Masterclass AMA Session | Showcasing the business value of data governance use cases

It depends on the use case we implement. So in case if we are talking about a use case focused on Issue Management then of course, it will be # of the critical incidents. But what else we can measure:

• # of data issues created in period (day, month, …)
• # of data issues per data domain, organisational domain
• # of data issues per data source
• avg. time of the data issue resolution
• # of assets with data issues

On top of that, we would like to check what will be the impact on the business from fixing data issues faster and it could be:

1. Increased quality of the business decisions with clear, trustworthy data
2. Reduced IT cost via minimised errors and eliminated rework 
3. Less time to resolve data-related issues via streamlined issue-resolution processes
4. Reduce the number of service desk licenses
5. Increased productivity: Support Engineer (20%) > this one can be used as a basis for the annual capacity saving calculation

Usually the calculation of savings or cost avoidance come from measuring the impact of reoccurring issues (e.g. low data quality leads to weekly manual data cleaning of 2 hours), and then how the DG project or action will prevent this in the future (e.g. higher data quality + monitoring & alerting leads to avoiding those 2 hours per week). Or in short, the value would be how much preventing the problem is neutralizing reoccurring mitigation. 

# of critical incidents worked on is a valid directly measurable metric, and based on that ideally you can calculate a financial impact based on cost of mitigation, frequency, and additional risks avoided.

Let's say your organisation would like to increase revenue next year, and one of the initiatives for the year will be - increase efficiency of the marketing campaigns. If this is the business goal, we can think about data use case that will address next need "Get access to customer data insights for tailored marketing campaigns and product offerings". And here some examples of the data challenges that we can address as part of the data use case implementaion:

• Customer data is in different formats and spread across different sources
• Inconsistencies and errors in customer data impact marketing campaigns
• Difficulty in locating and accessing (compliantly) the right customer data
• Challenges in validating data for AI models used for market segmentation
  and personalised product offerings

As a result of delivering the solution that will address these issues we can expect the next business value:

• Improved Customer Satisfaction
• Customer base growth
• Increased sales in a core product
• Increased Productivity: Product Manager (18%)
• Increased productivity: Data Analyst (13%)

@Community_Alex​ Dmytro already mentioned great examples here, and I would only add that aligning business value with a data use case can be a co-creational effort with business stakeholders - they would know best what kind of decisions or actions need to be informed or driven by data, and you can together define the right metrics that can quantify those, and start measuring and monitoring those. 

Value measurement is an ongoing process (see below the process flow). But there will be definitely some milestones when you would like to check the improvement rate and compare "Before" and "After" states. For instance, you could check "How long does it now take to find the right data set / data product for analysis? (hours/days)" before Data Marketplace implementation and after. Then you could calculate the time savings and multiply it by the number of the data product searches a year. It will be a good demonstration of the saved time with Data Marketplace.

@Community_Alex​ Additionally to the great answer from Dmytro I would say that the frequency of reviewing value per use case should be synchronized to business review processes. There are often weekly, monthly or quarterly business review processes of business functions, synchronizing the cadence into a cross-functional process is not only less disruptive, it's also a great chance to communicate data and business in a more closely connected way.

@Community_Alex​ I love this quote "People don't resist change - they resist being changed." The reason people don't like change is usually because the efforts needed to change is higher than the perceived benefit of the change. Sometimes convincing others to change needs someone to role model it for them because the benefit is too theoretical - and that might be the reason in your case. Try to live the change in a smaller, feasible scale, ideally with enthusiastic collaborators, and then spread the word of the self-experienced success story - it might help a little with convincing your coworkers.

There are different approaches how to drive the adoption of the platform (see the screenshot below). In case if you have time I would recommend using the "carrot" and partnering with your users, building solutions that will address their needs and will resolve their challenges. To do that I would recommend involving stakeholders and champions in the conversation early to make sure that use cases that is being implemented will help users be more efficient. If we deliver solution that will help users with their problems and they will confirm them before the use case implementation you won't spend a penny on adoption because users will be waiting for the solution already. So transfer responsibility for the user stories included in the use case on the side of users and do this on the stage of use case design.

Also, I would recommend our blog - link

@Community_Alex​ In my experience, the only way to measure the dollar value of a DG program is once the DG program is actually solving problems. Everything else is theoretical and conceptual, and is often then seen as just an increasing cost factor - which then of course puts the pressure even more on measuring the dollar value. 

Having a defined or at least planned scope of the DG program in terms of business use cases allows you to calculate the value on use case level and then aggregate it. The more funding and scaling the DG program becomes, the more use cases will be supported and generate more value - but usually then with lower relative costs (e.g. you don't have to introduce more data catalogs or more data owners), so it can in the long run become a successful story about return-on-investment increasing.

Totally agree with @tiankaifeng . Everything we do with data is to help businesses be more successful, solving the data challenges that don't allow or impede the successful journey.

We can measure the business outcomes "Immediate (personal) benefits realised from new capabilities" because we can measure "Before" and "After" states. For instance "How many data products people create a month". At the same time, we would like to demonstrate the impact on the business and the benefits that occur as a result of the direct outcome. This task is more difficult because there are a lot of things that impact the business targets such as "Revenue growth".

Another technique is to show as a value - avoided mistakes that happened in the past. Let's say Organisation X had a data problem that cost it $1M. If you build the solution that won't allow this to happen again you can show $1M as a potential future savings.

@Community_Alex​ I've heard various synonyms of "Governance" already, such as "Excellence", "Enablement" or "Empowerment" and many more - my short answer is that an organization should just use a word that works best culturally. We know from being human beings that certain words and concepts trigger negative connotations within us, and that happens often with the word "Governance", so if that negative trigger can be avoided, it's not a bad thing. 

AI Governance might run into the same situation, and I've already see companies already name it "AI Excellence" or "AI Enablement" equivalent to what they called data governance before, so as long as it's easier to get accepted and adopted in an organization, go for it. 

I worked with one organisation that banned the word "Governance" because it "Scares" people. If you ask me - did it change completely how the program was going? Then the answer - no, it was not a game changer. Of course, a good name could help to build a good internal marketing campaign and, probably, less people would refuse your invite to the "Data Excellence Program Kick-off" but at the end of the day - if it looks like a duck, swims like a duck, and quacks like a duck, then ... :) 

We have a team of experts who can help you with some questions during the implementation. Please reach out to your point of contact in Collibra with this request.

@Community_Alex​ ... and if you want help with the people & process part around Collibra, feel free to reach out to me directly!

@Community_Alex​ Yes that is indeed possible, and when you go for process-impact timeliness is usually an important aspect too. The prerequisite for that method of impact assessment is that business process management is taken seriously in your organization and not just another theoretical concept that was never implemented in reality - otherwise data governance unfortunately very quickly gets associated with being "just" a theoretical concept too.

Prioritisation based on the impact is one of the quite common approaches in DG. For instance, Financial organisations in order to comply with BCBS 239 should have identified CDE (Critical Data Elements) and be able to show all the data around them, with quality, who has access to it etc. etc.

So what we could do in your case, is to create the repository of the business processes and identify the most critical of them. Then in those processes, you could define what data is being consumed and produced at every step of the process. At the same time, we could define the data policies, standards and rules that would allow you to monitor the data on the input and output. So in case the quality of the data will fell down below the acceptable threshold you could setup automated notifications for all impacted stakeholders. And I want to add here that all of this can be easily achieved with Collibra capabilities :) 

@Community_Alex​ Thanks for reading my book! It will definitely help you align to business units / business value - and while you're at it, try to use the concept of "co-creational relationship" from the Collaboration chapter to ensure aligning to business value becomes a smooth and fun collaboration :)

@Community_Alex​ The most difficult aspect of determining business value is often missing the business context, and business stakeholders not being available for collaboration to provide that context. It then goes back to mindset, because everyone should understand that data needs to be valuable and that any data efforts should lead to that value - so before any calculations and measurements happen, it's important that people think and behave with the right goals and motivations. 

Agree with @tiankaifeng on 100%!

The most difficult aspect of determining the business value of data projects/initiatives is to connect it to the real business goal that would explain - Why do anything? Why now? Why is this strategically important to the organization? What’s the impact of not achieving the goal?

As soon as we have a business goal we will be able to identify the data challenges we need to fix and solution elements that needs to be delivered and it will be then very easy to explain the business value.