Operational analytics uses real-time data from operational systems to inform the most immediate and appropriate action for any business situation. The data used for operational analytics typically comes from business systems such as POS (point of sale), ERP (enterprise resource planning), IoT (internet of things), and CRM (customer relationship management) systems.
Operational analytics differs from business intelligence analytics, which uses historical information and complex algorithms to produce periodic reports for strategic decision-making. Instead, operational analytics makes insights available to business users in real time so they can use them to make decisions faster and take action immediately for the most significant impact.
Because of its focus on immediacy, operational analytics can help improve any process where information comes fast and data changes rapidly. Such processes include customer support, retail merchandising, industrial manufacturing, agile development, and many others.
This page will cover:
Why is operational analytics important?
Operational analytics is designed to help organizations make faster decisions using situational awareness. It enables employees to use business systems data to respond more effectively to events in real time.
For example, in a service call center, a support agent can better decide how to handle a customer inquiry if they know the customer’s demographics, account status, previous support cases, past purchases, and geographic location. This information, accessed during a live interaction, can inform the agent of the optimal response trajectory. Is the customer a gold level account holder? Then they get preferential treatment in the queue. Have they already engaged about the issue through other channels? Then they’re more likely to be dissatisfied, so it’s best to escalate the case proactively. By minimizing time to insight, operational analytics helps an organization take the correct actions quickly and mitigates problems caused by a lack of information.
In many cases, predictive analytics are used to enhance operational analytics by predicting likely outcomes based on data. For example, in the call center scenario, a predictive algorithm could assess the caller’s likelihood to purchase an upgrade based on their account status, age, purchase history, and location. The algorithm might also recommend an offer for the agent to make in real time.
Use cases for operational analytics
Operational analytics can be applied to nearly any complex or dynamic data-driven business process. Common use cases include:
- Customer support
- Fraud risk and detection
- Retail point of sale cross-sell/upsell
- Predictive maintenance
- Marketing campaign optimization
- Supply chain management
- Manufacturing floor optimization
- Fleet management
- Hospital patient care
Operational analytics benefits
The ability to use relevant operational information in real time can benefit the business in a variety of ways.
Improved customer engagement
Based on the premise that customers respond positively to highly personalized service, many organizations apply operational analytics to customer-facing processes that guide employees through engagement steps. Making offers and recommendations based on a specific customer’s profile makes it easier to upsell that customer and create loyalty. By analyzing operational data in real time, an organization can also detect issues that cause customer dissatisfaction and take corrective action before customers are affected.
Improved business processes
By monitoring and analyzing the state of critical business systems as they operate, an organization can detect and correct issues before they become problematic. For example, an organization can address maintenance issues with a high-speed assembly line before they lead to equipment failure and costly downtime.
Using data in real time as part of an operational process may eliminate the need to gather that information manually. And by gaining timely insight into potential issues, an organization can proactively mitigate problems, keep processes running smoothly, and maximize uptime.
Faster time to action
Because of its focus on providing information in the moment, operational analytics provides vital situational awareness that allows an organization to take the most beneficial action immediately. In comparison, traditional historical analysis is less impactful because it delays action to a future time.
Operational analytics challenges
The journey to successful operational analytics can be tricky because accessing data from multiple business systems for real-time analysis presents significant challenges.
Analyzing data without impacting operational workloads
Operational analytics requires data from systems that are critical to keeping the business running. For example, you need your POS system to process transactions quickly and accurately. But, if you’re also running analytic algorithms against the data on top of every transaction, you’re likely to overload the system, slowing it down and risking issues or failure. You need a way to analyze operational data without impacting the performance of systems that produce and use it.
Real-time access to operational data
To consolidate multiple data sources and minimize the impact on operational workloads, many organizations use ETL (extract, transform, load) processes that move data into a data warehouse where it’s analyzed. While this technique can be useful for isolating analytic workloads and reducing the impact on operational systems, it significantly delays time to insight. ETL routines must be developed carefully to maintain data quality during transfer, and they can take days or even weeks to complete. What’s needed is a way to analyze operational data in place without the delays of moving it to another system.
Turning insight into action
While analytics generally excels at clarifying what has already happened, a big goal of operational analytics is to recommend what to do next. Adding predictive capabilities to the analytics workload often requires integrating another technology. Additional technology, however, makes the environment more complex and prone to delays. What’s needed is a way to incorporate predictions and recommendations into analytics without complicating the technology stack.
Couchbase Capella for operational analytics
Couchbase Capella is a cloud-native, distributed NoSQL document Database-as-a-Service (DBaaS) that combines multiple database models into a single technology. Capabilities include:
- Processing of key-value data in memory for hyper-fast responsiveness
- Distributed storage of JSON document data for flexibility and resilience
- Mobile and IoT support
- SQL query support
- Full-text search
One of the most unique features of Couchbase Capella is its built-in analytics service.
Couchbase Analytics – Isolating operational and analytics workloads without data movement
Couchbase Analytics is a parallel data management capability for Couchbase Capella that uses a massively parallel processing (MPP) architecture to deliver insights at the speed of transactions. Couchbase Analytics is best suited for running large, complex queries involving data aggregations on large amounts of data.
The Analytics Service automatically creates a shadow copy of the operational data housed in Couchbase Capella, isolating it specifically for analysis. Couchbase Analytics can also source data from AWS S3 and Azure Blob Storage. Because the Analytics Service data is inherently and automatically linked to the operational data, changes in the operational data are reflected in analytics data in real time. And because the shadow copy of data is isolated, you can query Analytics Service data without impacting operational workloads.
Read more about Couchbase Capella Analytics in this blog.
Couchbase Analytics also supports user-defined functions (UDFs), which allow you to leverage machine learning algorithms to derive powerful insights from the data. With UDFs, trained ML models are called as functions in analytics queries that can evaluate the operational data and return predictions that are added to the result.
Read more about Couchbase Analytics UDFs for predictive analytics here.
Benefits of Couchbase Capella for operational analytics
Operational query latency and throughput are protected from slowdowns caused by the analytical query workload. Capella accomplishes this without the complexity of operating a separate analytical database.
Data is always current, and no ETL is required
Couchbase Analytics uses DCP (database change protocol), a fast memory-to-memory protocol that Couchbase Capella nodes use to synchronize data. As a result, Couchbase Analytics runs on extremely current data without ETL.
Common data model
Couchbase Analytics natively supports the same rich flexible-schema document data model used for your operational data in Capella. You don’t have to force your data into a flat, predefined relational model to analyze it.
The Couchbase Capella advantage
With Couchbase Capella and the Analytics Service, your organization can have the best of both worlds: a scalable and resilient operational data platform and a fast, powerful analytics platform. Capella combines both into a single system that consumes less infrastructure and needs fewer copies of data, resulting in a lower total cost of ownership.