To optimize the outcome of any business transaction, operational analytics are needed.
But what do we mean by operational analytics?
Operational analytics applies to the day-to-day activities of a business and relates to real-time decision-making at the moment of risk, or moment of opportunity. Throughout a business transaction there are multiple times when an important decision needs to be made.
For example, when creating a purchase requisition, a business user needs to select a reliable supplier to deliver goods on time, or choose the optimal quantity to order. These are the moments when a business user needs the support of analytics to help make real-time decisions in the flow.
Operational analytics is not just for data analysts. Operational analytics supports all roles in an organization and at all levels, from the C-suite to the shop-floor workers.
Launch the video for an example of operational analytics:
In the past, analytics were separated from operational transactional processing. This was largely due to the limits of the technology at that time. Key transactional data would be collected throughout the day and then copied to a dedicated analytics system, usually during the night when the transactional tables were not being accessed by business users. The next day, business users would have access to analytics. For strategic analytics, where long-term decision making on highly-aggregated data is needed, a delay of one day is not usually an issue. But for operational analytics, where we are making real-time decisions within individual transactions, a delay in the provision of analytics has a negative impact. For example, it might mean we have to delay the confirmation of an order, or the raising of an invoice until we have access to the analytical data that supports decision-making in these tasks.
The latest developments in technology have enabled operational analytics. One of the most impactful developments is the in-memory database. In-memory databases can process complex analytics using huge amounts of data in real-time, and present them to the business user inside a transaction. Data for analytics needs preparation. This means aggregating, filtering, sorting, calculating and converting data. These tasks are usually performed in a separate, dedicated analytical system. This means data needs to be copied from the transaction system to an analytical system. In-memory databases are able to perform these analytical preparation tasks in real-time using the same database tables that are used for transactional processing. This means that the data used in operational analytics is always live and there is no need to first copy data to a separate system.
Another technology development is the support for powerful mobile devices which means analytics can now be in the hands of any user wherever they need to make decisions. For example, analytics can be in the hands of an account manager who is visiting a customer to discuss volume discounts based on real-time customer data. Or in the hands of a warehouse operator so they can choose an optimal location for stock put-away.
Today, analytics and transactional processing are no longer separated. Transactional applications often contain embedded analytics and vice-versa. For example, a business user might begin by creating a purchase requisition transaction that is supported throughout with analytics to support key decision making . On the other hand, a business user might begin by analyzing a list of missing parts needed in manufacturing with the ability to take immediate action, perhaps by creating a requisition transaction. So you see, it works both ways.
To summarize, modern, operational analytics includes these requirements:
- Access to real-time business data
- Fast response even on large and complex data sets
- Access from any device
- Analytics provided throughout a business transaction at key moments of decision
- Direct link from analytics to transaction and vice-versa, without disrupting the work flow