It's an exciting time for embedded analytics – in fact, it's an exciting time in the tech world in general.
Cloud and SaaS services are in abundance – virtualization and other new strategies are making it easier to untether systems from hardware. At the same time, machine learning and artificial intelligence are adding to what digital platforms can do.
Another very exciting point on the horizon is the evolution of embedded analytics – the idea that you can get the power of analytics in a wider range of resources.
When it comes to embedded analytics, not everybody has the exact same idea of what this means. In a very general sense, embedded analytics means that user-friendly, self-service applications have analytics tools built into them, instead of the analytics being sourced through a more centralized – and often less accessible – model.
However, developers and others are expanding that definition to talk about smart platforms and new infrastructure environments that will really enable a wider range of users.
“We look at embedded analytics more broadly than just what you would typically hear,” says Catherine Frye, director of global product marketing for embedded analytics at Qlik. This, she says, describes the latest approach to offering embedded analytics to customers, and suggests that in addition to “smart apps,” embedded analytics can be built into workflows, processes, websites and portals.
How Embedded Analytics Can Drive Decision-Making
Mentioning the acquisition of point-of-sale data and the need to integrate it with other types of customer data, Frye said embedded analytics can drive predictive capabilities and help businesses understand their customers better. These analytics platforms can be used to expand a CRM system.
"It was through our platform integration with R that customers or partners could deliver predictive capabilities, in concert with embedded, to provide more targeted, relevant and ultimately more successful marketing offers back to individual customers," Frye said.
Frye gave an example of a client that provides marketing insights to venues or clubs through a customer analytics platform – a marketing platform with specialized features that handles different types of data.
“They're looking at mostly customer types of data,” Frye said. “Membership data, loyalty data, gaming data – they're combining POS data, marrying it as well with Wi-Fi information … and food and beverage data.”
The typical end user, she said, is an analyst working at a venue or club.
“The customer behavior is their endgame,” Frye said, adding that the insights can help with marketing offers that drive revenue.
All of this, she said, is designed to help key operational people within a company make faster, more accurate decisions in the moment or day, and to take action on opportunities being presented by their customers' behaviors.
“If you're an operational decision-maker,” she said, “now you have your insights that are actually relevant – and these insights should be improving and accelerating your decision-making. That's what embedded analytics provides. That's the full purpose of it.”
Also, she said, the implementation of embedded analytics is often egalitarian – it helps less tech-literate people or “non-techies” actually perform analytical work by lowering the barriers to adoption.
How Embedded Analytics Can Leverage IoT
The Internet of Things (IoT) has been a buzzword for years, but one of the key challenges of collecting all this input data is organizing and providing it to users in a useful way.
Frye pointed to the idea of a transactional interface that has an analytics component, suggesting that embedded analytics can really leverage device and sensor data to make it easier to integrate with other data types and produce analytics in near real time in an expanded IoT platform, such as with their OEM partner.
Frye explains that expanding the horizon of embedded analytics means looking at new ways to tie system components together. “We're trying to expand the way we think about embedded analytics,” she said.
What's important, however, is not so much a technical criteria for data, but the business outcomes that customers want to achieve and the ways that embedded analytics can help them get valuable insights.
“Is it about their customers? Is it about their partners? Is it about hiring … resource allocation? … That's where you get your very specific drivers and what you want to integrate analytics with,” Frye said.
In this sense, she said, the process varies – it's not a one-size-fits-all model.
“We see many types of data being aggregated,” Frye said, talking about the process of getting data into an analytical engine and then feeding it into visual resources.
The Key Ingredients of a Successful Embedded Analytics Program
The best embedded analytics software has a few key attributes: openness, scalability and a simple interface. Qlik has a rich set of APIs available that Frye said helps with integration and additional customization of the analytical applications. They are open and available with samples, tutorials and documentation.
Also, she said, it's impossible to overstate the importance of intuitive interface.
“It has to be simple and easy to use,” Frye said.
However, she added, that doesn't mean that there isn't a deliberate process involved.
“A partner or a customer, when they begin to think through and engage an EA project … first they decide they want to do this, and then it's the requirements gathering – they'll go through a period of several weeks – they'll probably already have thought through (some of the most critical issues),” Frye said, adding that these types of projects often take a huge amount of time on the front end.
“It's how you go about thinking through your embedded requirements to fine tune and just improve the insights,” she said. “The insights should be improving significantly and be actionable. You're not going to have that unless you have very fresh insights – analytics that you have right within your working environment.”