September 7, 2021
- Posted by: admin
Part One of our data maturity model research demonstrated that organisations believe data has value, but they are yet to put a hard dollar figure on the work. Data Strategies are absent, and the executive stakeholders are understandably not engaged (why would you engage if all you see is cost?).
Part Two of our report is going to dive into how to go about improving data maturity. In order to do this, you’ll need funding. The funding will help you deliver on the promise set out in your data strategy.
Here’s a quick recap on what the Cognopia data maturity model is all about. Consider which of our 5 stages you’re at now, as this will help define the approach you take to improving data maturity.
- The Initial level describes organisations that have no formal data improvement program in place
- The Repeatable level is where most firms are today. At this level, organisations have defined data problems and are trying to fix them, but have not yet embedded these behaviours throughout their organisation.
- The Managed level describes firms that have a formal data management program in place. They have identified and funded programs for data improvement, but issues exist in compliance and execution.
- The Defined level occurs rarely, with less than 7% of our respondents achieving this score. These organisations have well managed, well-funded data improvement programs and are able to monetise their data assets.
- The Optimising level is the highest score, where data is managed as an enterprise asset and optimised across all business programs. No firms hit this top score, which is unsurprising given the difficulty of embedding “data-first” behaviours across large enterprises.
Read on to learn how you can improve your data maturity.
Chapter 1: Funding to improve data maturity
The first thing we need to improve data maturity is the budget. We’ll need support from key stakeholders in our business so we can make changes to the way we handle data.
Where do you get budget for data governance?
In order to improve data maturity, you must have a budget. Unfortunately, we bring bad news when it comes to data governance budgets. A whopping 33% of our respondents tell us they have no budget for data governance. A further 37% state that they use existing project budgets, for ad-hoc work if necessary. The companies that have no budget also have limited stakeholder engagement. Of those with no budget, 71% also reported having either no engagement or limited engagement with senior stakeholders.
What we discovered about data governance budgets
Very few companies have a dedicated budget set aside for data governance activities. Many of those that are running a data governance function are reliant on ad-hoc project budgets, or taking money from the IT team to spend on basic data cleansing activities. Companies with no budget assigned unsurprising scored very low on our assessment, with an average score of 2.1 (0.51 below average).
Budget does not exist
In the 33% of companies that had no budget:
- 57% of these firms only look at data in terms of the costs associated with it.
- 64% of these firms had no business case for data governance, and;
- The remaining 36% had basic business cases tied to ad hoc projects (e.g. improving marketing campaign effectiveness).
- 97% of the companies that lack a good business case for data governance also report;
- an unwillingness of the business to engage in data improvement activities, and;
- that the data governance team is isolated in IT or nonexistent.
Business cases save the day
Of the 14% of respondents that had allocated a dedicated data governance budget, the average assessment score was 3.69 – a massive 1.08 points above the average.
For firms that have a good business case for data governance, 38% report using it as a strategic enabler for business performance and operations – a 12.7x increase in business engagement and appreciation of data governance vs those without budgets.
Out of all respondents, we only found one organisation that had a good business case allocated for data governance yet still housed the function within I.T.
- Whilst the IT team plays a pivotal role in managing your data, it’s not a great idea to leave them to run the function without business engagement.
- The business teams should always be accountable for the meaning behind their data – after all it’s the business team that creates most data in the first place.
Key takeaway – build a proper business case, aligned against organisational strategic goals if you want to gain funding. This is integral to your ability to improving data maturity, and without it you will struggleNeil Burge, Cognopia CEO
Who owns the data in your organisation?
A data owner plays a crucial role in governing data. They ought to be the person with the most to lose if their data is incorrect. This role must be accountable for the quality of the data flowing through the organisation, and typically you need a senior business stakeholder (e.g. Head of Finance) to own data within their specific domain (e.g. Tax Codes). Get this role right and the rest of your data governance roles will fall into place too.
What we discovered about data ownership
The good: 7% of organisations have populated data ownership roles in the business for all their critical data elements.
The bad: In 16% of organisations, no one owns the data. In 25% of organisations, IT owns the data.
Unless a senior business person owns data, you will end up with errors. When no-one owns data, it’s chaos. If IT owns the data, how are they supposed to know whether records are right or wrong?Neil Burge, Cognopia CEO
Case Study: A major Australian Telco listed on the ASX
3.86 – “Managed” Level
At 3.86, the Telco was close enough to Level 4 to be classified at the Managed Level. The specific category score breakdown is listed below:
Case study background
This Australian Telco had been implementing data governance for a number of years. They scored inside Cognopia’s top ten in terms of data maturity in the past year, but issues still remained.
An aspirational start
The detailed analysis of their data maturity scores showed that they had maxed out aspirational scores and truly believed that data was an enterprise asset. As such, they had comprehensive risk management policies that were kept in line with changing regulations and sponsored by an executive at the highest level of the organisation.
The major challenges
The Telco had invested heavily in technology, using Informatica for data quality, Alex Solutions for their metadata management and Quartz for data modelling. Unfortunately, their experience with these solutions did not match the aspirations of the business. Data governance roles were populated largely by IT, meaning the maintenance of the data glossary was run on an ad-hoc basis. This led to limited agreement between teams on terminology and best practice, defeating the objective of a data glossary.
Losing their way
The team’s stated policy of data quality being the responsibility of all data citizens was not adhered to. This led to data being cleansed at the point of use. With IT running the data governance function the organisation failed to unlock the value in the policies and Technologies they had invested heavily in creating. Policies were not being policed.
NPS driving improvements
Cognopia’s Telco Pulse 2020 report analysed publicly available online reviews of Telco companies. We used these to identify data quality issues that directly contributed to 1-star negative reviews. In spite of their high data maturity score, this Telco still had 1-star reviews due to Account, Billing, Truck Roll, Customer Contact and Product data errors. Focusing to improve this data would drive up their Net Promoter Scores (NPS). The Telco states that increasing NPS is critical to their executive remuneration, meaning this focus will gain high traction from those with power, budget and influence to drive improvements.
When basic errors creep in around customer names, addresses or when you get their bills wrong, your customers leave in droves. This is a sure-fire way to grab the attention of your business peers!Neil Burge, Cognopia CEO
How could this Telco improve data maturity?
Firstly, they needed to recognise that their data maturity was not actually where they thought it was. Rather than being 3.86, the organisation had a maturity closer to 2.0 – just because they had written some rules about how data should be handled did not mean they were actually following those rules in practice.
This is a common challenge when improving data maturity. You might get the right people in the right role, and you might buy them the best technology to perform that role. But if you’re unable to change the behaviour of these staff then you’ll never succeed.
Their annual report made big claims about a desire to be customer-centric. They remunerate their executives on NPS scores. So there’s a clear opportunity to demonstrate how bad customer data is causing customer dissatisfaction. Bring together the right data to serve that customer at the right time, and they’ll be well on their way to improving data maturity. They will get stakeholder support and buy-in because they’ll have aligned their data ambitions with the overall goals and objectives of the company.
Chapter 3: Fixing what’s broken
Once we have the right executive sponsorship, we have to actually get on and deliver against the plan we create. This means getting the operational aspects of data governance rolled out properly. Let’s take a look at what we found.
How is data quality handled?
Unless you’ve got quality data, any investment in analytics or advanced AI is likely to be wasted. Various sources place the cost of poor data quality for the economy in the trillions of dollars a year. Cognopia’s own “Cost of Poor Data Quality” calculator has been used to find hundreds of millions of dollars in waste and re-work. Delivering high data quality is one of the most obvious, beneficial and critical outputs of a data governance program.
What we discovered about data quality
The good: Only 2% of our survey respondents believe that data quality is not relevant to their work.
The bad: 37% of respondents state that they know their data is poor quality but have no processes in place to fix it.
IT is over-burdened
28% of firms have IT responsible for data quality. This might make intuitive sense, because IT looks after the systems in which data is stored. A deeper look makes this less sensible:
- IT can help you fix the column level data – e.g.:
- What structure should our customer data table look like?
- What data type should be used to store this type of data?
- IT is unlikely to be able to help with the row level data – e.g.:
- I have 2 records for Mr. Robert Walsh, are they the same person?
- I have 17 different formats for Zip Codes, is that likely given we only ship to the continental United States?
- Is the phone number for our best customer accurate and working?
Shared responsibility is key
21% of businesses state that every person touching data is responsible for its quality and they are actively improving it. This is a fantastic result, and should be the ambition of any data governance program. Unfortunately, when we dig deeper this is not always implemented correctly:
- 33% of these respondents state that data is cleaned at the point of use by whoever is using it;
- If your data is cleansed at the point of use, your organisation is wasting money on re-work that does not need to be done
- 22% of the respondents say they have no process in place to document or track DQ issues;
- How can you fix the root cause of any data quality issue unless you have a mechanism to report it?
- How will you know when the data quality issue has been resolved?
To truly ensure that every person touching data maintains its quality, you first must ensure each person understands what that data will be used for. Build clear linkages between data creators and data consumers – and set expectations sensibly for both stakeholder groups. You cannot improve data maturity without improving data quality.
Have you defined your master data?
According to Scott Taylor, master data is “a common source of basic business data used across multiple systems, applications, and processes”. An example of this is a Customer. Unless you have a common definition of the data fields you need to serve your customers, your different systems and processes will fall out of sync. This can lead to some of the negative customer reviews we mentioned on page 26 of this report. Scary stuff.
What we discovered about Master data
The good: 35% of respondents have developed master data management definitions and are using them.
The bad: The other 65% lack a central, coherent view of master data across domains and systems. Running a business this way is fraught with risk.
So, does Master data matter? Well, yes. Check out the difference in data maturity between those with well managed master data vs those without:
Well managed Master data:
Badly managed Master data:
Master data matters!
In our assessments, those companies that had the most comprehensive master data management processes outscored the companies with the least comprehensive master data management processes by 0.87 points.
This should come as no surprise. Unless you have a good understanding of the key relationships your business has, you’re going to struggle.
Your business purpose is to deliver value to your relationships, through your brands, at scale.Scott Taylor, MetaMeta Consulting
Have you documented your metadata?
Metadata describes your data so others can make use of it. Think of column headings as an example. If I have a list of numbers: 9875 4686, 8988 6641, 8998 3443 … etc, it is essentially meaningless. If I put that in a column called “Phone_Number” you have a clue as to what the data might be. Documenting metadata ensures that the business can find and understand the data it owns, which allows you to then put that data to use.
What we discovered about Metadata
The good: 5% of respondents have documented their metadata across their entire enterprise – eliminating ambiguity and reducing risk.
The bad: 77% of organisations have either failed to document any metadata or have only done so in silos. Time will be wasted trying to understand data.
Data Glossaries are lacking
A data glossary is a repository of information about the data your organisation has. You can use this to add meaning to critical data elements – e.g. KPIs used in internal reports or submitted to regulators. The purpose of documenting metadata in a glossary is to ensure every stakeholder understands what the data is, how it can be used, and who is responsible for its upkeep.
- 65% of organisations have no data glossary whatsoever.
- 19% of firms have no one responsible for maintaining metadata in their organisation.
- 44% of organisations try to manage this on an ad-hoc or project basis, which leads to limited agreement on terminology.
Do you need a data glossary template?
One of the challenges in building a data glossary is often knowing where to start.
- What should I capture to describe my data?
- How should this be used?
- What format does it belong in?
We created a business glossary template that anyone can use to get started. If you’d like to grab a copy for yourself, follow the link below:
If you need help using it, please get in touch with the Cognopia team at [email protected]
Many “operational” activities within the data governance space are not being completed. Organisations tell us they struggle to get data owners and data stewards to participate in these key activities.Neil Burge, Cognopia CEO
How widely has Technology been adopted?
Technology is a huge enabler in most businesses. Data management technology is no exception. At the same time, many data professionals reach for a new tool as though it’s the silver bullet that will suddenly cause their senior executives to sit up and take notice. In our experience, this seldom happens. In this section, we look at 2 different technologies – data quality tools, and metadata management tools.
What we discovered about data quality technology
The good: 63% of respondents rely on in-house scripts for data quality. This is great when you have a limited budget, but not a long-term solution.
The bad: 14% of organisations have nothing to measure data quality. If you are not measuring data quality, you cannot manage it.
As you can see, companies are not heavily invested in data quality technology. This came as a surprise, as there are many high-quality tools on the market. Whilst we would typically not advocate investing in technology at the outset of a data improvement program, data profiling tools can help expose bad data to the business. Now let’s look at metadata management technology:
What we discovered about metadata technology
The good: 28% of respondents use manual systems like Word or Excel. Given the average maturity, this is a sensible place to start.
The bad: Just 2% of our replies stated “Enterprise-wide tools are deployed and are well used by the organisation”. The technology clearly has a hill to climb.
How you can use technology when improving data maturity
Rather than how the real question is “when should I invest in technology to improve my data maturity?”. And the answer is typically “later than you think”.
Most organisations rush to buy tools before they have the People and Processes set up properly. This leads to over-investment and often the failure of the project or program. Don’t do this.
We would typically recommend waiting until you score at least 3 out of 5 on our data maturity assessment before you spend money on technology.
Want a copy of the full report?
No problem, click here to get your 51-page PDF copy, including our full analysis, another full case study, and the recommendations you need to succeed!