What is real-time analytics?

Real-time analytics helps teams make fast decisions by displaying data as soon as it’s available—often, instantly. Real-time is an innovation upon old fashioned batch analytics, which feature a long delay between when users ask a question and get a result. With real-time analytics, teams can answer questions as soon as they arise. Here’s how to use real-time analytics at work.

Colleagues reviewing real-time analytics

What are the benefits of real-time analytics?

Fast answers help teams make fast decisions. A business’s success today is often a function of how efficiently it processes data, as well as how quickly it does so in relation to competitors. Companies that can instantly detect customer issues and adjust the product to make it more appealing evolve more quickly and are more successful.

While real-time answers may seem commonplace to anyone accustomed to iPhones and the 24-hour news cycle, computing wasn’t always so fast and in some fields, still isn’t. When electronic computing emerged in the 1950s, all computing was batched. Users fed punch cards into a computer and then left the machine to run overnight. It could take hours or days to get results. When users committed an error, it could take days to discover and correct it. Modern supercomputers still work this way, as do many analytics platforms used in business, especially if they’re working with particularly large datasets.

Real-time answers let companies improve faster, in everything from product design to marketing, business analysis, and customer support. For example:

  • Product teams can query user data mid-conversation to resolve a dispute
  • Marketers can see how users react to a campaign as it unfolds, and adjust
  • Analysts can produce reports in minutes, not days
  • Customer support teams can get instant alerts for customers who run into trouble

Need real-time insights? Try Mixpanel.

Sign up

Examples of real-time analytics

Teams can apply real-time analytics to any part of the business that produces data. Here are examples of real-time analytics use cases from several industries:

A consumer app makes its launch more successful

When consumer companies debut a new or significantly updated product, every minute matters. If an app contains bugs and doesn’t function properly, the team must fix the errors quickly before they damage the app’s adoption rate. Product and engineering teams can use a real-time analytics platform to detect bugs and anomalies—sometimes, automatically—and fix them in the moment.

A credit card company can fight fraud

Credit card and financial services firms can use real-time data to fight fraud throughout their networks. When a cardholder, for example, makes an uncharacteristic purchase—say, in a foreign country, or for a large sum online—the credit card company can trigger a workflow that suspends the card and alerts the user. The team’s fraud-detection and customer service department can monitor this data in real-time to provide proactive and satisfying customer service, and to forecast trends in fraud.

A SaaS website increases conversions

SaaS companies often use their websites to acquire leads. If visitors experience errors or the pages aren’t clear, it can cut into their profits. SaaS teams can use a real-time analytics platform to monitor users who arrive at their site and track their behaviors in what’s known as a user flow, or the series of screens users follow from entry to exit. If the team detects common points of frustration—say, users click the same button many times—the team can investigate and make the graphics, text, or call-to-action clearer, increasing conversions.

A major media site recommends content

Streaming media sites and apps can use real-time data to view how users interact with their products. User-level data helps them understand individuals’ preferences and they can use this data to train a recommendation engine to suggest similar content, as brands like Amazon, Netflix, and Spotify have done.

See how a Fortune 100 media brand increased its viewership using analytics.

How to implement real-time analytics

1. Gather requirements

Before teams implement real-time analytics, they must gather the companies’ requirements. No businesses are alike, and ‘real-time’ means something different to each one. For example, while Hulu and Netflix are close competitors, their business models and analytics needs vary.

“Our value proposition to the consumer is more about completeness than freshness,” says Ted Sarandos, Chief Content Officer of Netflix. “Having the complete season is so much more valuable, in our business model, than having last night’s episode.” As such, Netflix can focus more on querying past data than Hulu, for whom it’s more important to forecast emerging trends and offer fresh content.

To perform an analytics audit, teams can ask:

  • What does ‘real-time’ mean to each team?
  • What are each team’s goals with real-time reporting?
  • What data sources matter?
  • What tools does the team already have?

2. Identify the data

Once the team knows what types of data the company wants to analyze, they can identify the data points and their sources. For example, they may ingest data from:

  • A customer-facing website or app
  • A customer relationship management system (CRM)
  • An enterprise resource management tool (ERP)
  • A third-party data provider
  • A customer support system

3. Set up the data ecosystem

The back-end of a real-time analytics system is known as its data ecosystem, or the collection of systems needed to ingest, store, and query the data. This may include a cluster of servers, known as a data warehouse, data cleansing tools, and monitoring and detection tools.

Teams typically build their data infrastructure in a hub-and-spoke model, with one central analytics application that provides most of the required functionality, with a series of bolt-on supplementary tools. This gives the infrastructure one central repository for data and reduces the chance that there are inconsistencies between different but overlapping datasets.

Teams dealing with particularly big datasets may need to invest in specialized data storage tools, known as non-relational databases. Older relational database systems struggle to retrieve information quickly because they store data in tables and columns, like in a spreadsheet. Non-relational databases can write and read data more quickly, and are ideal for big data real-time analytics.

4. Build or buy analytics

Analytics require some sort of visual user interface (UI) so teams can query their data and see results. This can range from the most basic—a text editor where users who understand data coding languages such as Python enter commands and receive results—to code-free graphical interfaces with buttons, reports, and charts.

The more user-friendly the interface, the more widely it can be used within the company, and the more the company benefits from insights. Mixpanel’s user analytics, for example, are designed for use by both experts and amateurs, and the e-signature software firm DocuSign boosted new user account creation 15 percent after granting more than 100 of its employees access.

Curious about real-time analytics? Try Mixpanel for free.

Sign up