Holy contextual bandits batman!

Netflix is at it again – schooling us all on what personal really means.

For a long time, Netflix has been perfecting personal recommendations on what to watch. Now it’s delivering a new feature to enhance how it makes those recommendations – personalized artwork.


So OK, that’s cool enough thinking about the thousands of titles, millions of users and all the potential key art variations needed to meaningfully personalize content. But what’s equally cool is their approach to measuring the performance of recommendations. It’s basically impossible to control for all the variables behind personalized artwork to understand what works best. So Netflix employed a methodology called Contextual Bandits.


You’re going to have to read the blog post to really understand it (and then explain it to me!) but here goes: contextual bandits are a class of online learning algorithms that trade off the cost of gathering training data required for learning an unbiased model on an ongoing basis with the benefits of applying the learned model to each member context. In other words, rather than waiting to collect a full batch of data, waiting to learn a model, and then waiting for an A/B test to conclude, contextual bandits rapidly figure out the optimal personalized artwork selection for a title for each member and context.

Anyway, it’s all pretty fascinating. And you can read more about it on the Netflix tech blog.

Why It’s Hot
Netflix takes the idea of dynamic creative to a whole new level, continuing to set the bar for 1-to-1 marketing.

Petlandia – Personalized Story Book About Your Pet!

Petlandia is a service that allows users to edit an avatar to look like their pet and then puts them into a story book. I went through it with my cat, Miso.

Miso, for reference.

I started off by editing an avatar to make it like Miso. Colors and pattern seem to be the major customizable features.

Then I entered details about him and his owners (me and my girlfriend Meg).

And that was it! I had a book all about my kitty.


There’s about 30 pages in the book that I could purchase it for $30.


I browsed their site about after purchasing the book as a Christmas present to my cat (I hope he likes it!) and found that they have an app where you can make stickers for your pets to send in chats. …I had to try it.

I downloaded the app from the App Store.

There’s no logging in, so I had to make Miso’s avatar again. Look at him with his cute lil face!

I sent a sticker to Meg on Facebook, but she reminded me that we have two cats and I can’t play favorites.

So back to the app which allows you to create multiple pets to make some Allister stickers.

Meg was happy with the results.

Why it’s Hot:

  • Quick and easy way to personal merchandise
  • The app is like Bitmoji, but for pets

Here’s their site so you can make your own pet books: https://www.petlandia.com/usa/





Spotify keeps on rocking with Lastminute.com partnership

There’s always one song that brings back memories of a really great holiday. This intrinsic link between music and travel is the idea behind Lastminute.com’s new partnership with Spotify. And according to data, a large percentage of lastminute.com consumers are also Spotify users, meaning the collaboration appears to be a win-win for both brands.

The partners are producing a series of interactive maps, playlists and podcasts – each one linked to 10 different destinations.



Each city is broken down by area, with playlists bringing to life the distinct sounds of each one. For instance, East London’s playlist includes songs by local artists like Dizzee Rascal and Katy B. Alongside this, the campaign will include a series of podcasts, each featuring an international artist giving insight into the music scene of their home city. And each user will receive personalized recommendations aligned to their music and travel tastes.


Why It’s Hot
Spotify continues to set the bar for leveraging data in meaningful ways to deepen relationships with customers.

better living partying through chemistry technology…

[about 2:05-2:45 should do it]

It’s not just a clever name, PartyBOT is your “Artificial Dance Assistant”, or ADA for short. Debuted at SxSW last month, ADA learns about party-goers musical tastes, drink preferences, and social savvy. Then, it uses facial and voice recognition to monitor the room, playing tunes tailored to the interests of those who aren’t partying hard enough as determined by their expressions and conversations. As described by its creators…

“The users’ relationship with the bot begins on a mobile application, where—through a facial recognition activity—the bot will learn to recognize the user and their emotions. Then, the bot will converse with the user about party staples—music, dancing, drinking and socializing—to learn about them and, most importantly, gauge their party potential. (Are they going to be a dance machine or a stick in the mud—the bot, as a bouncer of sorts, is here to find out.)

Upon arrival at the bar, the user will be recognized by PartyBOT, and throughout the party, the bot will work to ensure a personalized experience based on what it knows about them—their favorite music, beverages, and more. (For example, they might receive a notification when the DJ is playing one of their favorite songs.)”

Why it’s hot:

Obviously this was an intentionally lighthearted demonstration of using bot and other technology to improve an experience for people. Apart from knowing you like Bay Breezes, imagine the improved relationship brands could have with their customers by gathering initial preferences and using those to tailor experiences to each individual. Many times bots are thought of in very simplistic – question/answer/solution form – but this shows how combining AI with other emerging technologies can make for a much more personally exciting overall experience.

I Bet You Guys Are Wondering Why I’m Writing About Dresses…

Hopefully that headline got you – my colleagues in the NYC office – to stop on my post even though I’m not there to present it.

Without leaving you in too much suspense, it’s because Google recently partnered with H&M’s “digital fashion house” Ivyrevel to create something called “Coded Couture”.

The fashion influencers currently (beta) experiencing this new phenomenon will end up with completely bespoke dresses, designed based on data gathered from their activities over the course of a week.

How it ostensibly works is – they will download an app (to be released more broadly later this year), which will monitor “who they are”, what they do, and where they go, and the data captured will inform a design meant to reflect their unique personalities.

In the words of one of Ivyrevel’s co-founders, “The Data Dress enables women around the world to order a dress made entirely for them, that reflects the way they live their lives.”

Why it’s hot so hot right now (pls read in Will Ferrel Mugatu voice)

The idea of creating physical things informed by digital data is yet another example of digital transforming the physical world. Most often previously, data has helped personalize digital experiences, this is obviously data informing a physical object. If this became commonplace (/when it does), instead of choosing from mass produced options, everything we wear and use could be completely tailored for our individual lives/lifestyles.

But most of all, I’m just really curious how I would look in a data dress.

What do you guys think?

Twitter Launches New Logged OFF Experience for Users Across the Globe

Every month, over 500 million people visit Twitter to see what’s happening around the world. From conversations around live events to breaking news to pandas playing in the snow, and they want to make great content accessible for everyone — even if a user is not signed in. Twitter announced improvements to twitter.com that will bring those great Tweets to even more people across the globe.

Twitter rolled out a home timeline to people across 23 countries* who visit the twitter.com homepage on their mobile devices. Before, a user could see individual Tweets but it was hard to discover stories and conversations happening on Twitter without signing in. Now, you can check out a news story as it unfolds, dive into the play-by-play discussions around a game, and then come back again to see that exchange between two rappers everyone’s been talking about. It’s real-time and straight from the source, just like the Twitter experience for those who log in.


Twitter has also expanding their refreshed twitter.com homepage on web, which is already available in the US and Japan, to every country noted below. Now, anyone can explore and discover different topics and stories as they occur, including some that are tailored just for you based on your location and activity on Twitter.



Why It’s HOT: Twitter has grown beyond driving users to sign up for an account and has expanded their relevance to all mobile and desktop web users as a credible news source. For marketers, it will mean understanding the parameters for content to be chosen by Twitter to feature on the homepage blog. Currently, Twitter has only also gone the extra mile of personalization, including the ability for users to discover content organically as well as content tailored to their location or activity on Twitter. This will dramatically impact the way in which content “lives” on the platform. Content Discoverability will exponentially extend the life cycle of a tweet. Marketers will need to adjust their content strategy to consider it’s long term relevance when appropriate.



Source: Twitter

Retailers are tailoring their Web sites and promotions for you. Just you.


After years of using customer data to fine-tune their marketing efforts to smaller and smaller groups, retailers are now making a massive and expensive effort to tailor their Web sites and promotions for an even narrower target: an audience of one. Retailers and experts say the industry is in the early stages of a push toward personalization, in which individual shoppers have different experiences on retailers’ websites and receive highly customized e-mails, coupons and special offers. The technology retailers are using now is much more sophisticated: They are studying a wider range of activities — not just the last Web sites you visited — but also whether you opened or clicked on their emails, read a company blog post or previously redeemed a certain kind of coupon.

Why It’s Hot

In March, Barneys rolled out a new website in which nearly every page —the homepage, category pages and individual product pages — feature personalized content that is served up based on data from both a shopper’s in-store purchasing and online browsing behavior. “For one million users, we want to have one million different site experiences,” said Matthew Woolsey, Barneys’s executive vice president for digital.

This combined analysis of in-store and online patterns is particularly leading edge, industry experts said, as many retailers are still struggling to funnel these data sets together. The unified data can have powerful results. For example, Barneys has learned that many of the women who buy fine jewelry in its stores have previously browsed for it online. If Barneys purely looked at these shoppers’ web browsing history, Woolsey said they might deduce, “She’s never buying anything, so let’s try something else.” In fact, by looking at this shopper’s behavior across channels, Barneys learns it is indeed valuable to keep showing her digital jewelry lookbooks: She’s interested in the products, she’s just closing the deal in person.

Some retailers, such as CVS, are also paying attention to what you’re not buying at their stores. CVS’s personalization efforts are centered around its ExtraCare program, which was used by more than 90.8 million households in the last year. Even if you haven’t bought vitamins or toothpaste at their store, their data might still determine you’re a good candidate for a coupon for those products.

Retailers are now experimenting with this personalized, digital experience and discovering the potential of this technology.

Mattel’s New Barbie is Listening to Your Kids

Mattel’s new “Hello Barbie” hits store shelves later this year and has more tricks up her sleeve than just saying hello. With the press of a button, Barbie’s embedded microphone turns on and records the voice of the child playing with her. The recordings are then uploaded to a cloud server, where voice detection technology helps the doll make sense of the data. The result? An inquisitive Barbie who remembers your dog’s name and brings up your favorite hobbies in your next chitchat.

Why It’s Hot: 
The overall concept is interesting — a toy that listens and learns your child’s preferences and adapts accordingly … a true demonstration of personalization. The obvious downside is that nobody really knows what is being done with all that data (collected from kids!) and puts a level of accountability on the company. For example, children confide in their toys – so what if a child admits that they get hit by a parent? Should Mattel be on the hook to report it?

Uber and Spotify partner up

Uber riders can now blast tunes from Spotify during their commute. After hailing a car via the handy app, you can decide what music you’re in the mood for, and when the car arrives to pick you up, it’ll already be playing inside. You’ll need to connect that paid streaming account inside Uber’s mobile software to opt in, but doing so not only sets the music beforehand, but allows you to control it for the duration of the trip. The collaborative effort is set to launch on November 21st in London, Los Angeles, Mexico City, Nashville, New York, San Francisco, Singapore, Stockholm, Toronto and Sydney with a widespread rollout in the weeks that follow. Of course, the driver will need to connect their phone to the car’s stereo for you to take advantage, but Uber says those folks are excited about sorting your playlists. If the car you hail chooses not to play music, the option to play Spotify won’t show up in the Uber app.

Source: engadget

Why It’s Hot

Could Uber be more cool? They certainly are looking at their audience and developing pretty awesome initiatives, from partnerships with Starbucks and Amex, to stunts like UberWEDDING and UberKITTENS. In this case, it’s just a little more added value into an already sound pricing and convenience strategy.

Random – Predictive Content Discovery

“Random balances relevance and serendipity. It starts by suggesting topics that might be of interest to you. You then choose and get new content in front of you – be it a a blog post, video, photo or a news article. And then you choose again to find new things, at your own pace. You do not need to type in anything. You do not need to follow anyone. You do not need to sign in. The system learns from you and also brings up stuff from the periphery of your interests. Everyone has a unique journey when exploring new things.” — Jarno Koponen, Co-Founder


Why It’s Hot?

Random uses a new discovery engine that they call “predictive discovery.”  An algorithm focused on collecting behaviors within the application instead of outside influences like social connections, search queries, or personal information.  This sort of predictive modeling has huge implications when thinking about the future of information discovery and how it is delivered to you.  Leaning less on linear delivery system to more of a non-linear, predictive and adaptive system that is tailored to your unique interests.


Gap’s New Personalization Tactics Are “Cool, Not Creepy.”

Gap is piloting an in-store tracking app of sorts next month and increasingly tailoring its web presence for individuals. It’s facing new mall competitors as well as Amazon, the master of online personalization.

So what does it mean for shoppers? Online, that means different landing pages based on your browsing history, ads targeted to where you live — it’s called “geosniffing” — and more targeted emails. As for the physical world, the company will pilot a program at “a small group” of Banana Republic stores in two markets next month that will incentivize customers to identify themselves when they walk in, either through a smartphone app or via email addresses and phone numbers.

The in-store program will be opt-in, which will help Gap avoid the pitfalls other retailers, most notably Nordstrom, faced after testing in-store tracking without customer permission last year.

To that point, a major part of Gap’s presentation focused on personalization online, where it’s seeing tremendous growth — e-commerce sales soared 21%. Much of that tailoring is coming from cookies in the average person’s browsing history, as well as information they might voluntarily provide on Gap’s websites. Recently, Gap has started modifying homepages based on who’s looking at it, something most customers probably don’t realize is happening.

If you’ve told them and your cookies show that you’re really primarily a baby shopper, then Gap would probably go and put a baby landing page in front of you rather than landing on the generic brand landing page.

Why It’s Hot

If you’ve noticed personalization have become more frequent and more accurate nowadays compared to even a couple of years ago. It’s hot because it is data-driven, saves time and allows the marketers to become more personal with their clientele.