Operational Analytics

Operational analytics explained

Operational analytics is the process of finding insights from your data sources to improve decision-making for the daily operations of a business.

Development and product teams find insights from application usage data to inform product decisions and roll out features and improvements quickly. Marketers translate insights from website and shopping cart data to digital campaigns on the fly. Operations teams track order and shipment data to pinpoint adjustments needed on a daily basis. Support teams improve their customers’ experience by finding insights in tickets, issues, and call data.

Regardless of the use case, the data that informs these insights is being captured in applications — so companies need their operational analytics to be driven by application data. The required capabilities to achieve application-driven operational analytics include:

• Insights on fresh, real-time, or near-real-time application data with minimal complexity
• Ability to quickly adjust as data requirements change
• Efficient analysis without sacrificing application performance

Operational analytics can be challenging

Doing operational analytics on real-time data is not always easy — or fast. Most companies send their application data to a centralized data repository — likely a data warehouse or data lake — via a batch process, which means a one-time send at the end of the day or week. The data is then usually transformed within the central repository into a more usable format that is consistent with data that has come from other sources. This xtract, transform, load (ETL) process not only takes time but can be complex to set up and maintain. Often, the data is out of date by the time it has been processed, reducing a company’s ability to make operational decisions.

MongoDB makes it easy

Our document database–centered developer data platform allows you to combine data from multiple sources to create a single, refined dataset. This can be used for real-time analytics use cases, where insights from fresh data and low-latency queries are critical. MongoDB offers:

Flexible data model: Build with speed to meet market demand while maintaining agility as data requirements evolve and new data is introduced.

Aggregation framework: Surface insights faster and more easily integrate them into your apps and processes to enable better digital experiences for your customers.

Scalable platform: Ensure timing and latency requirements are met across real-time systems and applications as they grow.

Unified interface and API: Eliminate data silos so you spend more time making data work for you and less time working for your data.

Hybrid transactional-analytical processing (HTAP): Exercise greater business agility with HTAP for real-time data.

Find insights on live application data

In the simplest form of operational analytics, developers or other stakeholders want insights from a single application to help inform decision-making. The questions operational analytics can address are endless; the answers can be found by doing basic aggregations, sorts, searches, filters, etc., on a single dataset. MongoDB makes it easy to find those answers with the tool of your choice.

• Atlas SQL Interface, Connectors, and Drivers enable you and your team to use popular SQL-based tools, such as Tableau and PowerBI, for analytical queries. Make Atlas data easily accessible to data analysts and other users who prefer doing deeper analysis with a SQL-base tool without any data engineering or ETL processes.
• Atlas Charts is built for the document model and fully integrated with Atlas. It’s quick to get started, build data visualizations, and share powerful insights from with the same UI as Atlas.
• MongoDB Query API allows you to build modular, multi-stage aggregations on your data with your preferred coding language.

Combine several data sources for deeper analysis

Many insights that enhance decision-making at the operational level require blending data from multiple sources. For example, a support team will want to combine customer order data from their ecommerce site with deliveries from their shipping data to better triage customer issues.

• Atlas Data Federation allows you to seamlessly query, transform, and aggregate data from one or more MongoDB Atlas databases and AWS S3 buckets. Data Federation enables operational analytics via Atlas SQL Interface and Atlas Charts as it is fully integrated with both. Furthermore, you can convert data into an analytical format and persist it into an S3 bucket with the $out stage for downstream systems to use.

Optimize for analytics without disrupting transactional workloads

As the amount of data required for your operational analytics grows, you shouldn’t have to sacrifice simplicity and cost-efficiency to maintain performance in your analytical queries. Whether your data comes from live and tiered data from a single application, or multiple data sources federated together, MongoDB provides a couple of different levers to pull to optimize your analytical workloads.

• Atlas Database has analytics nodes with distinct infrastructure tiering. Analytics nodes have a replica data set of your primary node, but are isolated so they can never be elected to be the primary node, nor will any queries slow down the performance of your primary node. You can also choose an analytics node tier larger or smaller than the rest of the nodes in your cluster. This added level of customization ensures you’re getting the performance required for your transactional and analytical queries without over- or under-provisioning your entire cluster for the sake of your analytical workload.
• Enables isolating live application data for analytical queries, with fixed costs chosen separate from the rest of the Atlas cluster.
• Atlas Data Lake (in preview) is a fully managed storage solution that provides the economics of cloud object storage and has the ability to reformat, partition, catalog, and capture statistics about the data in order to provide the best performance when queried.
• Enables storing ingested application data inexpensively while optimizing it for analytical queries, with pay-for-usage compute.