Getting started with CSDM is easier if you have a simple model for implementation. We think that the 5-Step Model we present here is helpful. It works well with any data model, and it works well with CSDM alignment and the CMDB.
Maintaining high Data Quality in CMDB has always been challenging and probably always will be. Organizations have taken various approaches to organize their CMDB, with mixed results. The ways of maintaining the data are equally variable.
What is shared by all is the need for the CMDB to perform since an increasing number of processes rely on the shared data it contains. Furthermore, the need to align with CSDM is increasingly important. After all, ServiceNow has made it clear that many of the new features on the platform will rely on CSDM as a data model.
With the introduction of the Common Service Data Model, we now have guidance as to how Servicenow expects data to be in the CMDB. However, in practice, very few organizations can start from scratch and be “CSDM compliant” without a transition. For most, it is a journey during which existing data, or some part of it, is gradually aligned to the new model.
The 5 Steps Model
We’ve talked about our “5 Steps to Better Data Quality” for a long time. We published the first blog posts about it already in 2018. To put that into context, that was before the CSDM was even published!
Looking at the 5 Steps Model now, it is as relevant now as it was then. The 5 Steps model relies heavily on the concept of data ownership and the Consumer – Owner – Provider model. If you’re unfamiliar with the COP model, you can read more about it here.
There are two quick conclusions to be made:
- The 5 Steps model works very well for CSDM Alignment
- Since the principles are generally applicable, the 5 Steps Model can also be used for any data outside CMDB.
Like any model, the 5 Steps model is a simplification of reality. Nevertheless, we know it works because we’ve seen it repeatedly happen with our customers who use the model to figure out how to get started with Data Content Manager.
The 5 Steps in Practice
Focus is key. To quickly summarize how to apply the 5 Steps model in real life using Data Content Manager:
- In Step 1 we select a data domain to focus on. We’ll choose Business Applications and decide to go from Zero to CSDM Crawl. The CSDM Crawl data model is, therefore, our Target Model.
- In Step 2 we Define Data Providers. They are the people in your organization who are expected to provide the data you need. We then create a simple Blueprint for our Minimum Viable Data and run an audit to see where we are.
- In Step 3 we Secure our Minimum Viable Data, which we defined in the previous step. We do this by auditing against our MVD Blueprint and gradually improving.
- In Step 4 our goal is to Reach our Target Model – CSDM Crawl in this case. We run the MVD Blueprint and the CSDM Crawl Blueprints simultaneously to always have our Minimum Data in order. That is what we build on when reaching toward our Target model.
- In Step 5 we Refine and Repeat. This could mean expanding on the existing data model(s) or focusing on a new Domain.
Next, let’s look at these steps in more detail.
Step 1 – Select Data Domain
Data Domains and Responsible Persons according to CSDM 4.0
We also need to select the most important data class. This class will be the so-called Root Class in the Blueprint you will draw in the following steps. Again in this example, it would be the Business Application class.
What is the “Most important” domain varies between organizations and perspectives. It may not be an easy choice. Nevertheless, we need to choose a domain to get started. Things to consider that impact the importance should include data criticality, volumes, current quality, required effort, and people’s availability.
When we have chosen the domain, it is set up in Data Content Manager together with the domain responsibilities. Below is an example of the Business Applications domain set up in DCM: it is owned by Abraham Lincoln and managed by Fred Luddy.
Step 2 – Define Providers
Step 2 Target: Define where to find data provider information for each record and ensure that every record in the selected domain/class has an active data provider.
First, we define how to find the responsible person or provider for the data. In this case they would be the people responsible for Business Applications, since that’s our focus. We can draw a Blueprint into Data Content Manager using the Blueprint Designer. The most important class we selected in the previous step – Business Application – will become the Root Class.
Starting from the Root Class, we extend the Blueprint to where we expect to find the data provider information. In this example, we drew an “IT Application Owner” reference from the Business Application to a User record. In addition, we drew an additional “Managed by Group” reference to a Group. Quite often, it’s one or the other, depending on whether these responsibilities have been assigned to groups or individuals.
Here’s what the simple Blueprint now looks like:
Once the Blueprint is complete, we run an audit against it for an initial analysis to see how many of our “active” Business Applications are missing data providers or if the providers are no longer active.
Once we have a comfortable level of provider information on our records, we can move to the next step.
Step 3 – Secure Minimum Viable Data
Target for Step 3: Define the Minimum Viable Data requirements in addition to Data Provider details and enable Data Providers, who are now responsible for filling in the minimum information for their CIs.
We can adjust the data provider Blueprint to include our minimum requirements for Viable Data. We will use the adjusted Blueprint to continue auditing the provider details together with the rest of the minimum requirements.
The additional minimum requirements can be related to the identification and basic details of the Business Applications, like
- a unique name or another unique identifier
- a description
- a business criticality
- an architecture or an application type
Attributes listed above can be defined in a Field Setup that is part of the Blueprint. The Unique Identifier can be set up for duplicate checking. Visually, the only differences to the previous blueprint version are the Field Setup and the Duplicate Checking icons in the Business Application class. Links to Data Providers will remain the same.
Below is what the Blueprint would now look like:
We will then create a scheduled audit for the new Blueprint version. We can run the audit for all Business Applications that are in use. We additionally defined that there is no need to audit retired applications. However, we have chosen to check applications that are under development or at some other stage before the operational stage.
When we’ve run the audit, we can check the results and inform Data Providers of deviations. Because we now know who should provide the data, we can make data quality personal. Personal data quality KPIs will likely significantly impact their motivation.
The Audit Messages created by the audit can (optionally) be used to automatically create tasks for the data providers. For records that are still missing the data provider, we address the deviations to the domain owner, or we can create tasks about the missing providers for the owner as a to-do list.
Step 4 – Reach Target Data Model
Target 4 Target: Reach the Target Model after data providers and minimum required details are secured. Set targets for Data Providers to increase compliance over time. Celebrate achievements and actively communicate progress and wins.
In Step 4, we will focus on our initial target data model. In this example, the target is to reach 95% compliance against the Crawl phase data models described by the Common Service Data Model. For Business Applications, it means adding the required CI relationships to an Application Service. CI relationships can be direct or go via an SDLC Component.
Here is the new iteration of the Blueprint:
We will leave the previous audits running on their schedules since they provide valuable KPIs about where we are. Since we now have insights into our data quality, it is a great time to collect feedback and identify quick wins from the previous steps.
- What have we accomplished with better data quality so far?
- What can we learn from it, and what can we do better?
- And, importantly, can we already see where we should focus next?
Once we know, we can continue to develop yet another Blueprint version with more mature requirements or increase our scope. For example, it may not make sense to audit all Business Applications against the Target requirements initially – perhaps limit that to the critical applications.
Step 5 – Refine and Repeat
We should also consider requirements for particular use cases and see if creating separate Blueprints for those makes sense. For example, if we know we need a specific set of attributes and relations for a report to function correctly, we could create a Blueprint to validate that data.
You can create different Blueprints for the same data based on the use case so that different points of view and responsibilities can be addressed accurately.
With a systematic approach such as this, and with Data Content Manager providing real-time metrics on your progress, you can easily communicate your current data quality, target models, and progress to the different stakeholders. As a result, you will have a much better chance of winning trust and support. CMDB initiatives, like any other project, require management support to succeed.
The 5-Step Model is simple, robust, and applicable to all kinds of scenarios. However, as you can see from our example, it is straightforward to do with Data Content Manager. With DCM, you draw the Blueprints, and the rest will follow – including dashboards and reports that will tell you exactly how you are doing against your target models, what you need to focus on, and with whom.
It will not do the leg work for you, but it will guide you and make life easier with your stakeholders. Book some time with us, and we will show you how to get started, regardless of your maturity level!