Data Governance Metrics

Introduction

The KPIs and metrics presented in this document do not intend to be all-inclusive, nor do they pretend to be ubiquitous, given that every organization is different and needs metrics tailored to their situation. These metrics are example metrics and should be selected with care. To be useful, metrics should be available on a frequent basis, and thresholds, to drive measurable progress, should be agreed on. Metrics can be aggregated over domains and divisions if the organization has a multi-layered structure. 

A KPI, or key performance indicator, evaluate the success of an organization or of a particular activity (such as projects, programs, products, and other initiatives) in which it engages. In our case, the data governance initiative.

Business Asset Metrics

Quite too often the main purpose of implementing a business glossary is cited as “to be able to speak a common language across the organization”. In reality, creating a business glossary is more complex as it first seems, as a lot of things can go wrong along the implementation road. Introducing a business vocabulary into an organization is a convoluted matter and it is often difficult to measure and monitor the value the program creates. Success is not only determined by the amount of information entered but also the acceptance and actual usage of the glossary by a wide community (adoption across and within the business units), how the glossary matures over time ( quality amelioration and richness of the information), to which extent it leads to efficiencies and how it improves integrity in relation to data and systems. To prove and enable the glossary’s value within the organization the data governance team needs to consider different KPIs and metrics, preferably before starting up the business glossary program. Otherwise, the team risks releasing a paper tiger without the wished-for effect. Only by applying the right metrics and KPIs, it will be able to monitor, prove, and increase success over the lifecycle of the glossary.

Terminology

  • A glossary is an alphabetical list of terms in a particular domain of knowledge with the definitions for those terms.
  • Taxonomy is the practice and science of classification, the terms in a glossary can be classified to exert more structure.

Number of Business assets entered

The number of business assets entered within the Collibra Platform

  • Number of Assets entered (per asset type/domain/community or per day/week/last month etc.)
  • Number of Assets planned

Number of Business assets approved/certified /curated

The number of business assets approved by the responsible business steward

  • Number or percentage of Assets approved
  • Number or percentage of Assets under review
  • Number or percentage of Assets entered (Candidate)
  • Number or percentage of Assets rejected
  • etc.

Number of Business assets articulated

Any Business Assets is valuable only if it has a definition

  • Number of Assets with definitions/description
  • Number of Assets with articulated score = 100
  • Number of Assets validated / invalid

The additional metric can be added here – Status of the Assets, e.g.:

  • Number of Assets with definitions/description and Approved

Number of Assets with responsibility/authority assigned

The number of business assets with an assigned responsibility. There are different responsibilities classes, such as the business owner, the business steward, the subject matter expert, etc. This metric represents a measure of governance ownership and stewardship. Establishing Data Governance it’s always about people. More people involved in the Data Governance processes better quality fo the data we will have in the end. But at the same time if too many people will be involved in the same process we can be at risk that this process will never be finished.

  • Number of Assets owned vs Assets without owner/steward assigned
  • Number of Users per role
  • Number of Assets per user per role

NOTE: Combination of metrics is the key for the understanding of how successful your Data Governance Programm

  • Number of Assets with the assigned owner/steward, approved but without the definition
  • Number of Approved assets without the assigned owner/steward

Number of Duplicate asset

Having several assets with the same name is an absolutely common scenario for all companies. At the same time, dozens of assets with the same name can increase the time required for the searching. In addition, we also can not exclude the situation when stewards can create duplicates by mistake.

  • Number of Assets which have duplicates, or mentioned three, four times and more
  • Number of Unique assets per asset type

Number of Assets with data element related

The number of business assets related to a data element, like a column, data attribute, report attribute or a field. This metric determines to which extend business terms and data have been related. This metric is especially useful if all of the terms also have a data counterpart in one or other system. If this is not the case, it might be more useful to measure the number of data elements related to a business term, as theoretically each data element, at least each logical data element, should have a business counterpart.

  • # Business Assets related to Data Elements

Number of Assets with business rule related

The number of business assets related to a business rule. A business rule defines or constrains some aspects of specific business data. It is intended to control or influence the behavior of the business data. Examples are:

  • a customer number should be a uniquely identifiable
  • a country should have a valid list of values
  • a customer country of origin should be in the valid list of countries

A business rule normally is tightly linked to a data quality rule and/or a data quality metric. If the organization does not make use of business rules it can use a data quality rule or the data quality metric as a proxy, given that the corresponding data elements are already related to the respective business assets. In addition, one can also measure the business assets with regard to different quality dimensions, like completeness, accuracy, conformity, consistency, timeliness, redundancy, and integrity.

  • Number of Asset with related business rules

Issue Management Metrics

Number of Assets issues

The number of resolved business asset issues with respect to the total number of created issues. A higher score might be indicative of a better working data governance program, given that the issues are not structural in nature or are not caused by the data governance program itself.

  • Number of Assets issues resolved/unresolved
  • Number of Assets issues identified
  • Percentage of resolved issues to unresolved

Average number of days to resolve an asset issue

The time between the moment the term issue was logged and the term issue was resolved. A higher score might be indicative of better working data governance program, given that the issues are not structural in nature or not caused by the data governance itself.
( SUM(Issues Resolved) – SUM(Issues Logged) ) / SUM(Issues)

People Metrics

Number of people trained

The number of people trained in using the business glossary. Communication and training are very important when launching a data governance initiative. Training is an important first step in enabling and roll out data governance and the business glossary across the organization.

  • Number of people trained (by role)
  • Number of population
  • Number of training per role

Number of users who use the business glossary

The number of people making use of the business glossary. The number of visits and return visits are important to measure the adoption and growth of the community.

  • Number of unique active authors (per product/community/role etc.)
  • Number of unique active consumers (per product/community/role etc.)
  • Number of unique users who logged in only once

Number of users who use data governance processes

The percentage of users who effectively use the new data governance processes related to the business glossary. It takes time to roll out new processes across the organization and this metric helps to track where you are in rolling the process out.

  • Number of People who had at least one workflow task in the last month/etc.
  • Number of People who proposed new terms
  • Number of Active user tasks per workflow/community/role etc.

Data Governance Coverage

Percentage of data domain coverage of the organization

The percentage of domain coverage relative to the organization. Ideally, all data domains (Customer, Vendor, Client or Product, Project, etc.) and all parts of the organization are covered and shared across the organization.

  • Number of described/covered data domains
  • Number of identified data domains

Depends on the organization structure understanding of the data domain coverage can be also achieved by next simple metrics:

  • Number of communities and sub-communities per community
  • Number of domains per community/domain type

Number of authoritative data sources

In order to support all line of businesses and achieve the full transparency of the data in the company Data Catalog should contain information about all data sources which may contain critical business information or which may help to business people make the right business decision 

  • Number of Authoritative data sources described in DGC
  • Number of Data Assets without data source
  • Number of Authoritative data sources without the business or system owner

Number of unique users/roles

Progress of the Data Governance Programm also can be assessed by a number of people officially involved even if they do not take any actions in the tool. This metric most of the time should be presented with one of the people metrics.

  • Number of unique Owners/Stewards/SME/etc. per community/data domain or in total
  • Number of unique roles with assigned users

Number of custom visualization components

Progress of the Data Governance Programm also can be assessed by the user activity. And the number of visualization components created by users for their personal use or shared with colleagues will be a perfect metric for this

  • Number of asset table views private/public
  • Number of search filters private/public
  • Number of traceabilities private/public
  • Number of dashboards private/public

Processes Metrics

Average time for asset reviewing 

It’s important to understand how much time people spend on the reviewing of the one business term. It will help to estimate the required efforts and identify bottlenecks in asset workflow streams.

  • Time from one status (i.e. Candidate) to another (i.e. Approval) per asset type

Number of workflows

During the implementation process especially if there are a lot of use cases in the pipe-line and community/domain structure is complex it will be useful to control workflows development and configuration process.

  • Number of workflows enabled per asset type/domain/community/globally

Summary

As stated before, not all metrics will make sense for your organization and you probably can think of a few other ones for your specific use case. Metrics can be very helpful if they are used frequently, consistently, and interpreted by the right audience. But most importantly, a metric only really shines if it triggers actions to remediate or improve a situation over time.
Common data governance goals you would like to measure are:

  • Coverage, to reach as many users in the organization over time
  • Ownership, to assign ownership to the right people for all business assets over time
  • Completeness, to add and complete as many business assets as possible
  • Efficiency, to achieve cost and/or time efficiency and therefore also reach more profit
  • Integrity, to relate and connect business asset to other data quality and data elements
  • Maturity, to curate and improve assets quality over time

You have to login to comment.