As our lives become increasingly digital and connected, the world is drowning in data. Combine this with the emerging Internet of Things (IoT) and the vast amounts of data produced only promises to continue to grow.
Enter the Algorithm Economy. As described by Peter Sondergaard, Senior Vice President of Gartner Research, the Algorithm Economy is our future where algorithms become increasingly important marketable assets that have a dramatic impact on organizations’ bottom lines. This future would include a marketplace, much like an app store, where out-of-the-box algorithms with specific functions are sold.
Examples of these marketplaces are already starting to emerge — like DataMapper, which focuses on algorithms for the agriculture industry, and Algorithmia, a more general Algorithm-as-a-Service.
The Potential Impact
For ecommerce companies, the opportunity of an out-of-the-box algorithm that provides recommendations for customers is huge. If effective, this would give smaller ecommerce companies the ability to personalize their ecommerce shopping experience to a degree traditionally reserved for giants like Amazon and Zulily.
They could purchase and integrate a recommendation algorithm-as-a-service for their site and would likely see increases in average order value and revenue. Major companies are already using custom-designed algorithms to reduce manual sorting and improve efficiency.
The Long Road Ahead
The algorithm markets as envisioned by Sondergaard will likely contain general purpose algorithms. But since most algorithms require feedback mechanisms to “learn” and optimize in order to be effective, they will still need to be refined by those with the industry and user insights to make them truly valuable for specific businesses.
A recent example that highlights the amount work and capital involved in producing an optimized machine learning algorithm is the trending news feed section on Facebook. Looking beyond the controversial outcome of the algorithm, there was still a lot of effort needed to create it.
For two years, a team of over 20 curators worked to provide input so the algorithm could “learn.” Over time, the news feed algorithm did become more effective to the point where a story’s rank is now dependent upon factors as liking, clicking, commenting and sharing and, even the amount of time users spend reading an article.
Algorithms Have Already Proven their Value
Ecommerce company Snapdeal has seen great success using algorithms. They have played a crucial role in boosting the company’s sales with 35–40% of its sales now being driven by algorithms, as compared to less than 15% a year ago. They have a team of 50 engineers writing algorithms that incorporates user data to determine the right products for the right people. For example, someone going on a holiday to Aspen may see ski gear on the home page and someone else researching a trip to Maui may see whale watching tours.
A New Source of Value
Algorithms will help ecommerce companies not only improve and personalize their user experience, but also allows businesses to make more data-driven decisions. For these reasons, ecommerce companies should be investigating how algorithms can help them optimize certain parts of their businesses from better understanding their customers’ behavior to driving more business and increasing revenues.