Avoiding the Garbage-In, Garbage-Out Problem with Your Descriptive Analytics Workflow

Last Published: Aug 05, 2021 |
Russell Andrade
Russell Andrade

Sr Principal Product Manager

We all understand the importance of deriving the right insights from our data. In order to stay competitive and thrive in today’s business climate, organizations need to have a holistic, 360-degree understanding of their customers and their operations. Sometimes, however, we can often get carried away with the whole process of collecting and interpreting our data and forget to ask a more fundamental question, “Is the underlying data correct in the first place?” This is a big challenge that is becoming even more acute, especially given the large volumes of data being generated and processed. Analyzing incorrect data can lead to making decisions that can be completely wrong. To ensure we make confident decisions from analytics, we need to have the right people, process, and technology that ensure confidence in data. And this is where a Master Data Management (MDM) system can help: as the single source of the truth for all your business-critical data, MDM is the glue that binds your systems and information together and helps provide the context you need to drive successful analytics initiatives.

master data management is your single source of truth for buisiness-critical data

Because descriptive analytics (taking historical data and providing the necessary set of capabilities in the form of charts and dashboards for interpreting this data) in particular is considered table stakes, end users expect support for this from their business systems. The reason why descriptive analytics is so powerful and valuable is because it is simple to understand and answers the question, “What happened?” This is a crucial step, because before you use AI and predictive models to ask questions of the future, you really need to understand what happened based on historical data. Ensuring your data is clean is another key step you must take before you employ any type of analytics. While most customers are aware of the powerful cleansing and matching capabilities offered by Informatica Multidomain MDM, they may not know it also provides mechanisms for data analysis.

Clean your data prior to Analysis

Once the MDM system has merged and cleansed the data from source systems (such as your CRM or accounting system) it generates the single source of truth for all your data. Your data is now ready for analysis. There are many architectural approaches you can pursue to leverage cleansed data for your analysis. One approach is to publish the data back to the source systems and leverage the analytics capabilities provided by the source systems. Another option is to route the data to separate target systems with analytics capabilities such as Tableau. Both options work and can be good practice. But in many cases, it might be simpler and more straightforward to just leverage the MDM system itself for analysis. Many organizations also leverage a hybrid approach, where they use the analytics capabilities of the MDM system for certain types of analysis, in conjunction with the capabilities of the source and target system for more advanced analysis. Either way, we understand the importance of enhancing the analytics capabilities of our product and we are working hard to meet your requirements in that area.

Customers have used Informatica MDM to build valuable operational insights from the trusted data and with each release we look to add additional support for descriptive analytics.

In release 10. 3 of Customer 360, for instance, we recently introduced out-of-box charts in Customer 360, which provide 360-degree insights on customers, source systems and tasks. Product enhancements like these are just some examples of how we are continuing to invest in our analytics capabilities in the platform so that any client application can leverage them.

Report on your Clean Data

One of the interesting challenges inherent in analyzing data within an MDM system is the fact that each customer may cleanse and consolidate different sets of data and have reporting requirements specific to their business. Even when their use cases are similar, the type of business entity data they may be interested in can be very different. A service provider, for example, may want to understand how many devices they have added in a given year while a consumer-focused business may wish to understand the number of households added per year. The way we address this is by providing out-of-box reports on all entities in the system whether it is a customer, organization, household, or some other entity specific to the business. We also provide filtering capabilities with our analysis tools so that a Data Steward can filter out entities they are not interested in and only report on entities of interest.


Figure 1 Selecting Business Entities for Visualization


At the same time, we understand that every business has unique reporting requirements that can go beyond standard use cases. So, in addition to our out-of-box reports, we allow you to report on any clean data you have centralized within our system.

Create Visuals to Gain Insights

Most of us love visuals, and no descriptive analytics capability would be complete without powerful charting and dashboarding capabilities that can be customized to your business, and this is another area where Informatica continues to invest.

Meaning that, once you have decided on the data you wish to report on, you can create charts of various types from it and customize everything from the type of chart (bar, column, pie, stacked bars, etc. ) to your chart titles, labels, fonts and tooltips. You can even change the colors to suit your personal preference or corporate branding!

When viewing a chart, we often need to drill down into a specific entry of the chart and view its details. For example, as a Data Steward, I may be viewing a chart that shows a listing of my global customers by country. While valuable in its own right, I might then want to drill down further, to my US customers and a view that lists of customers by state, and then further drill down by city. Support for drilldowns is critical in understanding “what happened,” and you can expect to see us add this in an upcoming release.

Finally, you can lay out your charts and build your own custom dashboards using our powerful layout engine. More importantly, you can feel confident that the data you are visualizing is correct, thanks to the MDM system that fuels this data.


Figure 2 Customized Home Page Chart Layout


Action your Insights

Once we have a good understanding of “what happened” we often need to act on it. For instance, as a Data Steward, I may be viewing a chart that shows me a number of operational tasks organized by task status and priority. Let’s say I notice a large number of the overdue priority 1 tasks have been assigned to my overworked colleague Bob. I can now do something about it, and perhaps reassign some of Bob’s tasks to someone else. You can’t do this from a chart, however—in order to do this, you would have to go to an operational screen like the Task Manager.

This is where deep linking comes in handy. By providing the facility to set up navigation links from the chart to operational screens such as the Task Manager or Search Results, you can navigate to those screens directly from the chart itself. You can set up your navigation links to be contextual to the data you were just viewing—in this case, when you navigate from the chart to the task manager, it would show a filtered list of Bob’s overdue tasks. You can now leverage the Task Manager to view additional details about Bob’s tasks and update them accordingly. Informatica MDM provides not only provides you with the tools to clean and understand your data, but also facilities to act on it.

Hopefully, by now you understand the importance of reporting on “clean data” and the capabilities MDM offers for data analysis. By making MDM a key component of your analytics strategy, you can eliminate any doubts about the trustworthiness of your data and instead focus your efforts on the actual analysis and execution. If you have questions, please leave me a comment. I would love to hear more about your reporting and analytics use cases. I would love to hear more about your reporting and analytics use cases.

First Published: Nov 05, 2019