In this blog post we will discuss why a help desk, customer service and just-in-time approach to data governance and data intelligence should be taken by organizations. We’ll provide some examples, discuss data steward involvement, and talk about content prioritization, content creation, and points of engagement. We have mentioned this approach before and wanted to re-emphasize just as Aaron did in his recent blog post “A Help Desk for Your Data”. Additional videos, blog posts, and recorded webinars about this approach can be found in this resources blog post. Data governance and data intelligence work regardless of how you want to spin it. But if done right, you save a bunch of time and effort and money by having good documentation and improving data literacy. But to get data governance content created takes effort. You need people to do work to create content, to write definitions, and to curate reports. And you want this content to be engaged with, and the suggested approach makes this happen. Think about data governance like you do about a technology help desk system, which is something that most of you are very familiar with.
Some help desk and customer service examples and discussions:
A tech support team for a company had a big backlog in calls. And there was a big initiative to close more calls per day, and to kill that backlog. And they really did as they greatly improved the number of calls closed in a day. But the backlog did not go down, which caused frustration. But then the manager said, "Before we had few calls and had a backlog. Now we have many calls with a backlog which is good. More people are using the support system." The point was that if you are responding and helping people and they are finding what they need, more people will engage with that content. The reality was that the word got out that if you called and had an issue to tech support, that it would get an answer and more people were calling. The backlog number stayed high because people were using it, benefiting the organization. This will also be the case with a help desk, customer service, and just-in-time approach to data-related requests.
But nobody expects a help desk to anticipate and have an answer to every possible problem anyone is ever going to have. You might have a knowledge base of frequently and previously asked questions. But I have a problem where there is not an existing document explaining how to fix it. Now I can submit a ticket saying. I could not find a solution online, could someone help me? That is going to get routed to a help desk person to get back to me and help me through this process. The problem is resolved, and the resolution is put into the knowledge base so that the next employee who has the same problem can find that. This was a help desk, customer service, and just-in-time effort to satisfy my issue and update the knowledge base with new information. You have a question about a report that is not documented or it is not curated. You are trying to get information on a data system that is not yet inventoried, whatever it might be. If you cannot find it, you have a way to submit a request.
When I submit a data request, I might get routed to the right person who knew how to fix my issue. Maybe instead, I am going to get routed to a first-tier support person who is a capable customer service person, has some general knowledge, but is not a subject matter expert. But what they do know is how to find that person in the organizaition, at a second level, who is an expert to help answer questions or create the necessary content.
I am a new employee. I am working remotely, and I cannot connect to the VPN to access my files or the shared drive. I know that there is a help desk application so I log in. I search for information on how to get my VPN to work. I find some documentation. Maybe that is the first positive engagement. The documentation tells me how to get my VPN set up correctly. That is a positive engagement with preexisting content. Let’s change this to a data-related example. I have a report, and I am not sure how a certain column is calculated. I search for the report specification and find it. The information I require is in the specification. This is a positive engagement with the data governance knowledge base.
Say you purchased a scanner. You tried to get it to work. No success. You went to the manufacturer’s website and followed their troubleshooting steps. Still no success. And then you called up the scanner tech support folks. After a few questions the support person said "You know what? We sent you a bad scanner. We are going to send you another one. In a few days, you'll get a new scanner in the mail." After receiving the new scanner, it worked fine. And because of this experience, you will probably become a loyal consumer of their product because they fixed the problem. They interacted with you, and they solved the problem. You get better engagement from people if you fix a problem, then if there never was a problem before, and it just worked. Do not worry if the way that people get data governance content is by requesting it.
An objective for data governance is to assist staff in their daily activities and make them more data literate. And to accomplish this you need different types of data governance content such as:
- business glossary
- data system inventory and data models
- documentation of reports and ETLs
- data lineage information
- data requests and resolutions
- information about reference data
- data quality issue reporting and resolution
- data quality rules
- data policies
A solution like the Data Cookbook contains this content in a knowledge base. Content only adds value if there is engagement with this content. A happy staff is one who can find information about data easily and if not found, can get an answer to their question or issue quickly. Feel free to check out additional resources on points of engagement by accessing this blog post. Think about how users engage with data governance content. The types of points of engagement with this content are searching and discovery. For example, let us say someone is trying to do some self-service reporting where they wanted to include the employee's original hire date. They wanted to figure out do we have that? Does that exist in the glossary? If it does, does it tell me where it is and how do we define that and where does it live, in what data system and how do we pull that? Or they are searching for a report. They are going to search through the knowledge base and see if it exists. And if they find it, they are going to be able to discover and get that information back. In another type of point of engagement is looking at a report and I have a question. There is a link that explains the curation and documentation for that report. Now, if you do not find what you are searching for then that leads to another point of engagement - a request for information. Not finding the information is not the end of the world. This request will get routed to a data steward for curation to either review, approve, author, write, or to create that content.
Here are some points of engagement to data governance content:
- searching for a data model in your data catalog
- searching for a term in the glossary
- requesting a new term for the glossary or a new functional / technical definition
- reporting a data quality issue
- seeing if there is a data quality assessment on an issue that you have seen
- looking for documentation or specification on a report
- requesting that a particular report be curated
- requesting a new report
- trying to understand the lineage of an ETL process
- suggesting a new data quality rule
- needing access to a particular report or data system
- investigating set of valid values, reference data, or status codes and what they mean
- trying to figure out who is the owner of a particular set of data
- wanting to view a particular data standard or policy
All of these are points of engagement which should initiate a workflow or process to route to the right person (data steward or subject matter expert) to resolve that issue or to create the content as necessary. And if no data steward is defined, then the request should route to an oversight group. Staff need to be able to easily search for information and request information. If you are a casual user or consumer of data within your organization, you are not going to remember that there is this data governance tool available. We recommend putting searchable links on websites or on reports. Put points of engagement where the data users are usually working.
If you have an existing, IT ticketing system then you can use it for data-related requests, if you do not have a solution like the Data Cookbook. And you can use APIs and tools to let people search from the tools being currently used for reporting. I am looking at the report, I can pull it up and click on it and it will pop up the information report. Then you have your analyst and report creators who maybe don't know how things are defined, but they are asked to build reports. And then you have data system owners and data tool owners, and then your data governance committee. They all want to have ways to interact with this content. A tool like the Data Cookbook is helpful to route requests, and manage requests including approvals and routing back to the requester for closeout.
As part of the process, do not forget to close the loop back with the person who made the request. Come back to the requester and say here is the answer to your question. Did that answer your question? Or they reported a data quality issue. Come back and say you were right. There was a problem there and we fixed it, or we are going to fix it, or we cannot fix it, or whatever. Or sometimes, with a data quality issue, you are going to come back and say we investigated this, and we think there is a misunderstanding. Let me explain how this report works and why it gives you the numbers differently than you think. The whole point is you want to loop back to that person, and that is going to help build their trust and use of the data.
We assume that your organization has limited resources and time to curate and create data governance content. If that is the case, how do we prioritize the data governance effort, and still be sure that we are supporting the necessary engagement to meet goals? How do we reach that balance with engagement to have enough content or to provide the content in a timely fashion when it is needed? You should prioritize content creation based on requests. You must have enough content that there is a reasonable chance requesters may find something. It does not even have to be that much. If someone is coming in and looking for something and they cannot find it, then they can submit a request.
The measure of use and adoption of data governance should be considered more of a balance and exchange between content and engagement. Are you able to have enough content that people who are doing searching and discovery are finding what they want? And is a satisfying experience for them? If not, they request that information that you can now addit that content in a timely fashion to be of use to them. If you just roll out a minimum amount of content and people go and search and do not find it, but you do not give them a way to say, "Oh, I didn't find this. Can you get this information for me?" And have them get that information or answer in a timely fashion, then you are not striking that balance either. In an ideal world, you want to have this balance between what we call a content tipping point, enough content that people feel like there is some chance of finding something, but if not, having a very responsive request process, that they can get an answer back when they need something.
Let’s talk more about data stewards. Sometimes we hear that organizations could not get adoption of data governance or get people to use the knowledge base. We ask, what do you mean when you say that people are not using the knowledge base? And the most common answer was around content creation. Their data stewards are not writing the glossary definitions, documenting reports, or curating valuable content that would provide the necessary information.
One example is definitions. Let’s say you have 20 data stewards, across different domains. And each one of them is asked to create 50 glossary definitions for the most important and critical terms within the organization. And those definitions are not getting created. There are lots of reasons for this. The request was hypothetical and very backward-looking. It is creating content that is not tied to a goal or a specific request. We recommend that rather than asking to create 50 glossary definitions instead, say, "Hey, can you look at your 10 most used reports and define the definitions of items on these reports?"
A solution like the Data Cookbook helps prioritize the data steward’s time and improve the data steward's experience of interacting with this content and creating content. Let them see how it applies to the goals of data governance, and why you are doing data governance. Also, from a business glossary standpoint, you are going to get better definitions if you are doing it in real world applications, as opposed to a hypothetical.
In conclusion, none of us want to invest our time and energy in creating documentation that is never looked at. It is not uncommon for an organization to create a business glossary or some other documentation, and it is physically put in a binder and put on a shelf, where no one ever looks at it or uses it. What you want is data intelligence-related content that creates value and supports organizational goals. If the goal is data literacy and you have defined some business glossary terms, the only way that those terms are adding value towards your goal of data literacy is if someone in your organization at some point in the future is asking a question and finding the answer in the knowledge base. But if no one ever asked that question, there is not a lot of value in documenting that definition. Prioritize this content creation on requests coming in. Anything new that you are doing, document that as you go. Get more buy-in from the data stewards, authors, and creators. This also gives you an unexpected return in trust with your data requesters and consumers. The point of this concept of just-in-time, help desk, and customer service approach to data governance, is to have as much content as you can create, but do not kill yourself. Do not wait for perfection to roll it out. Think of content as a continuous improvement project and prioritize content creation based on requests. Educate your user community to let them know that if they request something, they will get a response.
Hope you found this blog post beneficial. Link to recorded webinar on this subject and link to DG Resources Page
Hope you found this blog post beneficial to you and your organization.
IData has a solution, the Data Cookbook, that can aid the employees and the organization in its data governance, data intelligence, data stewardship and data quality initiatives. IData also has experts that can assist with data governance, reporting, integration and other technology services on an as needed basis. Feel free to contact us and let us know how we can assist.
Photo Credit: StockSnap_HWLNQD65VR_PeopleHelping_HelpDeskApproach_BP #B1179
