In his just published book, Recommendation Engines, Michael Straight provides a wealth of information, insights, and counsel while sharing the lessons he learned about the power and impact or recommendation machines.

More specifically, he “defines, discusses, and details what recommendation engines are and what makes them special…[Then] places recommendations in broad historical contexts ranging from oracles and astrologers of antiquity to contemporary curators and self-help gurus…[Next he] reviews the history of recommendation engines themselves, their academic origins, commercial engines, and multitrillion-dollar global impact…[And then he explains] how recommendation engines work [and] examines the challenge of converting implicit, explicit, and side data into structures that can be algorithmically converted into recommendations…[before he] examines recommenders with user experience in mind…[and then] offers three brief incisive case studies that illustratively bundle the algorithmic and UX exposition…[before concluding] with apocalyptic/aspirational visions of possible, probable, and inevitable recommendation engines futures.”

Here’s a “baker’s dozen” of Shrage’s key points:

o “Recommendation systems make mass personalization technically and economically possible.” (Page 12)

o “The greater the recommender influence, the greater the temptation to manipulate them.” (31)

o “The history of recommendation is the history of how people pursue and perceive advice.” (36)

o “Divination may be anachronistic, but it foreshadows the role recommendation engines play in contemporary decisions. Both look for, listen to, and feel patterns that simultaneously inform and inspire.” (41)

o “Existing technologies couldn’t effectively help people prioritize their increasingly complex information needs. Managing digital document flows would be enhanced by collaboratively engaging their colleagues.” (66)

o “Automating and algorithmically aggregating collaborative filters profoundly changed how recommendations could grow.” (70)

o “‘Team-based’ similarity recommendations could reliably compute better, faster, and cheaper than ‘user-based’ ones.” (77)

o “Humans and machines alike would need to embrace new ways to learn from each other — and the data — to dramatically improve their collaborative abilities to predict what people really wanted.” (89)

o “As recommender systems become more pervasive, they increasingly frame the value and values of what — and who — they recommend.” (104)

o Algorithms transform data into relevant recommendations by finding, calculating, and ranking the most interesting correlations and co-occurences for users.” (114)

o “All recommendation engines, whether by explicit design or tacit default, are creatures of behavioral economics.”  (156)

o “Machine learning algorithms can create content, not just recommend it.” (191)

o “The point and purpose is not slavish adherence to these suggested enhancements [of specific recommendations]; it’s empowering people to literally see, hear, and feel what possible versions of themselves could do.” (229)

Schrage explains each of these insights within a context, a frame of reference, that enriches his narrative.

Those who share my high regard for this book are urged to check out Schrage’s previously published books. Also two others: Steve Brown’s The Innovation Ultimatum: How six strategic technologies will reshape every business in the 2020s, and, Daniel Kahneman’s Thinking, Fast and Slow.

* * *

A research fellow with MIT Sloan School’s Initiative on the Digital Economy, Michael Schrage’s research, writing and advisory work focus on the behavioral economics of models, prototypes and metrics as strategic resources for managing innovation risk and opportunity. His published works include the  award-winning The Innovator’s Hypothesis (MIT Press 2014),  Who Do You Want Your Customers To Become? (Harvard Business Review Press 2012), and Serious Play (Harvard Business Review Press 2000).  His new book, Recommendation Engines, was recently published by MIT Press as part of its “Essential Knowledge” series. He’s run design workshops and executive education programs on innovation, experimentation and strategic measurement for organizations all over the world. To learn more about Michael and his work, please click here.

Here is a direct link to my review of Engines of Recommendation.