Connect with us


Recommendations for all of us



If you live in a household with a communal device like an Amazon Echo or Google Home Hub, you probably use it to play music. If you live with other people, you may find that over time, the Spotify or Pandora algorithm seems not to know you as well. You’ll find songs creeping into your playlists that you would never have chosen for yourself.  The cause is often obvious: I’d see a whole playlist devoted to Disney musicals or Minecraft fan songs. I don’t listen to this music, but my children do, using the shared device in the kitchen. And that shared device only knows about a single user, and that user happens to be me.

More recently, many people who had end-of-year wrap up playlists created by Spotify found that they didn’t quite fit, including myself:

Learn faster. Dig deeper. See farther.


This kind of a mismatch and narrowing to one person is an identity issue that I’ve identified in previous articles about communal computing.  Most home computing devices don’t understand all of the identities (and pseudo-identities) of the people who are using the devices. The services then extend the behavior collected through these shared experiences to recommend music for personal use. In short, these devices are communal devices: they’re designed to be used by groups of people, and aren’t dedicated to an individual. But they are still based on a single-user model, in which the device is associated with (and collects data about) a single identity.

These services should be able to do a better job of recommending content for groups of people. Platforms like Netflix and Spotify have tried to deal with this problem, but it is difficult. I’d like to take you through some of the basics for group recommendation services, what is being tried today, and where we should go in the future.

Common group recommendation methods

After seeing these problems with communal identities, I became curious about how other people have solved group recommendation services so far. Recommendation services for individuals succeed if they lead to further engagement. Engagement may take different forms, based on the service type:

  • Video recommendations – watching an entire show or movie, subscribing to the channel, watching the next episode
  • Commerce recommendations – buying the item, rating it
  • Music recommendations – listening to a song fully, adding to a playlist, liking

Collaborative filtering (deep dive in Programming Collective Intelligence) is the most common approach for doing individual recommendations. It looks at who I overlap with in taste and then recommends items that I might not have tried from other people’s lists. This won’t work for group recommendations because in a group, you can’t tell which behavior (e.g., listening or liking a song) should be attributed to which person. Collaborative filtering only works when the behaviors can all be attributed to a single person.

Group recommendation services build on top of these individualized concepts. The most common approach is to look at each individual’s preferences and combine them in some way for the group. Two key papers discussing how to combine individual preferences describe PolyLens, a movie recommendation service for groups, and CATS, an approach to collaborative filtering for group recommendations. A paper on ResearchGate summarized research on group recommendations back in 2007.

According to the PolyLens paper, group recommendation services should “create a ‘pseudo-user’ that represents the group’s tastes, and to produce recommendations for the pseudo-user.” There could be issues about imbalances of data if some members of the group provide more behavior or preference information than others. You don’t want the group’s preferences to be dominated by a very active minority.

An alternative to this, again from the PolyLens paper, is to “generate recommendation lists for each group member and merge the lists.” It’s easier for these services to explain why any item is on the list, because it’s possible to show how many members of the group liked a particular item that was recommended. Creating a single pseudo-user for the group might obscure the preferences of individual members.

The criteria for the success of a group recommendation service are similar to the criteria for the success of individual recommendation services: are songs and movies played in their entirety? Are they added to playlists? However, group recommendations must also take into account group dynamics. Is the algorithm fair to all members of the group, or do a few members dominate its recommendations? Do its recommendations cause “misery” to some group members (i.e., are there some recommendations that most members always listen to and like, but that some always skip and strongly dislike)?

There are some important questions left for implementers:

  1. How do people join a group?
  2. Should each individual’s history be private?
  3. How do issues like privacy impact explainability?
  4. Is the current use to discover something new or to revisit something that people have liked previously (e.g. find out about a new movie that no one has watched or rewatch a movie the whole family has seen together since it is easy)?

So far, there is a lot left to understand about group recommendation services. Let’s talk about a few key cases for Netflix, Spotify, and Amazon first.

Netflix avoiding the issue with profiles, or is it?

Back when Netflix was primarily a DVD service (2004), they launched profiles to allow different people in the same household to have different queues of DVDs in the same account. Netflix eventually extended this practice to online streaming. In 2014, they launched profiles on their streaming service, which asked the question “who’s watching?” on the launch screen. While multiple queues for DVDs and streaming profiles try to address similar problems they don’t end up solving group recommendations. In particular, streaming profiles per person leads to two key problems:

  • When a group wants to watch a movie together, one of the group’s profiles needs to be selected. If there are children present, a kids’ profile will probably be selected.  However, that profile doesn’t take into account the preferences of adults who are present.
  • When someone is visiting the house, say a guest or a babysitter, they will most likely end up choosing a random profile. This means that the visitor’s behavioral data will be added to some household member’s profile, which could skew their recommendations.

How could Netflix provide better selection and recommendation streams when there are multiple people watching together? Netflix talked about this question in a blog post from 2012, but it isn’t clear to customers what they are doing:

That is why when you see your Top10, you are likely to discover items for dad, mom, the kids, or the whole family. Even for a single person household we want to appeal to your range of interests and moods. To achieve this, in many parts of our system we are not only optimizing for accuracy, but also for diversity.

Netflix was early to consider the various people using their services in a household, but they have to go further before meeting the requirements of communal use. If diversity is rewarded, how do they know it is working for everyone “in the room” even though they don’t collect that data? As you expand who might be watching, how would they know when a show or movie is inappropriate for the audience?

Amazon merges everyone into the main account

When people live together in a household, it is common for one person to arrange most of the repairs or purchases. When using Amazon, that person will effectively get recommendations for the entire household. Amazon focuses on increasing the number of purchases made by that person, without understanding anything about the larger group. They will offer subscriptions to items that might be consumed by a whole household, but mistaking those for the purchases of an individual.

The result is that the person who wanted the item will never see additional recommendations they may have liked if they aren’t the main account holder–and the main account holder might ignore those recommendations because they don’t care. I wonder if Amazon changes recommendations to individual accounts that are part of the same Prime membership; this might address some of this mismatch.

The way that Amazon ties these accounts together is still subject to key questions that will help create the right recommendations for a household. How might Amazon understand that purchases such as food and other perishables are for the household, rather than an individual? What about purchases that are gifts for others in the household?

Spotify is leading the charge with group playlists

Spotify has created group subscription packages called Duo (for couples) and Premium Family (for more than two people). These packages not only simplify the billing relationship with Spotify; they also provide playlists that consider everyone in the subscription.

The shared playlist is the union of the accounts on the same subscription. This creates a playlist of up to 50 songs that all accounts can see and play. There are some controls that allow account owners to flag songs that might not be appropriate for everyone on the subscription. Spotify provides a lot of information about how they construct the Blend playlist in a recent blog post. In particular, they weighed whether they should try to reduce misery or maximize joy:

“Minimize the misery” is valuing democratic and coherent attributes over relevance. “Maximize the joy” values relevance over democratic and coherent attributes. Our solution is more about maximizing the joy, where we try to select the songs that are most personally relevant to a user. This decision was made based on feedback from employees and our data curation team.

Reducing misery would most likely provide better background music (music that is not unpleasant to everyone in the group), but is less likely to help people discover new music from each other.

Spotify was also concerned about explainability: they thought people would want to know why a song was included in a blended playlist. They solved this problem, at least partly, by showing the picture of the person from whose playlists the song came.

These multi-person subscriptions and group playlists solve some problems, but they still struggle to answer certain questions we should ask about group recommendation services. What happens if two people have very little overlapping interest? How do we detect when someone hates certain music but is just OK with others? How do they discover new music together?

Reconsidering the communal experience based on norms

Most of the research into group recommendation services has been tweaking how people implicitly and explicitly rate items to be combined into a shared feed. These methods haven’t considered how people might self-select into a household or join a community that wants to have group recommendations.

For example, deciding what to watch on a TV may take a few steps:

  1. Who is in the room? Only adults or kids too? If there are kids present, there should be restrictions based on age.
  2. What time of day is it? Are we taking a midday break or relaxing after a hard day? We may opt for educational shows for kids during the day and comedy for adults at night.
  3. Did we just watch something from which an algorithm can infer what we want to watch next? This will lead to the next episode in a series.
  4. Who hasn’t gotten a turn to watch something yet? Is there anyone in the household whose highest-rated songs haven’t been played? This will lead to turn taking.
  5. And more…

As you can see, there are contexts, norms, and history are all tied up in the way people decide what to watch next as a group. PolyLens discussed this in their paper, but didn’t act on it:

The social value functions for group recommendations can vary substantially. Group happiness may be the average happiness of the members, the happiness of the most happy member, or the happiness of the least happy member (i.e., we’re all miserable if one of us is unhappy). Other factors can be included. A social value function could weigh the opinion of expert members more highly, or could strive for long-term fairness by giving greater weight to people who “lost out” in previous recommendations.

Getting this highly contextual information is very hard. It may not be possible to collect much more than “who is watching” as Netflix does today. If that is the case, we may want to reverse all of the context to the location and time. The TV room at night will have a different behavioral history than the kitchen on a Sunday morning.

One way to consider the success of a group recommendation service is how much browsing is required before a decision is made? If we can get someone watching or listening to something with less negotiation, that could mean the group recommendation service is doing its job.

With the proliferation of personal devices, people can be present to “watch” with everyone else but not be actively viewing. They could be playing a game, messaging with someone else, or simply watching something else on their device. This flexibility raises the question of what “watching together” means, but also lowers the concern that we need to get group recommendations right all the time.  It’s easy enough for someone to do something else. However, the reverse isn’t true.  The biggest mistake we can make is to take highly contextual behavior gathered from a shared environment and apply it to my personal recommendations.

Contextual integrity and privacy of my behavior

When we start mixing information from multiple people in a group, it’s possible that some will feel that their privacy has been violated. Using some of the framework of Contextual Integrity, we need to look at the norms that people expect. Some people might be embarrassed if the music they enjoy privately was suddenly shown to everyone in a group or household. Is it OK to share explicit music with the household even if everyone is OK with explicit music in general?

People already build very complex mental models about how services like Spotify work and sometimes personify them as “folk theories.” The expectations will most likely change if group recommendation services are brought front and center. Services like Spotify will appear to be more like a social network if they don’t bury who is currently logged into a small profile picture in the corner;  they should show everyone who is being considered for the group recommendations at that moment.

Privacy laws and regulations are becoming more patchwork not only worldwide (China has recently created regulation of content recommendation services) but even within states of the US. Collecting any data without appropriate disclosure and permission may be problematic. The fuel of recommendation services, including group recommendation services, is behavioral data about people that will fall under these laws and regulations. You should be considering what is best for the household over what is best for your organization.

The dream of the whole family

Today there are various efforts for improving recommendations to people living in households.  These efforts miss the mark by not considering all of the people who could be watching, listening, or consuming the goods. This means that people do not get what they really want, and that companies get less engagement or sales than they would like.

The key to fixing these issues is to do a better job of understanding who is in the room, rather than making assumptions that reduce all the group members down to a single account. To do so will require user experience changes that bring the household community front and center.

If you are considering how you build these services, start with the expectations of the people in the environment, rather than forcing the single user model on people. When you do, you will provide something great for everyone who is in the room: a way to enjoy something together.

This Article was first live here.

Continue Reading
Click to comment

Leave a Reply

Your email address will not be published.


Amazon is testing a service that uses its Flex drivers to deliver packages from mall-based retailers, expanding the variety of goods available for fast shipment (Matt Day/Bloomberg)



Matt Day / Bloomberg:

Amazon is testing a service that uses its Flex drivers to deliver packages from mall-based retailers, expanding the variety of goods available for fast shipment  — Inc. is testing a service that uses the company’s sprawling network of gig drivers to fetch packages from mall-based retailers and deliver them to customers.

This Article was first live here.

Continue Reading


Dinosaurs Roam The ‘Prehistoric Planet’ In Exclusive Clip From The Apple TV+ Show



Footage of Mononykus hunting in the Apple+ new show. Credit: Apple TV+

Behold: Exclusive footage from Prehistoric Planet, the upcoming Apple TV+ show that offers viewers some of the most scientifically accurate depictions of dinosaurs to ever grace the screen. This clip depicts the methodical hunting routine of Mononykus, a petite, insect-eating theropod recognizable for its massive claws.

Narrated by David Attenborough, produced by Jon Favreau and scored by Hans Zimmer, Prehistoric Planet show is a five-episode series (one episode will be released every night starting May 23) that will transport viewers to five different habitats of the Late Cretaceous period, 66 million years ago, around the same time that a meteor impact wiped them all out. In case you missed the first captivating trailer for the show, it can be found here.

This exclusive footage is from the episode on deserts, and features a Mononykus on the hunt. In the clip, the roughly dachshund-sized dinosaur inspects a withered log in a desert, before rapping on the wood using a claw on one of its stubby forelimbs. Then the dinosaur pokes a hole in the log, and slurps up termites with its prodigious tongue. This depiction is in line with one theory about Mononykus’ ecological niche: that the dinosaur hunted like modern-day anteaters and pangolins.

Prehistoric Planet is not the first attempt to simulate extinct animals for the screen, but many of the specific reptiles the show features are being depicted on screen for the very first time.

Each animals’ behavior and appearance was produced with great detail given to their actual anatomy and biomechanics, so the show depicts dinosaurs in ways you may not have imagined before—and ways that science has only recently revealed they lived. There will be footage of dinosaurs in polar regions (a paper published last year described evidence of the animals nesting near the North Pole), and sweeping shots of massive dinosaurs moving in herds.

The show aims to depict the breadth of dinosaur biodiversity as never before—from duck-billed dinosaurs as shaggy as Dr. Seuss characters, to sauropods with massive, bubble-like accoutrements lining their necks.

You can catch Mononykus, Tyrannosaurus, and many other creatures of the Cretaceous when Prehistoric Planet debuts on May 23.

This Article was first live here.

Continue Reading


Arcade1Up Pinball Cabinet Review: A Great Start




  • 1 – Absolute Hot Garbage
  • 2 – Sorta Lukewarm Garbage
  • 3 – Strongly Flawed Design
  • 4 – Some Pros, Lots Of Cons
  • 5 – Acceptably Imperfect
  • 6 – Good Enough to Buy On Sale
  • 7 – Great, But Not Best-In-Class
  • 8 – Fantastic, with Some Footnotes
  • 9 – Shut Up And Take My Money
  • 10 – Absolute Design Nirvana

Price: $599

Josh Hendrickson

As a child of the ’80s and ’90s, owning a pinball machine wasn’t something I could fathom when I was young. But if it had occurred to me, I would have loved it. With Arcade1Up’s pinball cabinets, the company tries to thread the needle between accessibility and realism. And it gets pretty darn close.

Here’s What We Like

  • More affordable than traditional pinball
  • A working plunger
  • Super loud speakers
  • Mod potential for days

And What We Don’t

  • The display could be better
  • The fans are loud
  • DMD only uses half the screen

The first thing you need to know about Arcade1Up’s pinball machines is that they are digital and not true recreations. You won’t find moving parts or an actual ball beneath the plexiglass. Instead, you’ll stare at a display and play pinball games created by Zen Studios. That’s probably for the best, though, as actual pinball machines are notoriously difficult to maintain. Arcade1Up was kind enough to send two review units, the Marvel variant and the Star Wars version. We didn’t get to test the Attack from Mars machine, but functionally all of them are identical when it comes to hardware. Only the games and artwork change.

Some Assembly Required

It’s not an Arcade1Up cabinet unless you need to do some assembly. The bad news here is the box is super incredibly heavy. Get yourself a dolly and maybe a second person on hand. The good news is, putting the cabinet together is a mostly easy (if not tedious) affair. We’ve continually praised Arcade1Up for designing cabinets that are easy to put together, and that remains the case here.

You won’t even have to build the whole thing. The main “box” housing the glass, display, and computer arrives already assembled. Your job is to build the upper box that holds the speakers and DMD (Dot Matrix Display), connect the wiring between the two boxes, and attach the legs. You will need to turn the main box glass-side down to connect the legs, so I suggest getting a blanket to lay it on so you don’t scratch the system up. And though you’ll see me flip the entire unit on my own, don’t be like me: get some help. It’s quite heavy, very awkward, and I nearly dropped it.

One nice addition is that, while the main box arrives assembled, you can still take it apart. That matters because you may need to get to the computer to apply firmware updates or replace parts if they fail (more on that later).

The External Hardware Is Great

Two pinball machines, side by side
Josh Hendrickson

Arcade1Up did a great job putting together a cabinet that feels like a real pinball machine when it comes to the artwork and materials. It’s a tough call on whether the Marvel or Star Wars artwork is better, both are excellent and I suspect you’ll lean towards the setting you like better. I can’t speak directly to the Attack from Mars system, but the pictures suggest something a bit more bland and unremarkable.

The “flipper” buttons aren’t anything special, but they get the job done. To give the feel of “real flippers,” Arcade1Up placed a bunch of solenoids inside the box to give some feedback when you hit the flippers or the ball hits some other object in the game. I wish Arcade1Up went with something a little stronger, and I’m not alone there. You can find some pretty easy mods (which I’ll get into more later) for changing the positioning of the solenoids to solve that problem.

The faux coin door is a nice touch and a good place to put the buttons you’ll need to interact with menus, along with the volume and power controls. The volume rocker feels a tad mushy, but you probably won’t interact with it much. The machine remembers where you left the volume last, and unless you need to mute for late-night sessions, you’ll probably set it and forget it. The faux coin door would be more convincing if it had the fake coin slots to go with the fake coin return, however.

I also love the spring-loaded plunger to launch the ball with. Arcade1Up could have gone with a simple push-button, but this feels more authentic. As you pull the plunger, the digital “other half” moves to match—but only to a point. Eventually, if you keep pulling, the digital plunger stops. I wish this could have been fine-tuned a little more, but it’s still better than a push button.

And I appreciate the concept of lowering the display into the unit to give it a more “real pinball” feel. That includes a large black insert designed to draw your eyes to the gameplay, and it does enhance the effect. Unfortunately, the downfall is the display itself.

The Electronic Hardware Isn’t So Great

A closeup of a pinball screen with washed out colors
The display isn’t the greatest. Josh Hendrickson

When it comes to the electronics in general, I wish Arcade1Up had gone higher quality. Arcade1Up locked the 24-inch display at 720p, and it really shows at times. Menus can be somewhat hard to read, details are fuzzy, and despite going with 720p to obtain a better frame rate, the motion can occasionally be a little jerky. The colors aren’t the greatest either—you get a more washed-out look, especially on the Star Wars cabinet, since its tables tend to use more muted colors.

The problems continue on to the DMD at the top of the unit. Arcade1Up boasts that it’s a 7.5-inch screen but neglects to mention that only half of the display actually gets used. It’s a really odd choice, and you’ll essentially get a letterboxing effect with blank areas at the top and bottom. To top that odd choice off, the images aren’t even centered on the display. I just don’t understand why Arcade1Up did this.

An even closer look at a pinball screen showing washed out colors.
Josh Hendrickson

And then there’s the computer. Inside the main box, you’ll find a single-board Android-fueled SoC running the whole show. It’s not a Raspberry Pi, but you’re on the right track with that comparison. Android is a sensible choice, as that probably made bringing the Zen Studios pinball games over an easier process.

It’s not very powerful (as evidenced by the 720p lock), yet the fan that keeps the board cool is surprisingly loud. I thought I might have an issue with my review unit, but after Arcade1Up sent me a new computer and I swapped them out, it still didn’t help.

The second pinball machine is just as loud. It’s not an eardrum-shattering noise, but it’s loud enough I turn the machines off when I’m not using them to avoid the distracting fan noise. The weak computer takes a lot longer to load games than I’d like, though, with each new screen taking at least a minute to show up. On a somewhat related note, I wish I didn’t have to choose a language every time I turned on the machine. Arcade1Up, make that a persistent menu in the machine’s settings somewhere.

Arcade1Up also packed an accelerometer into the cabinet to emulate “tilting” the pinball. In theory, you can whack it, and the game will respond accordingly. Go too far, and you get a “tilt” which forces your ball to drop, and you lose a turn. In practice, I had to really shove the machine hard to get anything to register at all. Other digital pinball machines just use a second button for tilt, and that’s probably the better way to go.

One highlight of the electronics is the speakers. They sound really good (though they could use a bit more bass). And they get incredibly loud. I never go above a “two” on the machine’s volume scale of about 10 (it doesn’t display numbers), because it just isn’t necessary. If you turn the speakers all the way up, your neighbors will know which game you’re playing.

Oh, But You Can Mod This Thing

The inside of a digital pinball machine
Josh Hendrickson

Shortly after the first pinball machine arrived, one of the solenoids stopped working. Between that and how loud the fans were, I contacted Arcade1Up, and they sent replacement parts out along with guides. I swapped both the solenoid (which fixed the issue) and the computer (which didn’t fix the fan issue). Along the way, it became really obvious how easily you could mod these pinball machines if you wanted.

Those mods can be very easy, like moving the solenoids or replacing the glass for something higher quality (though I think the plexiglass is fine), or more difficult, like upgrading the display to something nicer. You can already find dozens of videos on YouTube covering potential mods, and the vast majority of them are probably easy enough for the average person, thanks to Arcade1Up’s design.

Even better, it’s possible to mod the computer that came with the machines now. Arcade1Up released firmware to fix a few issues that were present when the machines were first released—it used to be that holding a flipper button and then using the other would cause the solenoids to not fire. Some enterprising users have customized the software Arcade1Up released to make additional changes to the computer, like loading more games (but you should own those) or increasing the screen’s resolution to 1080p.

If you’re feeling really enterprising, you could throw in a more powerful computer, but you’ll need to add a controller box. However, the fact that you can mod the machine doesn’t mean you should—I’ve seen reports that unlocking 1080p led to overheating and dead computers. Failure is always an option if you don’t know what you’re doing, but having the potential is nice.

Buy a Pinball Machine if the Price Is Right

Two pinball machines back to back
Josh Hendrickson

If you’re wondering if you should buy one of Arcade1Up’s pinball machines, the answer is yes—if you can find it at the right price. When they first came out, these pinball machines they launched in the $600 territory. But not long after, the price increased to $800. Since then, we’ve seen it almost always on sale for $600. I can’t recommend the pinball machines at $800, but $600 is definitely reasonable. It’s a lot of hardware and includes a mixture of metal, MDF, and electronics, along with some great artwork. If you find one for less than $600, get that thing in your cart as fast as you can.

As to which one you should get, you probably instinctively know the answer. If you prefer either Star Wars or Marvel over the other, that’s probably the direction you should go. I like both franchises equally, but I’ll admit to liking the Marvel pinball games a little better. I prefer the way the tables play, and I think the colors look better on the washed-out screen. But that’s subjective.

However, if you want a pinball machine that feels the most like the real thing, spring for the Attack from Mars cabinet instead. The Marvel and Star Wars games show their “made for an iPad” colors and frequently break the illusions by having characters walk around or zooming to a particular section of the table. Attack from Mars tables play closer to the real thing.

And personally, I really hope Arcade1Up continues on with more pinball options. If the company releases a “second generation” version with better electronics (especially that display) and quieter fans, Arcade1Up might just have a perfect hit.

Here’s What We Like

  • More affordable than traditional pinball
  • A working plunger
  • Super loud speakers
  • Mod potential for days

And What We Don’t

  • The display could be better
  • The fans are loud
  • DMD only uses half the screen

This Article was first live here.

Continue Reading