For a brief moment, algorithm-based recommendations looked so revolutionary that they were likely to make the old ways of doing things obsolete. No longer forced to follow the lowest-common-denominator tastes of artificially defined demographic segments, brands were free to speak to an audience of one. Everyone from Amazon to Spotify to Netflix could tailor their offerings to what you and you alone were likely to enjoy.
As it turns out, this wasn’t the whole story. No one denies the power of algorithms for facilitating discovery of new bits of entertainment to consume, but brands are increasingly becoming aware of their shortcomings. There are ways of reaching customers that aren’t covered by either traditional marketing efforts or individualized outreach, and these have companies, including retailers, distributors and exhibitors, embracing generalized recommendations. Either through crowdsourcing or expert opinions, they are setting themselves up as tastemakers.
Amazon’s algorithm-based recommendation system is a long-standing and effective part of its online interface, accompanying a user review feature that is widely used if not especially reliable. But the company still perceived weaknesses in its ability to connect with readers and introduce them to new books, leading the retail giant to purchase “social cataloging” website Goodreads in 2013 for an undisclosed amount. At first glance, this looks like an odd fit: one of the main features of Goodreads is its review system, which was not integrated with Amazon’s. So what does the world’s biggest online retailer have to gain by incorporating a book website with only a moderate following?
What Goodreads brings to the table is heavy engagement with a community of dedicated readers who are ready to spend money and open to discovering new books. Baked into its platform is a wide variety of user-generated features that promote products via word of mouth. Famous authors and celebrities with a toe in the publishing world post blogs and talk up their own favorite books, while thematic reading lists and discussion groups offer innovative pathways for readers to encounter new material without the aid of algorithms.
At the end of each year, the Goodreads Choice Awards are presented to the best books of the year in a number of categories as nominated by site users. These awards are indicators of quality, not simply popularity like the New York Times bestseller list; but they are awarded by everyday readers instead of elite judges, and as a result they feel more populist and accessible than the Pulitzer or National Book Award. Goodreads offers users the ability to express themselves and form communities, but from Amazon’s perspective what they are ultimately doing is selling each other books without requiring any payment for their trouble.
Music streaming services compete heavily on the strength of their algorithm-based recommendations. While most services rely on algorithms to direct listening sessions, Spotify also automatically generates the Discover Weekly playlists, a feature that has become so popular that it is now accepting brand sponsors. There are any number of other playlists on offer, some created by third-party algorithms and many curated directly by users. Independent tastemakers build entire communities on YouTube and Spotify around their selections; Spotify in particular has done well at leveraging different modes of user-generated song recommendations in much the same way that Goodreads has with books.
A newer trend in music services is the turn to podcasts as mouthpieces for critical authority. Bandcamp, which serves primarily as a vehicle for lesser-known acts to gain exposure, offers the Bandcamp Weekly podcast to help users navigate its wide range of selections and encounter the best music available. Apple Music offers podcasts like “St. Vincent’s Mixtape Delivery Service,” which sees the indie darling showcasing her favorite music with guests like Run The Jewels and Haim. Tidal runs the Tidal On Air podcast initiative, with shows such as “Rap Radar” and “En La Mira.” Rather than restricting listeners to their own tastes as algorithmically interpreted, these podcasts give them the chance to experience new flavors selected by trusted names in music.
Apple Music offers its own year-end best-of list, but in this case the selections are made by paid editors instead of listeners. While the awards are presumably tailored to established listener tastes, the effect is to offer users a standard of objectively good music that they may have missed. This element of critical judgement stands alongside user selection as a good way to introduce customers to products they were unfamiliar with, this time out of a desire to stay culturally connected.
User reviews and critical authority are both utilized as selling points by digital film services, with Netflix providing each entry with a user star rating and iTunes displaying each film’s Rotten Tomatoes score—a convenient shorthand for critical consensus. But in some cases entertainment brands use recommendations not simply to increase sales, but to work toward transforming the state of the industry. One small example is Netflix’s recent decision to direct viewers to the series “Killing Eve” following star Sandra Oh’s Golden Globe win—despite the fact that the show streams on its archenemy Hulu.
The Alamo Drafthouse movie theater chain goes out of its way to promote independent films, and it takes pride in demonstrating how much better critically acclaimed indies do in its theaters compared to their nationwide box office. In addition to effective marketing for good indie films in general, the chain has also debuted the Drafthouse Recommends initiative, a way to specifically boost underpromoted films that the chain’s executives deem worthy of wider exposure. Business Insider tracked the performance of some of the films included in this program, suggesting big box-office gains for films such as Call Me By Your Name and Eighth Grade.
The conservative stance for film exhibitors right now is to focus on guaranteed blockbuster franchises such as the Marvel Cinematic Universe and Star Wars, while astute companies might gamble on lower-budget pictures that they expect to blow up. Where theater chains may want to emulate the Alamo Drafthouse is in aggressively working to connect potential indie gems with the audiences who will fall in love with them. It’s good for independent film in general, but it doesn’t hurt revenue either; engaged viewers are more willing to drop money while in the throes of discovery.
None of this is a knock against the incredibly sophisticated algorithm-based recommendation engines that keep your favorite entertainment brands chugging along. But these services are intended to turn you on to books, songs and films that you’re predisposed to enjoy. Getting you to step outside your comfort zone and possibly find a whole new obsession requires the intervention of real humans, whether they’re a large number of fellow consumers or a small number of experts.
The sociologist Mark Granovetter introduced the idea of “weak ties” as the crucial bonds that connect different parts of a social network to each other; for example, you’re more likely to find a job through an acquaintance than through a close friend. If you imagine a network map composed of both people and the entertainment products that they enjoy, it’s possible to understand why personal recommendations work where algorithms don’t. Algorithms act like strong ties, while critics and fellow customers you’ve likely never met (but whose tastes you’ve learned to respect) are weak ties that can usher you into a new world of entertainment. This is why the human element is necessary to keep consumers finding new things to get excited about once they’ve exhausted their old list of favorites.