Data Governance Is Data Strategy
By Michele Goetz of Forrester and Eric Dumain of ObservePoint
Hi. Thank you so much for having me. It’s great to be a part of the Virtual Analytics Summit of 2018. What I thought I would do today to kind of help kick things off is really talk about the relationship between data governance trusted data and how to really make data work for your organization. Where do you get value? How do you monetize it? Where are you going to see the best outcomes for your business? By really stuarding your data and governing your data appropriately, and in context of your business activities and your overall data management program.
So, the first thing to really look at is, what is it that your business stakeholders are looking to do? And top of mind is really, how do you simplify the organization and the operations? How do you make it easier to do business essentially? And this is really top of the list. Ya know, 67%, two thirds of organizations are really thinking about how to run their businesses better.
And to do that, they’re thinking about, where does technology play a role in this? And we’ve been digital businesses revolving to digital businesses over time from the very first data centers that we put up to the applications that we’ve adopted to moving to the cloud to thinking about how we interact and transact with our customers, but digital business is getting a new life. And organizations are thinking about not only how do you replicate the things that humans have done all of the time and do that in a one to one manner? But where can technology simplify the way that we’ve done things in the past and run our businesses differently both to be operationally efficient but to get the best outcomes?
And 61% of companies are taking this to heart and thinking about, “well how do we expand our digital investments for this new age as operating as a technically adept company?” And then secondly, it’s about, how do you use digital and technology to really accelerate your ability to operate in this digital world? So, again, not just about replacing on a one to one exactly the way that humans do things, but really using technology to rethink the way that we run our businesses.
This is not easy. And one of the forgotten things about moving to technology and becoming a digital business is that it’s all predicated on data. And data has been really hard. It has always been at the heart of why we have challenges within the technology and digital investments that we make. So, the first thing is, as you’ve heard about big data and we’re now coming onto a decade of hearing about big data and the v’s that are associated to it: the volume, the variety, the velocity. It’s always been so cliche, and a part of this big hype cycle. But now it’s becoming a true reality. Businesses are really struggling to handle all of this information and recognizing the fact that information is coming from so many different sources that it’s seemingly messy. How do you create consistency around this data? What can you believe when you see conflicting views? Or recognize the fact that those views are not conflicting, they’re both accurate.
And on the back end, to keep up with the pace of a digital business and how quickly organizations and stakeholders and executives are now able to make decisions or operate. You need the technology to keep pace with that. So how do you put in the right investments around data? That you allow data to keep pace with the speed of decisions and the speed of operations. Those three things always come to the core of insuring that you’re not accruing technical debt in your digital investments and getting the major benefits out of them.
Now the other side of this, is we’re always struggling under the changes that come from a regulatory perspective. It seems that every week there’s a new one that comes out that we have to handle. Well, the big one right now certainly is around GDPR. And that really thinking about how do we ensure the privacy and permissions and preferences of our customers? And making sure that we are protecting their data on their behalf. And stuarding that information in a way that they find acceptable and comfortable with so we can maintain our strong relationships or build relationships with our customers.
And when we talk to the B to B marketers, particularly, who are really thinking about ensuring that they have strong relationships and those relationships aren’t just with people, but they’re also with companies. 44% really see this as a major challenge. Certainly it’s a consumer problem, but the GDPR isn’t just about consumers, it’s also about the way that you share and protect information about your business partners.
And that sort of brings up the other challenge with why data is so hard. It’s not just that you have to instrument your data so that you can make sense and get value out of it. But you have to take into consideration, somebody else's permissions and preference requirements. And that can throw a wrench into the way that you capture, the way that you store, and the way that you deliver information into your business activities.
Data governance is a must. It is the way that you operate appropriately and effectively around your data. You can’t just leave it up to technology and IT to account for storing the information, keeping copies of it, and then delivering services or delivering data from point A to point B. You have to think about the context of how it’s used, you have to think about the context of the insights that you’re going to get out of it and how decisions are going to be made. And you also have to think about the protections like GDPR requirements or other data security types of considerations.
Data governance is that mechanism to balance between your objectives and your priorities and your risk management. It has to be offenses and defences. It has to be proactive and reactive to keep up with the speed of your digital business.
Now let’s kind of break this down a little bit. I found some really interesting things because often times when we’re thinking about where do we get started with data governance or trusting our data or getting value out of our data? The first thing we always ask is what are our overall business priorities? And we kind of rank them in order to get a perspective of where are the top three or the top five that we should really be looking at? Now when we talk to business stakeholders, what our business technigraphics data from Forester is telling us, data professionals are saying that overall, if you ask us, we’ve got to grow revenue, we have to improve customer experience, and we have to reduce cost.
Down at the bottom, you’re seeing number nine comply with better regulations and requirements. Governance we always think about the fact of, it should be supporting business value, but often times it’s driving us towards solving our regulatory requirements. Solving around risk management. Solving around efficiencies, the operational efficiencies. The reality is that when we orient the discussion around data and the investment around data to where those business priorities are or where data governances align to on those priorities, improve customer experience, driving big data and analytics for business decision making, and improving the products and services that we’re delivering all start to rise to the top.
So what’s important here? Number one, if we compare that to overall business priorities, data governance is starting to get number one, more specific about what data is going to do for the business. And number two, it’s starting to think about those offensive strategies. Those strategic investments from a digital perspective the business are trying to take. But you know when we talk about governance and the first perception of governance, it typically revolves around I know I’ve got budget set aside to tackle the regulatory issues like GDPR is the flavor of the day right now. If you’re in financial services it was around things like babble 2 and BCBS 239. And there’s a whole host of other regulations.
And certainly you can chase that from a monetary perspective, but what you’re really noticing about this data is regulations and requirements, those are table stakes. You have to do that. It doesn’t matter because if you don’t do that you’re not going to get the business value on the other side.
Now, the other thing that’s starting to come in to play is look at number 3 on reduction of cost. So that’s that efficient perspective: doing more with less. From a governance perspective, that’s also kind of moving down and becomes a little bit more table stakes because if you’re doing the right things and putting the right pieces into place, it should reduce cost because the data is enabling your digital business.
So don’t lose sight of the fact and don’t overcorrect on your governance to only focus it on things that are risk-associated or cost-associated for your metrics. Ensure that you’re also building a story and an objective for your governance to really solve the problems and achieve the results that digital strategies are trying to support.
Now if we change the subject a little bit and look at it more from a trusted data perspective. All those things around data quality and making investments there, here’s where it gets really interesting.
Most of the time when organizations are thinking about governance, they’re thinking about, “how do we improve and trust and build confidence in our data so that we can democratize data so that we can monetize it to get the most value?” And here what you’ll see is business priorities shift a little bit differently. Grow revenues was certainly at the top. Leveraging big data and analytics decisions is rising to the top. We saw that in the data governance side.
But look at this number two: improve our ability to innovate. Why is this important? Because as soon as you start thinking about monetizing your data and moving into the data economy to support digital business, your starting to be even more strategic in thinking about what is the future going to look like? This is a longer term investment where data governance is handling the black and tackle of today's challenges and requirements.
Orienting your governance programs to include quality and trusted data helps you to start positioning it to succeeding for some of the future goals and the complete return on investment of where your digital investments are going to go. And again what you find is things like reduction of cost and managing around risk and compliance again is a table stakes capability because if you don’t have that, it’s going to be very difficult to trust your data or to have others who are consuming that data or that you’re sharing it with or the regulators are going to ensure that you actually know what you’re doing around that data.
So really what you’re trying to do here is taking these business priorities and establishing governance, but doing so within the context of your business so that you can be both offensive and defensive and proactive and reactive in those exercises.
So let’s go back and start pushing on the things that actually make a difference here. Where can you look? How do you orient? It’s very easy sometimes to think about your CRM systems. They account for a big portion in supporting of what your business activities are going to do because things will revolve around your customers. Remember those business priorities, ensuring great customer experience, this begins in CRM systems supporting the processes and activities that are going to allow you to do that.
Organizations struggle with this. And number one, they’re struggling with the quality of information within our CRM system. This comes up over and over again, and I know within my almost thirty years of practice around data and data governance, it’s the same complaint. How do we keep up with the quality of information about our customers and our activities in our CRM system? And ultimately, the next thing is why? Because then it’s really hard to get insights out of it. How do you guide your decision making when you can’t trust your information?
Now on the other side what you see though, is as CRM processes kind of stayed the same, the quality challenges are going to as well. Why do we say this? What does that really mean? It means that a lot of the ways that we’re instrumenting our CRM systems are trying to coincide to our business processes. But in the end, every time you change a business process, you change a step, you introduce new automated capabilities, you want new views within that CRM system, that all has a downstream impact to you data.
So if you haven’t addressed the data in context of changing the changes that happen upstream in your CRM system, you have further drift in your data and you’re building up more and more technical debt around that data. It’s basically data quality debt within your system. And if you think about the fact that CRM is such a central force or center of gravity to the rest of the systems within your organization that rely on a system of record for customer information, you can see how this doesn’t just live within the sales and marketing environments or customer service and support environments, it’s hitting into how do your engineers and developers understand the way that they’re going to create products and align with what’s going on in the market. It’s going to have other upstream effects to the way that you’re managing around your financials and truly understand your business state. And certainly it’s going to make a difference to understand if you’re applying the right resources. And is HR hiring and training and maintaining strong skills that help you succeed in driving those great customer experiences and relationships?
So, while CRM is a microcosm, you can see it certainly trickles out so that relationship between process and data is a great example in here about how you have to keep them aligned.
To do so, first thing that you want to think about is orient to what is that relationship with your customer. Look at it in terms of the process with your customer. How does technology or digital investments align to this, support that, and automate that, and create great experiences. Then what is the type of information that is there that support that information, not only within the context of raw data, but what are direct insights that might be coming up through that as well?
And then lastly you can start thinking about your sources. This is a complete flip to the way we normally think about data. Which is, where is our data warehouse? What is that data lake we’re building? What system is data being collected in and living in? We think about our data and our architecture and the way we invest in it as a data museum. But if we want to activate our data, if we want to monetize our data, if we want to drive great insights and experiences with our customer and ultimately drive revenue growth, we have to think about what data is supporting. So orienting toward something like a customer journey or key customer journeys that you want to implement. Or you’re in a manufacturing and your top priority is the way you’re most easily able to and quickly build and adapt your products to market demand. Or thinking about your logistics system. All of those are processes. Start with the process. Build that down. And your data will follow and support those decisions and actions.
How do you identify where you’re going to focus? How do you prioritize? Keeping up with the pace of business is really important. This is something that you’re doing just naturally through some of the tactical decisions that you’re making on data to solve some of those ad hoc requirements or project requests on a quarter to quarter basis.
But annually, your organization is going through a planning cycle. From those planning cycles, our strategic plan that you’re c suite and senior leadership teams are pulling together that are describing how are they going to solve against the areas that they care about. What you need to do is step outside of the data for a second. Think in terms of that digital business. Talk to your business stakeholders. The senior leaders in the organization. Have them present their strategic plan. Listen to be able to solve and facilitate and then act upon those plans so that you’re keeping data not only aligned to what questions that they have about their marketing customers, but also about what is the type of information and insight that facilitates the automation of the business and the digital experiences that are going to drive business activities and customer engagement.
The third thing to think about is how do you scale data governance in a digital ecosystem? You can’t do that strictly by having a small team for data governance and stewardship. You really need to think about the mocrotization. And we often times, when we talk about data governance, talk about this notion of ownership. Who owns the data? IT says data governance teams own the data. Data governance teams say the business owns the data. Or maybe IT says business owns the data. Well, that’s great, but if you haven’t enabled your stakeholders and users and consumers of the data in ways that make it easy to do scaled out stewardship, you are going to fail. Small governance teams could never keep up with the amount of data that they manage, and can’t keep up with the drift of the data over time as your business changes, as your market changes, and as your competition might change as well, or even as your products change.
So rather than just managing and gathering data from a centralized batch and task oriented fashion, think about retooling some of the services and activities that you do within a stewardship and remediation environment and accommodate that from where data are changed or updated, deleted, or even consumed so that you’re citizens, those stewardship tribes that live in the business, have the ability to sell steward and take care of things immediately rather than waiting for that to happen in the backstream.
Really the key thing is how do you institute services in the application to guide data responsibility as part of day to day practices? Ensure that the environment that these stewardship tribes are living in accomodate for the types of self service steps that are going to happen, particularly in analytics. Because a lot of data transformation is occurring there to prepare data sets that will then be used and guided out. Think about how much you learning is going to support things. There are a lot of capabilities coming on to market within your data management environments where machine learning is helping you to do a lot of this automation, to be predictive and suggestive in the way that you’re going to manage and govern your data. And also to do mass interpretation and extrapolation of things like classifications and tagging so that you’re maintaining and improving and extending things like taxonomies, hierarchies, master data management models and so on.
And lastly, put a BI layer on top of the data that is allowing you to look at data in context of the workflows that occur within the stewardship tribes so you can maintain alignment between business process and data upstream in the applications rather than waiting for it to flow back into your warehouse or your big data environment.
And ultimately, what that’s bringing you towards is, the fourth thing that’s really important is tell a story. You can’t improve what you can’t measure. We’ve always thought about that in terms of, well I’m watching the business rules and am I passing or failing the business rules associated to my data that I’ve programed in? Those are really interesting and that tells you technically if you’re taking care of any of the more detailed granular issues that are coming up with your data.
But from a governance perspective, what does that mean? If governance is set down a set of core principles, a set of goals, that it needs to attain, or how are you taking the analysis around your data and showing that you’re complying and adding and achieving those data governance objectives that are going on that are highly policy based? And start touching upon those areas that are related to the business expectations for the data.
And then the last thing is, and where the story matters the most is, how by adhering to those policies are you making a difference in the digital business? How did you simplify operations? How did you speed up your ability to deliver on processes? Bring products to market. Handle customer needs and requirements. Adapt to the market. The aspects of the business, what you’re trying to achieve, all trickle down to what you’re policies are being stated as to the way that the data is related to.
So, again, thinking about earlier, that notion of process to application to the information, to the data sources. Your business intelligence layer on top should tell that complete story so that your technology teams know where to prioritize and act. Your data governance teams know where to prioritize and act. And your business has a good understanding of how data is helping them today and where potentially data can take them tomorrow.
So lastly, what can you take away from that? Four things.
Governance is your strategic framework for data. It isn’t just an operational risk management capability. Remember, offensive defensive, proactive and reactive. That helps you keep aligned to what your business priorities and expectations are.
Don’t forget about how data and process map together and are aligned to decisions and actions that are going to happen both within the digital aspects of your business, but also in the more manual or strategic planning areas that are going to happen. Always put data in context of what those processes, decisions, and actions are, and you’re ensuring that data is aligned to the monetization and value it’s going to bring.
Don’t leave stewardship to centralize teams and committees that can’t scale effectively and be adaptive and elastic to the needs that are happening in your digital ecosystem. Instrument services and enablement into the way that people do business so that data is stewarded at scale and data issues are quickly addressed and taken care of so that it doesn’t have an exorbitant impact on your business and you’re not creating additional data on your data.
And lastly, tell your story. Build that business intelligence view over your data operations similarly to the way that you think about building your business intelligence view around all of your business units and their operations. And then you’ll be able to tell the story of how data is supporting the digital business. You’re contributing to those key priorities around things like growing revenue and customer experience and creating that efficient and effective business. And you will be doing great.
So thanks for taking the time and listening to this. Hopefully you’ve got some good ideas and practices to take your next step around trusting your data. Having confidence in it and building the next generation of data governance capabilities. And what I’d like to do now is hand it over to Eric Dumain from ObservePoint, and he will tell you about other things you can be thinking about to be the best data professional you can be.
Well, thank you Michele, for your presentation. And thank you Brian. I will try within fifteen minutes to highlight how data strategy and planning are key success factors for data governance policy. And I will try within this short time frame to give you flavor or vision for structuring your data governance plan.
First of all, let’s answer one question. What defines data governance model? It could be data origin, data sources, data nature, data processes, or data usage. In fact, it could be many more options. The nature, the relevance, the quality, and the volume of data you will collect and aggregate will significantly impact lengths and efficiently of processing. As well as the potential of usability of such data.
My experience with data proved to me that in order to be long term and sustainable data governance, what really matters first is to define what you will want to do with your data. So data usage and goals are good ways to define a data governance model. And to build a data strategy accordingly.
One of the most complex and fascinating cases in data governance is related to digital analytics. And here are 7 components of complexity we could highlight. Content, campaigns, and releases that change in a high frequency. Multiple data sources and new devices. A huge volume of generated data covering multiple purposes. Most of the time, collected data are coming from unauthenticated sources. Compliance to many regulations. Near real time decision processes to match fast online saves our companies cycle. And finally, these martech related components are part of the first step of valuable data governance model.
This may lead companies to limit a number of variables they will use in their tags. Hence, the volume and nature of data they will collect which will lead them to inefficient, if not useless, digital analytics implementation. Other companies could decide to use hundreds of variables, potentially generating millions of useless data leaving them to inefficiency and real-time processing and significant extra costs of data processing.
A data governance model really begins with data strategy dedicated to online data collection. It’s no mystery if 69% of marketers say they feel their martech help them to do a better job.
While 42% of CMO say they couldn’t fully attribute sales results to marketing investments.
Anyone involved in a digital business dreams about recognition automation of converting patterns. Anyone dreams about collecting and processing relevant and accurate data accordingly. Most advanced contributors will want to seek your most pure profitable patterns, will want to improve average ones, or will want to quickly and automatically cut those that are inefficient.
In a digital universe led by instant marketing, speed and automation are two success factors. We are not far away from that. Actually, we have in our hands all the technologies methods and best practices we need to be there. Those who will be able to reach and align point for structuring that data governance and strategy will take a competitive advantage.
So, let’s try to figure out the changes generated by global digitization. Over the past three years we have seen the role of the chief digital officer rising. One of his core focuses is to get organizations more reactive, flexible, simplified, and flattened in order to address more efficiently fast changing environments.
Another core focus of the CDO is to get insights that match the needs of transverse organizations, compressed of the main company, it’s brands and product lines, the countries they cover, as well as their distributors and online partners.
So, in this sense, while a CDO flattens a companies organization, he enlarges the scope of digital ecosystem in order to secure end to end data collection strategy and governance policy. Securing an end to end data collection strategy and an end to end data governance policy will be our two goals today.
As we said, multiple data sources are another factor for complexity in building sustainable data governance plans. Spreadsheet management is not any more sustainable from governance perspective. Not to mention the data flavor of the day, which is not really appropriate for reactivity perspective in a digital universe that tends more and more towards more real time actionable insights.
Recently the GDPR introduced another component in the data governance question. Beyond consent management, it is about privacy by design, data collection methods, data processing, and data aggregation limitations. A good thing about GDPR is at least that it provides some items for data governance and it pushes multiple departments in a company to work together and to get aligned. Started by marketing, I.T., and legal departments under the control of the data privacy officer.
So now, this is time to throw the headache slide, what I call the headache slide. I hope everyone is ready for this one. We are going to address 3 axes for data governance model and strategy.
A data governance strategy can be defined by multiple components distributed on 3 axes. Alignment, processes, and control. Each of these axes is critical to build and to get an assessment of our potential of success. On the alignment axis one key component for data governance is the data model. Most companies usually set it in the process axis, and it’s not wrong as long as this is a small or centralized company. However if our companies are a wider organization then the data model will stand on the alignment axis. On the process axis the key component for data governance is control. Setting business processes without control is the same as an accounting department without a financial controller or an airport without a control tower. Too often companies say something like, “oh yeah we know it’s important, we should take that into consideration.” You know the kind of things said about I.T. security until they were under multiple attacks. Well, in the case of data governance for instance, the risk could come from a governmental organization in charge of GDP Reinforcements. And this is why on the control axis, the key component is compliance. Compliance to our data collection strategy. To our tagging plans. To our data layer. To GDPR. All in all, compliance to our data model.
Now, do you see this here? This little black marker? So if we don’t do anything with respect to alignment processes and control, this will be where we will stand. This will reflect our level of performance. We can say zero. The fact is that most companies stay focused on processes and at the start that’s a good start. Our marker is making some progress. Now look if we include the control model, we will then be in a much better position. However, this mostly applies to very centralized companies. One company, one territory, one website. Now, securing alignment together with processes could be seen as a good move. However, how can we make sure that we have a complete or at least a significant alignment on our processes? Without control there is no way, and we seriously may reduce our chances of success. Now look, adding the control access leads us to a sweet spot since it covers the three axis for data governance: alignment, processes, and control. Without control, alignment and processes are almost nothing.
Oh, and by the way, here, this applies to complex digital ecosystems and global organizations. This is what I will demo a bit further from now.
And good news, this is the end of the headache slides.
So, first access: alignment. Alignment applies to global organizations down to a simple project. Organization and project alignment help avoid episodes and silos. And at least it helps better and faster overcome these kind of issues when they show up. Alignment can be extended to ecosystem actors and contributors such as brands, countries, or distributors, as well as external actors such as firms and agencies. Of course, each of these actors and contributors may have their own goals and their own agendas and could be contradictory from time to time.
Which, I have to say, introduces the first dose of fun in the process. So let’s assume we are a group comprised of multiple brands in multiple countries. We want to have a global data governance policy. In this case, we may want for instance, to have a harmonized view of data collection produced by purchasing patterns in order to quickly make educated decisions and changes when necessary. And in order to fill in our data governance model accordingly.
To reach a point of harmonization, getting a full commitment on alignment is crucial. That’s not an easy one since any organization is made of real people, real human beings who are not yet replaced by artificial intelligence. So everyone will claim his organization or his data model is unique or is too specific to fit inside a global model. And beyond that, there is the emotional part of it that must be taken into consideration.
I personally have been running for ten years a consulting firm specialized in leading and deploying digital analytics projects across global organizations. You can trust me, I’ve never seen something different. And this is the reason why the global data model should not be the main company data model. What we will want to do is to include room for specificities in our data model. In order to make it global to get any component or an organization and contributor in a project to get their endorsement. Commitment on alignment is a must have for any component of the alignment axis.
So, here are seven recommendations for alignment on data collection. First one, set data governance at the first stage of your digital strategy. Then define your digital ecosystem boundaries and goals limits. Involve business owners in the plan phase. Define cross channel and cross device policy for data collection. Exclusively collect meaningful data or after process reusable data. Ensure collected data will lead to consistent and applicable insights. Make sure that data processing is GDPR compliant from collection, encoding, aggregation, import, or export perspective.
So, now let’s talk about process.
So, whatever an organization is, I mean centralized or global, it appears data fragmentation is still part of the DNA of about 100% of them. I think it is cultural if not historical. In a digital universe, based upon near real times and marketing an end to end business models, data fragmentation is the last thing we want to see. Fragmentation starts with fragmented data collection and leads to silo and irrelevant insights potentially able to jeopardize a global strategy. This is a model that should be forgotten otherwise our two main goals will be missed. I mean, harmonize data collection strategy and consistent data governance policy.
We will prefer to go forward for coherence and for a model that helps go for a global approach that is strict enough to match our data governance goals. Of course, we won’t forget to include the specifics on top of it. To seek your customization and instant marketing requirements.
Data collection is a never ending story on the web, on apps, on mobile apps, on social networks and others. A monitor process across teams and contributors will help secure compliance to our data model. Some steps may be monitor daily, weekly, or monthly. However, the first one stands at the heart of our cycle since it is built for the long term, and it is related to setting global goals and metrics. We will soon discover that our global process includes a quite significant part of control.
So, 7 tips to define simple and efficient processes for better data governance. Number one, build and share a global data model with room for specificities. Sharing is critical and room for specificities is critical for success. Build an efficient, sustainable, easy to maintain model for data collection. Identify data points by practicing your own websites and apps as if you were a first comer. This is an awesome way to really identify key data points. Collect only meaningful, reusable, and accurate data you really need. Then build sustainable tagging plans that match your ecosystem policy down to website sections. Set and automate control processes for data collection. And include GDPR and other regulations compliance in your monitoring model.
So, let’s talk now about the control axis, and this is where ObservePoint helps a lot.
As we said, the data governance policy begins with processes dedicated to online data collection. You will want to ideally collect and manage appropriate data only and to make sure that it is accurate. This is why to reach these goals we strongly recommend you use the tag governance framework.
Now, do you remember our slide about spreadsheet management for data sources? Well, the tag governance framework will help you manage these multiple data sources. Huge volumes of data and related processes easily and exhaustively inside one single model.
For instance, two key components of ObservePoint’s upcoming release, are dedicated to plan and build a data collection model. They will help structure a sustainable data collection strategy that will be able to follow the release cycle of a digital ecosystem and to help maintain a global data governance policy. I invite you to learn more about ObservePoint’s tag governance framework by participating in Chris Baird’s VP Marketing session today.
I thank you for attending our Virtual Analytics Summit Keynote, and I turn it back over to Brian.