Why Data Governance Can Be Overwhelming

Lots of people who come on my training courses say that they are feeling overwhelmed by the sheer magnitude of what they've got themselves into, and are confused by where to start their Data Governance initiative or (if they have already started) what to do next.

To be honest, that’s exactly how I felt when I was first starting out in data governance. I used to compare it to juggling.  Many years ago, before I discovered data, the Bank I worked for sent me on a leadership course. On that course, they taught us to juggle.  Some people found it just clicked and they were amazing at it, but I found it incredibly difficult.

However, at the end of a five day residential course, I did manage to successfully juggle three balls (for a short period of time). Other people on the course managed more than three balls, but we each discovered our limit of how many balls we could keep in the air at any one time.

When I first started out in Data Governance, I  felt the same. There is so much that you have to think about when you're doing data governance, that it can feel just like you are juggling.  There are too many balls for us to keep in the air at any one time.  I came to the conclusion that you can't do everything at the same time. More importantly, I worked out that you shouldn't be doing them all at once anyway. That might be good news, but where lots of people struggle is knowing which activities you do need to do and in what order.  

When I was working on my methodology I noticed that you need to do certain things in the same order for your initiative to be successful. Having said that, there's also a number of other things that will vary depending on your organisation and exactly what you're trying to do.  

The juggling analogy has stayed with me because when I had a video made to promote my online training course last year, the juggling balls came to mind. You can see that video here if you're interested.  

My methodology takes you through everything you need to consider and do in the right order. But that doesn't mean that Data Governance isn't without its challenges. There are many challenges and the biggest one is the culture change you need to instigate. At the moment, most people in your organisation probably aren't thinking about data being part of their job at all. To be successful at data governance, you need everybody in your organisation to start thinking about the data they're creating or using, worrying about the quality of it and whether they should be using it for the purposes they are.

That's a big challenge on its own and there are lots of activities, communications and training that you're going to need to do to affect the culture change you require.

A second challenge I frequently see is that data is not a top priority. Perhaps you are lucky enough to work for an organisation that is focussing on “becoming a data-driven organisation” or is embarking on a digital transformation.  This sounds great as clearly your organisation is finally interested in data, right?  Sadly most people focus on the exciting outcomes and don’t understand that their data needs to be well understood and of good enough quality to facilitate these lofty ambitions.  If this is the situation you are facing, it will take a lot of effort to convince your stakeholders that they need to implement Data Governance first so that such initiatives can be successful.

You've got to make the case for data governance before you can even start designing and implementing a framework (as I've mentioned many times in the past, there is no such thing as a standard data governance framework!) 

When you do get approval to start, there is a lot of work to figure out a framework that will suit your organisation’s structure and culture and once you’ve done that you're going to need to do an awful lot of stakeholder management.   

You need to engage your stakeholders, identify and train Data Owners and Data Stewards. You are going to need a detailed communications and training plan.  I often say that you cannot do too much communication about data governance because you are trying to affect the culture change that I mentioned earlier. These communications need to be good quality, well written targeted communications and briefings. This in itself is a mammoth task.  

So having considered all of this it's no big surprise that a lot of people get overwhelmed with data governance and just don't even know where to start.

That's one of the reasons I started offering data governance training five years ago.

I went through the pain of my first few data governance initiatives before I worked out my methodology. I realised that I could help people avoid some of the pitfalls and the pain that I had gone through by sharing my methodology and that's exactly what my training course does. I take you through stage by stage what you need to do and in what order. I also share skills, tips and techniques to make you more successful. 

What is more all attendees of my training course (both in-person and online) get a copy of the actual checklist I use when implementing data governance for my clients.  So they really don't need to feel like they're juggling.

My next public course is in London next March and if you book before the end of January there is an early bird discount available.

If you want to make 2020 a great year for your Data Governance initiative, why not come along (and if you’ve got unused training budget, why not book before the year-end to make sure your budget doesn’t get forfeited).

If you have any questions about the course and whether it is right for you please feel free to schedule a call with me using the button below:

Schedule Appointment

Rupal Sumaria - Data Governance Interview

I am really pleased that one of my clients has kindly agreed to share what it is like being new into a Data Governance role. Rupal recently changed roles from Business Intelligence Support Manager to Data Governance and Analysis Manager at Penguin Random House UK. In her role as BI Support Manager, she saw first-hand how Data Governance impacts Data Quality and focused her efforts on fixing issues, improving the process and wherever possible educating users on data rules. In her new position, Rupal will now focus on embedding good data culture and sharing existing best practices in Data Management and Quality across Penguin Random House UK.

 How long have you been working in Data Governance?

 I’m very new to Data Governance, taking on the role officially only a month ago!

 How did you get into this area of work?

My boss Pete Williams, Director of Data and Online at Penguin Random House convinced me! He pitched Data Governance to me (warts and all) and I was sold. I’m still figuring out my role, and being able to set my own direction and path is very exciting.

What is driving your org to invest in this?

The publishing and media industries are undergoing huge changes as they respond to changing consumer behaviours, growth in our digital presence and new data-savvy competitors like Netflix and Amazon Prime as we compete for consumers’ leisure time.

In order to grow our business and ensure we are as data-savvy as our competitors, it is vital we maintain clean data and embed a framework that supports our future.

 Why is data governance important to you?

Data is created, stored and used in every aspect of the publishing business, but data concepts can feel really abstract in a highly creative industry. As a result, we need to empower and engage users to have honest and open discussions about data.

Part of data governance is to change the mindset that data quality, management or storage is solely a technology problem, as data issues affect decision making across the board. It’s exciting to lead conversations about Data Governance and hopefully make an impact on the business.

 What tips have you been given so far?

The top tips I have been given are:

1.    Create a strong business case that aligns to your business strategic goals so that Data Governance resonates with your senior leaders and they support the initiative.

2.    Don’t shy away from challenging areas that think they have perfect data.

3.    Don’t focus on what you call Data Governance; it’s the practice, people and process that are important.

4.    Don’t worry about who takes on what role, i.e. Data Owners, Stewards, etc. It is more important to have a process that allows for Data Quality Issue Resolution.

and were they useful?

 The advice is all very sound and simple but harder than it seems!

 1.    The business case proved to be the easiest of the tasks as we were very prepared and our senior leaders responded positively to the initiative.

2.    It can feel quite intimidating working with senior leaders but luckily Penguin Random House leaders are very friendly, understanding and patient.

3.    The Governance term felt very heavy-handed and as our business is very focused on the power of words, we changed the name to Data Management and Quality.

4.    This has been the most challenging, we very quickly got caught up on trying to assign roles without really taking stock of our Data Domains. We are now working to re-focus on the process, making sure to give everyone the support and resources they need for their roles.

 What tips would you give to someone at the same stage?

 Data Governance can be daunting. Talk to people that already work in Data Governance to seek advice and make sure you have a sponsor from your leadership team.

Also, don’t underestimate how much time you might spend organising people into workshops and arranging meetings to get Data Governance off the ground.

How Long Will My Data Governance Initiative Take?

In this blog, I want to answer a question that I am asked several times every week. To be honest, it’s not an unreasonable question, but it’s not an easy one to answer!

Before I go into any detail trying to answer the question, I want to make one thing very clear: there is no end date on Data Governance.

Data Governance should be something that you are implementing and embedding within your organisation, so that it becomes part of business as usual. For this reason, as anyone who has worked with me or attended my training courses will know, I make a point of impressing upon everyone that Data Governance is NOT a project. If you truly embed Data Governance into your organisation it should never end.

However, having said that, it is entirely possible that you may want to do a project (or project-like initiative) in order to design and implement a Data Governance Framework in the first place. So perhaps the question should be “how long will it take to design and implement a data governance framework and start delivering some benefits?”

But to be honest, that questions isn’t any easier to answer and you could say that both are “how long is a piece of string” questions. Last year, I was lucky enough to be on a panel debate at Data2020 in Stockholm with David Dadoun from Aldo and Andrew Joss from Informatica. Whenever I participate in a panel debate, I always start with a sense of trepidation as to whether my fellow panelists will have the same views as me or not. In this case I did not have to worry because both David and Andrew were very experienced in Data Governance and had seen many of the same challenges that I had over the years. This meant that we all agreed that there is no such thing as a standard Data Governance Framework or a standard approach to implement it. It also meant that— much to the frustration of the Chairman— we took it in turns to answer many of the questions with “it depends.” The panel debate was filmed and you can watch it here if you’re interested.

The reason I tell you this is that whenever I am asked this question, I am always tempted to respond with “it depends.” However, this would not be useful for the person asking the question, so instead, I have to follow up with some supplementary questions. These will include things like:

• Do you have an agreement to commence a Data governance initiative?

• How many resources have you got to work on the initiative?

• What is the scope of your initiative?

• How big is your organisation?

• How open to change is your organisation?

And depending on the answers to the above, I may well ask “is your organisation ready for Data Governance?” Please note this final question is not the same as “does your organisation need data governance?”

Back in 2014, the Data Governance guru Gwen Thomas (founder of the Data Governance Institute) wrote a fantastic article called “When You’re Not Ready for Data Governance.” I frequently direct people to have a look at this post to help get their head around whether now really is the right time for them to commence Data Governance, because sometimes you just have to accept that now is not the right time.

So having asked the first round of supplementary questions (detailed above), if I am convinced that an organisation is ready and able to commence designing and implementing Data Governance, then I need to answer further questions. These are around what they are aiming for and where they are starting from. To help answer these questions, a lot of companies turn to a data governance maturity assessment of some kind. These are very valuable tools in helping an organisation decide how mature they need to be, and in identifying where they currently are.

Please be aware that sometimes organisations can get tied up in “analysis paralysis” and spend inordinate amounts of time and effort on completing a maturity assessment. This is not useful, and care should be taken to only go to the level of detail needed to understand what capabilities your company is hoping to attain, plus identifying its current state.

There are multiple different maturity assessments available. As with all things Data Governance  I prefer a simple approach and you can download a very quick and easy Data Governance Health check questionnaire for free here. If a more detailed assessment suits the culture of your organisation better, I recommend you look at the freely available maturity assessment published by Stanford University. Sadly they recently removed their assessment from their website, but Alex Leigh has created an excel spreadsheet version that you can download from his website.

It is only after you have gone through the analysis outlined above that you will be in a position to estimate how long implementing Data Governance is going to take in your organisation. Now clearly the timescales are going to vary, but in my experience, it is going to take you the best part of a year (and probably longer) to design and implement a Data Governance Framework over at least some part of your data or organisation. This doesn’t mean that you won’t be able to deliver some quick wins during this period, but it will take a reasonable amount of time and effort before your Data Governance Framework starts to deliver value on a regular basis.

I don’t say this to put you off starting in the first place, but I have seen so many people underestimate the amount of effort and time that a Data Governance initiative takes, and it is vital that you manage your stakeholder’s expectations from the outset.

So whilst I can’t give you an easy answer that works for everyone, I hope I’ve given you some insight into how to work out the answer for yourself.

What is the impact of a poor data culture?

In this blog, I want to look at how a poor data culture can impact on staff in your organisation. Most of the articles I write focus on what you need to do to implement data governance, but I had such a great response to my post on why the data governance business case is so hard to get approved, that I thought it was worth delving a little bit more into the topic of poor data culture.  This could also be described as a lack of data literacy in your organisation.

All too often I come across organisations who have a very poor data culture. By that, I mean, that they don't really think about data at all. I see this improving all the time but awareness levels are still low.  Some industries (those where they make, move and sell things) I can understand that perhaps the focus is on the thing that is being made, moved or sold and less about the data around these processes.  However, for many service industries like financial services the products aren't tangible. The products don't exist in the real world, they are only data and yet many such organisations still suffer from a poor data culture.

It's absolutely vital that we get everyone in our organisations, whatever sector we work in, to start thinking about data, and the impact that poor data is having on our organisation. So in this blog I want to consider the impact a little more.

In my last blog, I advised you to look for your data quality issues. Maybe you found examples of where poor data quality has caused a loss.  Identifying a measurable cost to your organisation is fantastic, but if you found examples of poor data culture how are you going to measure that?  To be honest, it’s not something that often gets mentioned in a business case.

I can't tell you how many times over the years I've had people tell me that part of their job is to fix and cleanse data. I can think of one instance in particular when a student actuary was spending two weeks every quarter, cleansing and fixing a spreadsheet before it could be loaded into one of their complicated actuarial models.

I was aghast that he was wasting eight weeks every year fixing data in a spreadsheet. This person had been doing the role for 18 months and had been told that this cleansing and fixing the data in the spreadsheet was part of the process that had to be followed!  Actuaries are very intelligent and possess impressive analytical and statistical skills. Do you think it's good value for them to be removing duplicates from spreadsheets or reformatting data in spreadsheets? I certainly don't.  

This is just one example but I think it's fair to say that there are probably intelligent individuals doing monotonous routine tasks like this in most organisations  What impact is this having? You have a company not able to fully benefit from these skills and added to that these individuals are going to get disenchanted with the role and be less productive or even worse may look to move to another organisation.

Sadly I see this on so many occasions across all sectors and business areas.  Where individuals tell you that there's no point telling you about their data quality issues, because they've been there forever and nobody is ever going to fix them.   This defeatist attitude not only creates a poor data culture, but soon impacts the culture of the whole of organisation.

Think how much more engaged and efficient your staff would be if they didn't have to fix broken data or poor quality data before they could do their ‘real job’.

I've come across loads of similar examples over the years, but I was keen to see if other people’s experiences were similar.  I asked for input on LinkedIn earlier this month and had some great responses.

Obviously, this blog has to be a digestible length so I'm not able to include all of the examples but I wanted to share with you some of the impacts on culture that were disclosed.  Don't forget to keep reading to the end because I've saved the best/worst one for last!

Many examples raised the common issue of a culture of tactical or short term fixes that create data issues and build or reinforce data silos.  This means that organisations are then not able to take advantage of new technologies to use that data, some people shared examples of investment having been made in Artificial Intelligence or Machine Learning Tools to then find that the data wasn’t good enough to use them.

One example mentioned the care sector, one with a heavy dependency on people but which doesn’t take the time to train them in the importance of data.  This results in well intentioned people but poor quality and poorly managed data.  The management then can’t rely on that data so seek workarounds, perpetuating the poor data culture, increasing inefficiencies and increasing staff turnover.

And the final example is a sad but excellent example of what can happen if you neglect your data culture.  Someone shared that a 3 year regulatory reporting project involving approximately 100 people and significant investment in technology had failed because of a lack of data culture.  Data analysis and data quality had been de-scoped from the project and the end solution would not work because of poor data quality.  The person described the situation as like trying to make a chocolate cake without any chocolate – an excellent analogy.  This culture took its toll and the individual concerned ended up resigning taking their valuable skills to another organisation.

Please take these examples as a useful warning.  It does not have to be this way, if you get it right a good data culture will empower your organisation to see data as an asset and managing data as an asset will enable you to use data to focus on and deliver value.

One of the respondents, to my request for input to this article, summed up the situation nicely:

Leaders need to work to create an environment that is conducive to a behavioural shift and that is what a good data culture does. It is the foundation of successful change.

Helping you improve the data culture at your organisation is a key part of the Data Governance Journey that Alex Leigh and I have created together, you can find out more about that here.

Why is a Data Governance Business Case Hard to Get Approved?

When I'm running training courses, one of the early topics we cover in the day is the various challenges of implementing data governance, but to be quite honest, the challenges start long before you even start designing and implementing your data governance framework.

It can be a real struggle to get your data governance initiative approved in the first place. So I wanted to have a look at the reasons why this might be the case so that you can both plan for and mitigate them.

I think the challenges span four main themes:

·      data governance is rarely considered a top priority

·      it is hard to measure the value of data governance

·      it is hard to measure the success of your data governance initiative

·      your organisation may be successful in spite of its data

So let's have a little look at each of these in turn.

Data is not a top priority

This has been a common issue for many years, although the more recent focus on data and the drive for organisations to become “data-driven” has meant that this is getting slightly easier. But, be warned, it comes at a price as many senior executives hear about the “cool” data initiatives such as AI and Big Data Analytics and they want those without putting in place the data foundations (i.e. Data Governance) that enable such initiatives to be successful. 

 In order to overcome this, you need to sell Data Governance in terms of the outcomes it will deliver, and you need to tie those outcomes to the things that are a priority for your organisation (a good place to look for these is in your corporate strategy).

It's hard to measure the value of data governance

The problem is a lot of the benefits which Data Governance will deliver can't be measured in advance. They are intangible.  You could say that you will protect the company from things like reputational damage or investigation and censure by regulators, but your stakeholders could respond with something like: “We've never had data governance before and have not faced those particular issues”.

However, I think it's fair to say that every organisation I have helped implement data governance has achieved significant cost savings.  Most organisations experience many inefficiencies as a result of data not being available or accurate.  Alex Leigh (a fellow consultant I often work with) always says that these inefficiencies cause a lot of organisations to be “data fix factories”. 

Reducing inefficiencies is just one cause of an increase in profits after implementing Data Governance.  Many of my clients report that they've been able to better identify new opportunities or provide better customer service because the quality of their data is better.

These are good examples of where data governance has helped find and resolve issues.  The trouble is that you don’t find and resolve them until you have put Data Governance in place. At the time you are asking for your business case to be signed off, it feels a bit like you're gazing into a crystal ball. You don't know what the issues are that you are going to solve.

My advice in these circumstances is to go on a hunt for your data quality horror stories. Try to get some examples of real things that have gone wrong in your organisation and cost money.  Without having some real concrete examples, you are building a business case based on an unquantifiable value that may be delivered at some point in the future!

It is hard to measure the success of your data governance initiative

This is very similar to the point above because if your potential future benefits are unknown it is hard to agree on indicators to measure the success of your initiative.  You will undoubtedly make cost savings and increase profits, but if you don’t know where these will be, you also don’t know what to monitor at the time you are writing the business case.

Using the data quality horror stories mentioned in the previous section will help you articulate where you will be looking to measure success. Another area to consider is new systems. If your organisation will be implementing or designing a new system, you will not have to spend a significant amount of effort and analysis of data (as was likely the case). Data will be well documented and its quality understood in advance of the project. So try to get evidence of how much effort this has taken on previous projects.

Sadly there is no easy way to answer this is advance - you are in effect waiting for things to go wrong so that you can fix them.

Your organisation may be successful in spite of its data

Finally, an issue I've seen a number of times.  If your organisation is successful in spite of a lack of understanding and control of their data, it is hard for senior stakeholders to understand why they should invest in Data Governance, especially if there is no regulatory requirement for your industry to do so.

In this case, my advice is the same as point 2 – you need to find your data quality horror stories to provide evidence that poor data quality is having an impact on your organisation.

I don't want you to think that it's all doom and gloom. Creating a successful business case for data governance is possible, but I want you to go into it with your eyes open and aware of the challenges facing you. Having help from someone who has done it before, is a great way to make your business case more successful.

Alex Leigh and I have been working together for the past couple of years and because we have complementary skills (and like working together!) we have decided to formalise some joint product offerings. Helping you develop a successful business case is the first service that we are pleased to launch.  You can find out more about it here or get in contact if you would like to discuss it with us.

How to use a Lean Approach to Data Governance

Getting a data governance initiative started can be extremely challenging.  This is especially true if you work in an unstructured manner, trying to start too many tasks at the same time or doing things in the wrong order.

Over the years I have developed my own methodology for implementing Data Governance successfully.  This is based on my experience of what has worked successfully (as well as what hasn’t) over many years of implementing Data Governance.  Although I got into Data Governance by accident when I was a Project Manager, I had never particularly considered whether a lean approach could be applied.  So, when I was first asked whether a Lean approach would work for implementing Data Governance, I decided to look into it.

It didn’t take me long to realize that I have incorporated some of the lean principles into my approach without realizing it and that a lean approach would definitely be a good way to structure a Data Governance implementation.

To be successful, data governance needs to be implemented iteratively and efficiently and that is why applying lean principles to data governance works so well.

What is Lean?

Lean was originally created by Toyota to eliminate waste and inefficiency in its manufacturing operations. The process became so successful that it has been embraced in manufacturing sectors around the world and is now used in many different industries as it can improve how teams work together.

The goal of lean is to eliminate waste, i.e. the non-value-added components in any process. The idea being that until a process has gone through lean multiple times, it contains some element of waste. When done correctly, lean can create huge improvements in efficiency, cycle time, and productivity, which leads to lower costs and improved competitiveness.

The Lean Enterprise Institute (LEI), founded by James P. Womack and Daniel T. Jones in 1997, is considered the go-to resource for lean wisdom if you want to learn more about the details, but for this blog I want to focus on why it is a good approach for Data Governance.

Lean principles are all about the following initiatives:

• Empowering small teams

• Reducing cycle times

• Gradually eliminating waste

• Focusing on value

So let’s look at each of these in turn:

Empower Small Teams

Lean improvements start with people, and the same can be said for data governance. Applying lean principles allows you to focus on your team first.

Instead of creating a huge project team to implement data governance, setting up a small central team to support users across your whole business will have greater success.

Rather than being responsible for all data, a small data governance team can focus on:

• Identifying and maintaining existing data management activities

• Providing a framework for managing and aligning existing data management activities, and planning for future activities

• Coordinating the implementation of the data governance framework

• Acting as a liaison between the Business and IT to verify that business requirements are fully understood by IT and ensure the business is fully engaged in IT led projects

Reduce Cycle Times

Many data governance initiatives fail because they are too big in scope, cost, and timescales. Working on small phases or projects will more likely lead to success.  For example, try executing one process in a business area. When that is completed, implement that same process in the next business area.  Don’t try to achieve too much at once; if the data governance programme is too large and unstructured, the benefits will not be delivered efficiently and the entire programme might get stopped.  If you focus on small areas of scope, you are likely to achieve small but consistent successes in the implementation of your data governance framework.

Gradually Eliminate Waste

Implementing a data governance program through small frequent phases (or projects) allows you to use the lean problem solving approach: The Plan – Do – Check – Adjust (PDCA) Cycle.

For example, the initial plan step would review the business area you are intending to implement data governance in and take measures to fully understand the current situation.  What are the priorities and challenges in that area? This knowledge will enable you to plan the implementation for a data governance framework that will benefit that area.

In the do phase, you will implement the framework and identify and brief various stakeholders about their roles and responsibilities. During this phase, prepare stakeholders to start following one of the data governance processes or activities, such as defining data items for a data glossary, or using a data quality issue resolution.

Next you check the results of the do phase, confirm they align with expectations, and identify what can be learned from the experience. Determine if anything should be done differently next time.

Finally based on the insight gained, you can either adjust your approach before planning to implement the framework in a new business area or move back to the do stage to make changes in the first business area.

Focus on Value

Taking inspiration from George Orwell’s Animal Farm, we could say that all data is equal, but some data is more equal than others.  Lean uses prioritization techniques to focus on areas where you can gain the most value, and the same approach can be applied to managing data.  You can’t achieve value from managing all of your data with the same level of monitoring and control.  The highest level of monitoring and control should only be applied to the most critical data, needed to successfully run your business (I talked about this in more detail in this recent blog).  Applying lean principles and prioritising data governance activities for data that adds the highest value for the lowest effort will help engage stakeholders and demonstrate the benefits of data governance, while long-term activities (i.e. high value and high effort) progress.

Business engagement is absolutely vital to the success of all data governance activities, and underpinning these activities with the solid foundation of a data governance framework will help you achieve lasting data governance success.

Don’t try to do too much at one time. If your data governance initiative is unwieldy, it will be too big to get started and too slow to deliver benefits.  Applying a lean approach to data governance can help you work iteratively, checking and improving as you go and focusing your efforts on activities that will deliver the greatest value to your organization.

My Data Governance Checklist gives you a structured approach to design and implement a successful Data Governance Framework. You can download the free version of this checklist here.

Critical Data and Master Data - what is the difference?

In a recent blog, I looked at why you should prioritise your data and identify the data that is the most critical or material to your organisation, so you can focus your data governance efforts on that data. In response to that blog, I had a number of people contacting me to ask what was the difference between critical data and master data, so I decided that it would make a good topic for this blog.

As it is common for the two to get confused, let's look at each in turn:

My go to reference point for all definitions to do with data is the dictionary published by DAMA International (the Data Management Association). However, I was disappointed to find that there is no definition of critical data in there. Luckily with the help of Google, I managed to find the following definition:

“data that is critical to success in a specific business area”

In other words, it is the data required to get the vital processes completed.  And don’t forget that data critical to one business area may not be critical for another.  When this happens you have to manage the data to its highest level of criticality.   

That seems simple enough doesn’t it? So let's look at the definition of master data. This time, the DAMA Dictionary of Data Management comes up trumps and provides the definition as:

“The data that provides the context for business activity data in the form of common and abstract concepts that relate to the activity. It includes the details (definitions and identifiers) of internal and external objects involved in business transactions.”

Now, this is a wordy definition and while technically correct, it may not be helpful if you are not already familiar with the term master data.

Whenever I teach courses on Master Data Management I use this definition but like to highlight the key words in bold as follows:

“The data that provides the context for business activity data in the form of common and abstract concepts that relate to the activity. It includes the details (definitions and identifiers) of internal and external objects involved in business transactions.”

 For me the most important idea is that it is the data which provides the context for a business activity. For example if you considered a banking transaction, you have two sets of data.  The data about the transaction itself (how much was paid in, where and when) and you have the contextual data that enables you to identify the customer and the account:

From the image above you can see that the transactional data will change every time there is a transaction, but the master data would stay the same.

Master data is sometimes described as the most valuable data shared across an organization. If you consider the case of customer data (a data domain frequently “mastered”), it is common for customer data to exist on multiple systems across an organisation and unless you “master” the data – i.e. match and merge that data to create one record you cannot understand the extent of your relationship with a particular customer.  In order to match and merge customers you need to look for identifiers (as mentioned in the official definition) such as date of birth, address or post code. This enables you to create what is often called a “single customer view”.

Customer data is not the only data domain that you might want to consider mastering, product data, employee data and supplier data are all typical areas where master data is found.

 If critical data is something that is vital to your organisation, 

how is that different than master data?

The confusion between the two terms comes about because they are used for different purposes. Master data is nearly always critical to your business, but your critical data could include non-master data.

I do hope this has helped clarify the difference and if you are embarking on a master data project of any kind you may find this checklist useful as it details the activities you need to consider when implementing Data Governance as part of a Master Data project.

Data Owners and Data Stewards - What is the difference?

My last blog about how you identify your data owners stimulated a lot of interest, but also a lot of questions. One question in particular, I have been asked many times over the years (in fact, I got an email asking the very same question while I was actually drafting this blog) is the topic of this blog:

What is the difference between Data Owners and Data Stewards?

This topic does cause a lot of confusion. If you do some research online you will find many articles that discuss Data Ownership and Data Stewardship as well as Data Governance. This could easily lead you to believe that there are two or even three separate data management disciplines being discussed. To clarify the situation - Data Ownership and Data Stewardship are important components of Data Governance (although not the only components).

I believe quite strongly (and may have mentioned it once or twice before) that there is no such thing as a standard Data Governance framework. But I do believe that there are three key things you have to include in your Data Governance framework for it to be successful:

The three things as you can see from the image are policy, processes, and roles and responsibilities and they form a key part of my methodology.   It is the last category, roles and responsibilities, which covers both Data Owners and Data Stewards 

To understand the differences we should look at what each of these roles do. Let's start with the more senior of the two: Data Owners. If you've been following my blogs for any time, you will also know that they don't have to be called Data Owners (if you face resistance using this role title, you should call them an appropriate name that works for your organisation).  You can read more about this here.  But for this article we will stick with the more common role titles. 

Data Owners are senior stakeholders within your organisation who are accountable for the quality of one or more data sets. That sounds nice and simple, but covers activities such as making sure there are definitions in place, action is taken on data quality issues and Data Quality Reporting is in place. 

To be suitable to be a Data Owner, they have to be suitably senior in your organisation. They need to have the authority to make changes and also have either the budget or resources available to them to undertake data cleansing activities. If they don't have that authority and resources available, they won't make an effective Data Owner. 

Now, you may be reading that thinking, “if they're that senior, do they really understand the detail of the dataand do they have time to do all the things listed?”  That's a fair point and why I use the role of Data Stewards. I ask Data Owners to appoint one or more Data Stewards to assist them in their responsibilities. 

For many years, I wrote separate role descriptions, where I diligently listed everything that both the Data Owners and Data Stewards have to do.  To be honest the activities were largely the same, I just changed the language from saying “accountable for”in the Data Owner description to “responsible for”for Data Stewards.  

A few years ago I realised that there was a far simpler way: I now just write the detail for the Data Owner role and include words to indicate that a Data Owner may appoint one or more Data Stewards to assist them to undertake these responsibilities on a day to day basis. 

 How does it work in practice?

If you were talking about writing a data definition, you would say that a Data Owner is accountable for that definition. In practice, you would expect the Data Steward to be responsible for drafting that definition and presenting it to the Data Owner for them to approve.

Or if you were looking at a data quality issue, I would expect a Data Owner to be responsible for investigating and agreeing remedial actions. In practice, the Data Steward would do the research and propose appropriate remedial actions to the Data Owner to approve.

Another related question I am often asked is: 

Do you need both Data Owners and Data Stewards?

There is no standard answer to that question as it depends on the size of your organisation. For large organisations you probably do need both roles. If you don't have a lot of staff, you may not. 

I've worked with two organisations who both had approximately 200 staff. When we worked out who the most appropriate Data Owners would be and asked them to nominate their Data Stewards, we were close to half the employees of the organisation being either a Data Owner or Data Steward, which clearly is not useful.  The solution was different for each company:

In one organisation, we changed the level of seniority of the Data Owners to the next level down. They still had authority, but also had the time and expertise to understand the subject matter in more detail. In that company, the role of Data Steward was not used.

In the other organisation the right thing was to keep the Data Owners suitably senior (i.e. the Finance Director was the Data Owner of Finance Data), but instead of having multiple Data Stewards per Data Owner, each Data Owner nominated one Data Steward to act as deputy and help them with their Data Governance responsibilities.

You may not need both roles,  it depends on the size of your organisation. You need to work out whether you need both (and what you call them) to make data governance successful in your organisation.

To summarise, Data Owners and Data Steward are not the same role, but they are involved in the same activities. The Data Owner is accountable for the activities and the Data Steward is responsible for those activities on a day to day basis. 

Identifying appropriate roles and responsibilities is only one of many things on my data governance checklist. You can download the free version of this checklist to help you design and implement a data governance framework successfully here.