WHY AUTOMATED
DATA TESTING

You’d never imagine that turning from manual data verification to automated tests can save you so much money and unlock so much potential.

The savings and outcome enable you and your business to walk an authentic journey to data-driven culture.

TRUST

Trust in your data is essential. The business world claims to be moving towards data-driven cultures. However, more than 50 percent of companies base at least half of their regular business decisions on gut feeling or experience rather than data and information.

A multinational company needed to increase confidence in the data regarding exchange rates. As a result of frequent changes, it was necessary to implement a set of tests that cover critical financial data marts. Those checks are executed with each exchange rate update.

With automated testing, the customer has immediate insight into exchange rates and the quality of related data marts. She knows she can trust the data and can thus make business decisions without worry.

EVIDENCE

Testing is nothing else than the process of improving
your confidence through evidence!
All your data stakeholders are fundamentally interested
in an overall view of how your data is doing –
what its actual quality, reliability and availability level is,
whether it is on track or due to be ready when expected.

A client was not sure in which parts of the developed application errors occurred and was threatened with not being able to deliver by the agreed deadline. Our automated tests suite, which also included a configurable report suite with test results, overall execution summary and trends provided an immediate overview of the current status of the project.

Thanks to well-arranged reports, the client identified critical areas of the project, and could thus prioritize accordingly.

INSTANT

Receiving feedback is one of the most vital elements of making progress. The more you automate, the more feedback you get. The more feedback you get, the more you can improve. The more you improve the more resilient your business becomes.

A client was developing a data solution that contained more than 120 measures.
It was crucial the metrics be reviewed each time the app changed. With nearly 500 automated tests the app was verified over 60,000 times in 50 days.

With automated tests that are performed on a regular basis, the customer is informed daily about the accuracy of the developed data app, and can thus refocus on a daily basis, too.

In addition, more than 600 days were saved on manual testing!

PERSISTENT

Testing is the critical part of any software development project. Each individual solution has a set of test cases to specifically evaluate its unique features, functionality and requirements. Running these tests is not a one-and-done affair; we have to ensure that the program will be evaluated throughout its lifetime, necessitating easy access and repeatability in its test set.

A customer had dozens of tests that needed to be run manually after each release. We delivered a set of automated tests that can be easily repeated every day, plus applied to various development environments.

People make mistakes, automatic solutions do not. The customer has the same feedback every day and knows she can rely on it.

Btw the full test suite was also executed tonight.

IMMEDIATE

The impact of a data issue on a production system can be immeasurable. Thus the sooner we are able to detect defects the fewer resources needed to fix it. The detection needs to be quick. It needs to be immediate. As well as the remediation.

As part of a development of BI solution, we identified the need to reduce the number of production errors found in order to save time spent by the development team on support. We implemented a set of automated tests that is executed immediately upon every change made to the solution. The feedback is available in 20 minutes!

Thanks to these tests, the development team can immediately identify and thus immediately x such errors. The number of errors found in production and related cost was reduced significantly.

I NEED IT IMMEDIATELY

THE TYPICAL SCENARIOS OUR CLIENTS LIKE TO TEST AUTOMATICALLY

Typically you migrate from one system to another and barely have a clue how successful the migration is. How much has been lost during the data transfer from one system to another? You are usually very limited in the number of retries, as you are most likely heavily into retesting those migrations. It usually takes a week(s) to gain confidence, so you can release your business people to test it.

Data must never get compromised. This is rule number one in any serious data solution. In this case, you need to ensure and be sure, that a user (or set of users) have access to exactly that data they should. But more importantly they must not see the data they shouldn’t. Automated checks validating user role data access is a great booster in data security.

Data consistency is essential. Because all your applications work with data, it must always be correct, otherwise you only amass damages. It’s very common that your data get broken as soon as a new change is applied, regardless if it’s caused by the system itself or any externally dependent application. Covering your data with consistency checks permanently ensures the data are fine and were not broken by any internal or external modification.

Data solution architecture always consists of many rules, various concepts and tons of patterns. It defines how data should be extracted, loaded, transformed, checked, agged, removed, enriched, stored, reported etc. The usual goal here is to build a sustainable solution providing business value incrementally. Automated architectural cross-checks ensure the development team isn’t compromising common and agreed design patterns and therefore is not compromising the future value.

Data quality is everything. As was already mentioned there’s no value in data without trust. Data quality checks, usually focusing on data through the lens of business, ensure the data fulfill the required business criteria. It’s continuous feedback to all stakeholders that data is of a quality they all expect.