Better Decisions with Metrics in a Business Context

“74% of the companies surveyed indicate that there is no or hardly any integration between IT and business metrics.”

Imagine a renowned insurance company in the year 2016. Business has hardly changed throughout the years. But suddenly customers want to get in contact and interact with their insurance representatives via completely new channels. The still static company website becomes insufficient, mobile solutions must be provided and the application environment, grown over time, must be expanded with innovative cloud solutions.

Furthermore, the insurance provider appears to be confronted with a new kind of competitor for the first time. A great number of startups compete with their increasingly successful, innovative concepts and take a significant share of the market.

The challenge: creating a business context for metrics

Confronted with these challenges, business managers recognize that they must react more flexibly to the dynamic behavior of customers. An information initiative is launched to use the data generated in the IT environment for better decision-making. And there is no shortage of data.

A variety of expert tools generate a flood of data. The insurance company implements dozens of tools in the areas of business analytics, IT development, and IT operations. Development tools, code management tools, data management software, service desk tools, performance management tools, logging tools, IT asset management tools, and business intelligence tools generate information more and more rapidly.

However, this information is not very useful for business decisions. CRM solutions generate data structured according to customer accounts, development tools according to "user stories" and monitoring tools according to entities or servers. There is no context, no common denominator. Due to the sheer volume and speed with which the data is produced, business analysts are swamped with data and are unable to put it into a meaningful context for the decision-makers. Moreover, several experts need to become involved for simple analyses. Consequently, analyses require long processing times for simple questions. Fast improvement cycles are thus impossible.

This problem does not just affect insurance companies. Gartner predicts: "In 2017, 90% of the information from big data initiatives will only be available in isolation and thus will not be usable." In particular, such isolated flows of data can no longer satisfy the paradigm shift towards a continually more integrated model of "business," "development" and "operations," or "BizDevOps" for short.

Giant information projects often run out of steam

But what is the solution? Large-scale information initiatives were launched in the past with a lot of enthusiasm. But the giant programs eventually ran out of steam due to increasingly significant problems concerning inconsistent and redundant data, antiquated data warehouse systems, and organizational trenches.

To avoid making the same mistakes, the insurance company establishes some framework conditions. The new data view should conform seamlessly to existing structures. Business capabilities and an application repository are already securely anchored in the reporting and organizational structure. Therefore, real-time data should be processed in this context.

A fast and iterative approach is the second important basic requirement. The resulting metrics must be easy to use for business analysts. For example, it must be possible to aggregate data to different time periods for every end-user. It must also be possible to add new reports quickly and without extensive expert knowledge. In the past, simple requirements quickly got out of hand and turned into long development projects.

Solution: Create a business context with Enterprise Architecture

The insurance provider has already focused on Enterprise Architecture (EA) as an important link between IT strategy, operations, and development for several years. The team's self-image has rapidly transformed over the years. While they merely used to be modelers of the IT environment, today they ensure effective communication and information exchanges between different stakeholders.

A multitude of expert tools generate data that land in silos without context

The leader of the EA team sees a major opportunity in the integration of real-time metrics into his existing EA information platform and reports to shape the digital transformation of the insurance provider more rapidly. He can now make future- and user-oriented suggestions for the first time: How does the new rate comparison on the website caught on with our customers? Which areas of our software for calculating individual policies are prone to error? Can the backend manage twice as many users?

The Project Steering Committee decides to initially integrate the real-time data from three IT expert tools into the EA solution LeanIX, which has been successfully employed already: The login data of the recently updated mobile app is linked to the number of implemented features from the development tool Jira. The availability of the ten most important applications from the portfolio is visualized with up-to-date data from the monitoring tool Pingdom.

To gain a better understanding of the risks of obsolete interfaces, all transactions for each interface are represented in a metric each day.

With the help of LeanIX Metrics, data points can be easily displayed in various diagram models in the context of fact sheets. A variety of applications are possible due to the increasing popularity of APIs in modern IT organizations. The procedure is simple: The time periods are defined with the corresponding metric characteristic based on the data source. These are sent to LeanIX Metrics via the API. Every data point is marked in LeanIX to ensure that it appears in the right fact sheet. Representation of the data can be configured without programming work.

The application in the EA inventory is combined with real-time metrics.

Result: Real-time IT data enables better decisions

The piloted initiatives bring initial results quickly. Data for decision-making is now available without manual effort in the information platform. More metrics are iteratively added, discarded, and further developed. Data and information now allow a dynamic development of the IT environment and no longer stand in the way. For the first time, the automatically-collected data allows analysts to identify interrelations between business, users, development, and operations.

Better and more reliable decisions with data

Much is already gained by effectively using and incorporating the currently available IT management data in daily decision-making. The insurance company's problem was not the lack of data. On the contrary, the sheer amount of data available was what made it so difficult to put it into a meaningful context. Information is made easily accessible with metrics in the context of the LeanIX EA platform—beyond functions, processes, and organizational units.
As a result, important decisions are now supported
by real facts.

Iterative improvement and orientation to the future

The lean metrics setup allows for the quick integration of new data sources and the testing of various visualizations. As a result, improvements can be made quickly in short sprints based on the needs of the business. This facilitates a must-have capability in the digital age: Short improvement cycles based on real user feedback. Now data is no longer analyzed retrospectively, but available information is extrapolated into the future in real-time.

Use of existing structures

Since the EA repository already exists and is anchored in the organization, a new logic based on which the data is structured does not need to be developed. Similarly, all data sources are already available. Gathering existing data can be achieved much more quickly than developing and building up from scratch. This saves a lot of money in comparison to large business intelligence solutions.

Easy usability for analysts

IT data analysts are in the position to perform their evaluations quickly and without help from experts. Diagram types can be easily modified. Data can be aggregated arbitrarily and applied to various time intervals. As a result, dependencies, which lead to long response times for simple questions, disappear. Even the training effort is reduced due to the intuitive and modern user interface.

Merging of functions such as "Biz" "Dev" and "Ops"

Business models change in digital times. Insurance companies must react to new competitors, communicate with their customers via new channels and convert their data into cash. The integration and exchange of information are now core competencies. New operation models that bring together development and IT operations more closely, DevOps for short, require agile teams and therefore subsist on integrated and freely accessible information.

In our next blog article, the third in the series, we'll be examining five specific use cases, in which real-time metrics were used to solve problems.