Horses, Warehouses, Analytics and Making Decisions

Horses, Warehouses, Analytics and Making Decisions

Do you get your business insights straight from the horses mouth or all round the houses? Are you a batch person or a real time person? Would you rather do your big shop at the supermarket, or have it delivered to your doorstep? I like horses, real time, at my doorstep. I have put together some arguments below that support my choices.

I started with Data Warehousing roughly 20 years ago. I don’t know if that makes me a dinosaur, or “overqualified” but I have certainly been around the topic for a while. Working with data and people who both create and consume information have kept me interested in the world of Analytics, and whereas some core concepts don’t change much, the pace of change in technology keeps us all on our toes.

My first experiences with data came through writing SQL in a support role for an Oracle based operational application and a related reporting solution. I could not understand why it was necessary to have one database for the operational application, and having to transfer the same data out to another database for creating reports. I was explained that there are technical reasons with database performance and different levels of locks on tables and rows etc and that you “just don’t do it”. Technical reasons I can understand, but attitude towards doing things differently I cannot agree with.

Consider where we are today: There aren’t many technical constraints that force you to process your data in a certain way, in fact, it’s quite the opposite. However, I feel the negative attitude towards doing things differently still exists. It is difficult to let go of some core Data Warehousing principles, where you typically extract the data from a source system, transform and load it to a data store, and finally analyze the information using a tool, which might once more store the data for its own particular processing.

strategic analytics

If you are consolidating vast amounts of data with history, from various types of source systems, with data quality issues, this is still a good way of processing your data before end user consumption. The downside of this approach has always been the number of systems and processing involved, and the amount of data storage due to data replication across the systems involved in the solution.

The above example of information processing and analysis is often called Strategic Analytics. Using a Data Warehouse is not a foregone conclusion, but if you are collecting data in this fashion for analytics that support real time business decisions, there are alternative ways of doing this.

Operational Analytics support business users’ decision making whilst in the middle of a business process. The information is available real time, at key points in a process that benefit from an analytical insight. This insight is based on the user’s input, or predefined analytical paths based on the process itself. Accessing this information does not require any additional tools, it is embedded in the user interface of the operational application. The data itself is the same transactional / process data that the end user has created as part of their job function.

Operational Analytics can be further enhanced by mixing the same Live data with information from other business areas, forecasts based on previous transactions, or combined with information from Strategic Analytics sources. This information, point of analysis or a specific finding can be shared real time with a team member or highlighted as a call for action elsewhere in the organization.

live embedded analytics

In the above scenario you don’t need to move the data anywhere and there are less systems involved…which typically saves costs in support, hardware and licensing. The end user can focus on making qualified decisions with relevant data available real time, and IT doesn’t need to support data extraction processes, storage systems and databases.

This type of an approach has of course been possible for a while, but is not yet widely adopted in my opinion. If you are, for example, considering upgrading, or changing your ERP solution, you have a great opportunity to consider what type of information you need, and where you turn your data into insight.

The above examples are quite high level, and there are several excellent ways of building an analytical solution. I challenge you to think of different (simpler!) ways of providing your business with valuable information.

We have a two part Webinar on the subject of “Where and When does data turn into an asset” coming up in March, stay tuned for more information on that!

Tags: , , , ,

No Comments

Leave a reply

Post your comment
Enter your name
Your e-mail address

Before you submit your comment you must solve the following arithmetic function! * Time limit is exhausted. Please reload CAPTCHA.

Story Page