a bit of a fuzzy definition but let me explain! You have various actual data such as reference data, listing countries and currencies. You have your master data, which is information about clients, products and so on, and then you have transactional data. These data answer different questions.
For instance, a particular order has been placed today, so today’s date is transactional data, while the client’s name is part of master data. Metadata describes how your data is formed – all the elements upon which your actual data is built. For example, the customer’s name might be defined as ‘an alphabetic string with 50 characters’.
Who is involved in elaborating a data governance framework?
It’s usually defined at a very high level – key stakeholders come from C-suite level. They give the initial steer and then local teams work on it. I am currently working with a client with various stakeholders but the biggest role in defining the data governance framework is played by the CDO.
…but not every organisation is going to have a Chief Data Officer?
If you’d said that 5 years ago, I probably wouldn’t have been able to say what a CDO was! They’re much more commonplace nowadays, especially with larger clients. Everyone has a CDO and there are often CDOs for different functions in the company. Companies are really ‘getting’ how important this role is. In fact, as an illustration of how important this role is, last year, there was a notorious case where a major US bank was fined a massive amount. The regulator’s findings concluded that there was a severe lack of risk management and data governance.
That’s surprising that a large bank would demonstrate failure in those areas…
Well, especially in banking, proper data governance is key for issues like money laundering, anti-fraud, customer data security, risk management…
So, as a data governance specialist, what areas do you cover in your role?
That’s a big question! The first challenge is to simply understand who owns the data, because they can help you to define the data governance framework. As I mentioned, these people are often CDOs or other business stakeholders. But it’s common that organisations don’t have a clear understanding of what data governance is and then it’s really difficult to make people accountable for data governance. What do I mean? Firstly, that involves providing the definitions of data. So, for example, what does ‘the active product’ mean? Is it SKU which we are selling, have in stock or does this definition also include the products which were sold in the past, but due to the warranty or extended warranty must still be considered as ‘active products’? These definitions may be different in different parts of the organisation. A key challenge in the data governance programme is to ensure that the common, unambiguous catalog of the business data terms is defined, agreed and owned.
There must be someone who owns a definition. This is a part of the data governance function.
So, do you need to fact-find to determine the level of data maturity in the organisation?
Exactly. Again, depending on the industry, the amount of work involved can be enormous. Usually, you need to gain an understanding of the source systems. How does the data flow between systems? How does the data evolve and what’s its lineage? – put simply, where does it come from? If we look at the product information management – we need to understand the entry points for product data. The entry points include external feeds from the suppliers, manual entries from the product managers. Entered data is then transformed based on the business rules, validated, reviewed, approved by various stakeholders and published on the web or the other systems in the organisation. Proper data governance helps in understanding the lineage of product information – it helps track the data elements from its origin, through PIM and all consuming applications and channels.
When you carry out the initial stages of this ‘investigation’, do you get pushback from those organisations who aren’t very well-prepared?
Very often, yes. In some cases, it is mandatory, but in others, business users don’t really care if data is replicated in other systems. A kind of attitude which says ‘I know my product portfolio’. This way of ‘siloed’ thinking is a path to failure. Data needs to be properly understood and shared across the organization. An unambiguous definition of the product attributes helps in building the holistic view of the product information and helps linking other datasets, for example: all product related complaints raised by customers or sales team may help in developing improvements in the product itself to increase the customer satisfaction, reduce the return’s rate, etc. Without proper understanding of the product data journey, building data insights is far harder. It’s still the case that awareness of its importance isn’t great, but a proper data governance may help the organisation in multiple ways including cost reduction, increased customer satisfaction, (better) product development.
To what extent is the robustness or otherwise of data governance connected to company culture? Siloed teams, making up their own rules, and so on…?
To quite a big extent, in the sense that data governance cannot easily be translated to money value. For instance, if I’m a decision-maker, I’d prefer to spend my budget on building a sophisticated data analytics platform to try and understand how my customers behave and to enhance my value proposition to them – cross-sell, upsell, and so on. They’d rather spend on that than investing in a tool which helps them understand how the data flows through the organisation. For people in the sales department, for example…
…they don’t see the tangible benefits…?
So, let’s say, you present your initial findings to the stakeholders and move onto the implementation phase. How does that happen?
The positive thing is that there are many tools on the market which help you to manage your business glossaries and link them with the technical metadata pulled from source systems. Another good thing is that governance processes are largely standardised. So, in a nutshell, how you build your business glossary and link with the metadata is pretty much the same everywhere.
A basic question perhaps, but what is a ‘business glossary’?
It means the set of business terms defining your data entities. So, again, what is a ‘customer’? What does ‘active customer’ mean? What is a ‘product’? An ‘SKU’?…What does a ‘customer name’ mean? How should it be treated? So, it’s a way of establishing some rules for quality standards for data entities. Ultimately, it’s like a dictionary for terms used across the organisation.
Does that take the form of a document that people can refer to?
Yes, normally, but typically, there are some data cataloging tools which can help people to organise these business terms, to ensure there is a genuine owner associated with them, and to connect the business terms with the technical metadata from the source system. So, if we look at the customer definition, it can be associated with various systems – let’s say, Salesforce, your PIM, your billing system – so you have full visibility about the use of the term for ‘the customer’ across the entire organisation.
Ok. That’s been a very useful overview of what data governance is and how exactly it works. Thanks a lot for your collaboration!
It’s a pleasure.
Start with Data consultancy services
It’s essential that your systems stay up and running and fine-tuned for maximum performance and to minimise business disruption. Our technology consultants are experts at managing these core systems and can handle day-to-day operations. We will deliver ongoing training of PIM system owners, supporting business and technical users.