As charming as the image of little Amazon fairies perusing your bookshelves and credit card receipts may be, the system which offers you “people who bought this also liked…” is a lot more complicated than a simple swish of a wand.

When Amazon makes suggestions of books and movies you might enjoy, it isn’t magic. It’s based on thousands of data points which have been gathered from a range of online sources and processed by a recommendation engine.

The real technology behind these suggestions, Recommendation Engines, analyzes mountains of data from previous purchases, items which have been clicked on, rated and liked, and what other customers have bought, before making suggestions which might tickle a user’s fancy.

Recommendation Engines Boost ROI

E-commerce platforms which have rolled out this technology boast incredible results, boosting overall revenue by between 5 percent to 15 percent.

Successful big data solutions were once stuck in the domain of huge enterprises with deep pockets, who were able to hire massive teams of data scientists. But now cognitive machines make it possible to build recommendation engines in a few hours rather than in months.

From a customer perspective, these engines make their lives easier as their search time is reduced and they spend less time browsing content that isn't interesting to them.

From a company perspective, recommendation engines empower them with ‘insider tips’ which can help them enhance the user experience, and boost sales.

But to date, aside from a few forward thinking big budget enterprises, adoption across other industries has been slow. So why are enterprises not jumping on the recommendation bandwagon? And what benefits can it bring to a wide range of industries, outside the e-commerce market.

Developing Recommendation Engines

Recommendation engines offer a means of funneling targeted content directly to potential consumers based upon their previous online behavior.

However, while the benefits of recommendation engines are attractive over a range of industries, the overwhelming burden in terms of funding, time and effort to roll out a recommendation engine have traditionally scared off everyone except from huge budget giants with billion dollar valuations.

“Companies that want such a recommendation system generally have two choices: build it themselves, or use off-the-shelf technology. Building your own is risky. In addition to being expensive, a recommendation engine that isn’t very good can be even worse than not having one at all,” said Reece Pacheco CEO of, a digital media company that chose to develop its own recommendation system.

There are two basic approaches to developing a recommendation system: the collaborative filtering method or the content-based approach. Both methods are extremely complicated, difficult to scale, and complex to build.

And the tools and materials are not the most expensive components. The experts needed to build the systems, monitor the results, and make sense of the data are becoming more and more difficult to find, as the ‘tech talent shortage’ of data scientists becomes more and more of a problem all around the globe.

A study by McKinsey projects that “by 2018, the US alone may face a 50 percent to 60 percent gap between supply and requisite demand of deep analytical talent.” Put simply, it doesn’t matter if you are a billion dollar valued giant or a five man team based out of a garage.

If there are no data scientists available for hire, your options are going to be limited.

Automating Data Science

The only way to solve this problem is to ensure that the recommendation engines are cognitive in nature. Automating data science offers a new level of speed, scale and repeatability which offers a competitive advantage to enterprises.

Cognitive systems are developed to automate data science using machine learning, allowing them to ‘learn’ from data. Once a system has been ‘trained’ it can pick and choose algorithms independently, without needing teams of hard to come by data scientists to do all the hard work.

This leaves the handful of highly trained data scientist to do what they are best at, creating forward thinking data strategies and discovering amazing insights from the data served to them by the cognitive machines, thus really creating value from the raw data.

It is becoming more difficult to hire the best data talent out there. Without these essential data scientists, enterprises with outdated big data management projects are left with piles of raw, unanalyzed data.

According to a 2015 report by Gyro many enterprise big data projects stall or fail due to the insufficient skills of those involved. Usually the people involved come from IT — and are not qualified to ask the right questions and gather the necessary insights from the data.

Building cognitive systems is the future of data science and will define the future forerunners in the field. According to Gartner, by 2018, more than 50 percent of large organizations globally will compete using advanced analytics and proprietary algorithms, leading to the disruption of entire industries.

Not Just for E-Commerce

In the world of e-commerce, recommendation engines allow retailers to stay ahead of the competition and in favor with consumers. A Forrester study shows that 73 percent of customers surveyed stated they preferred a personalized shopping experience.

Learning Opportunities

But while generally linked with e-commerce, cognitive recommendation engines can boost customer engagement over a wide range of industries.

They do this by eliminating manual tasks — for example not having to segment or sort a customer base — and through reducing reliance on human teams. They can also identify the best time to target customers and improve scalability providing an ongoing, continuous mechanism for interacting with and giving sound recommendations to a client base.

In the media industry, Netflix is probably the most well-known company to build its own recommendation engine, but the tech is also used by YouTube and Google News to filter content which users are likely to engage with. Content sharing platforms like Medium and Pinterest both use similar systems to match users with media content that really interests them.

Cisco's Big Data Connections

Cisco, a world leader in enterprise network solutions, has pushed data to the forefront of its business model and developed an in-house IT Hadoop Platform which works constantly using data from a wide range of sources. Cisco farms amazing insights from huge large datasets related to customers, products, and network activity and terabytes of unstructured data such as web logs, video, email, documents and images to maintain a competitive advantage in its sphere.

"Cisco UCS CPA for Big Data provides the capabilities we need to use big data analytics for business advantage, including high-performance, scalability, and ease of management," said Jag Kahlon, Cisco IT architect.

Cisco recently acquired Jasper, an industry-leading cloud-based Internet of Things (IoT) servicewhich allows companies of all shapes and sizes to launch, manage and monetize IoT services on a global scale. By harnessing the power of cognitive data science platform Cisco wants to learn from the data gleaned from potentially millions of individual devices connected to the IoT.

Micropact's Data First Approach

In another case, Micropact, the creators of industry-leading enterprise BPM software entellitrak, a unified platform for case management and business process management application has undertaken what CTO Mike Cerniglia calls a Data-First approach.

The entellitrak platform is based on the idea that the data needs to be the core of a solution’s design.

The Data-First approach to enterprise software solutions puts the control directly in the hands of knowledge workers themselves, offering organizations a powerful tool to manage important business operations and offering an innovative solution which can quickly and easily adapt to the changing needs and requirements of users.

Ikea Harnesses Social Data

Homeware giant Ikea is also using recommendation engines to improve marketing, customer service, and product development, by harnessing the power of social data from client behavior online and on social media.

When moving into new markets, the company uses data analytics to gather data from social media platforms. Information is gathered in order to choose where to base new stores, what advertising and promotions are most suitable, and how to structure stores and display goods in a way that will boost sales.

These are high risk business decisions that could cost a company millions if miscalculated, but they are made easier by cognitive machines.

Embrace Recommendation Engines

The talent shortage is a real problem which is not going to subside any time soon, so the sooner that companies realize that their current data projects are not producing the results they need, the better. While developing recommendation engines is costly, the ROI is significant.

Companies who are slow to jump on the bandwagon are going to continue wasting piles of potentially game-changing data, and watch themselves slip behind forward-thinking enterprises with the means to discover the hidden secrets that their data contains.