I attended Quantcast’s Supernova Big Data Summit last Thursday in New York. Most of the talks were less about data and more business/marketing oriented (talks like “Death of the Traditional RFP: Rise of the Programmatic Specialist” were not really for me) but there were a lot of great speakers.

Chris Matyszczyk, an advertising creative consultant and tech writer, emceed the conference. Matyszczyk was hilarious and made all the talks he moderated, as well as the conference as a whole, lively and entertaining. In his opening introduction, he joked about how much he loved digital ads, building up to “I want to lick the face of the guy who created autoplay video.” (**everyone from Gawker laughs nervously**)

Morning Sessions

Data’s Impact of Commerce and Culture: Fireside Chat

Matyszczyk moderated the first fireside chat, with Lisa Archambault from Zappos and Jared Cluff from Blue Apron. This was one of my favorite talks, because it was probably the most data-focused of the day.

Advertisement

Blue Apron looks at data on how many users ordered or passed on each week’s available recipes and the user ratings of the ones they order – users can rate the recipe 1-5 and also have an open comment box to leave feedback. The data team uses this data to look for ingredient trends and correlations (which sounds like an awesome job) to inform future meal offerings.

They also do some manual data work – the chefs read every single user comment, which makes me wonder 1) how big could their user base possibly be if they’re able to do that (Cluff wouldn’t disclose this number, he only said that they ship >3 million meals per month) and 2) how are they planning to scale if/when they do grow big enough that reading all the comments personally would be unmanageable? It seems like there are a lot of interesting ways they could use with the data they have that they haven’t started exploring yet (recommended meals for you based on your preferred cuisines, ingredients, etc.)

On the business side of things, Cluff said that word of mouth is essential to the growth of their business, so they aim to keep their customers happy so that they’ll tell their friends. On measuring marketing campaign success, Blue ran a subway campaign and to track how many users they gained from it, they had a unique offer code on the subway ads, but they also put up a survey on their site asking users how they heard about them, which users actually answered.

Advertisement

Blue Apron’s sales pitch focuses on the quality of their ingredients/meals and the experience of “getting back into the kitchen,” spending time with friends and family cooking. Cluff said that they thought that talking about price isn’t as interesting as the story of good food and quality experiences. . Matyszczyk asked if this focus pushed Blue Apron upmarket, and Cluff answered that they didn’t think good food was “exclusionary;” that everyone cares about quality. I’m guessing by “everyone” he meant “wealthy people who can afford to care about quality,” but tomayto, tomahto.

Zappos has some interesting analytical work going on – the company decided a year and half ago that they were going to focus their energies on helping and retaining the “best” customer, rather than all customers. Through clustering analysis, they found that their user base fit into 7 segments, and after analyzing the behavior of those segments, the team was able to find the “best” two (she didn’t clearly specify what made them the best, but I assume it was based on spend.) The two best segments were 1) younger people who spend 30% more on themselves than on others, and 2) time-pressed, pragmatic, older professionals. They use this information to create and target better marketing campaigns - the second segment wants something straightforward or “quick and dirty”, while the first is more likely to browse and engage on the site/with marketing. Zappos even uses the segments to tailor how call center representatives approach incoming customer calls. Archambault said Zappo’s is venturing into some offline projects to help solve their best customer pain points – not just shoe-related, but in life. So stay tuned for some IRL, non-shoe related thing from Zappos, coming soon (if you’re a fashion-loving millennial or a high-spending professional, that is.)

Internally, Zappos employees are looking at data all the time – Archambault said they have a master dashboard with all the metrics they need, updating in close to real time. One of the most important metrics they look it is product sell through rate, because it points to product success, but also informs inventory decisions. They also use this dashboard to monitor the ratings they receive on their retargeting ads – Archambault said Zappos was one of the first companies to experiment with retargeting ads and they started off just flooding their customers with them. No frequency cap, no time limit, nothing. These days users can rate the ad when they see it, and if Zappos average rating gets below a certain “negativity threshold” they rejigger the targeting settings. Zappos also uses their dashboard to look at “macro trends” and adjust their strategy to prepare for them.

Product Announcement

The big reveal at this conference was Quantcast’s new product AudienceGrid. Quantcast partnered with a variety of data providers – credit card companies, TiVo Research, and more – so they can marry online and offline data to paint a fuller picture of who your audience is. It’s free for customers of Quantcast Measure, the company’s reporting and measurement tool. I think they’ll probably make money off of this by selling their collected data to publishers/advertisers for ad targeting. But it looks like a pretty useful tool, and provides some interesting insights. Apparently Gawker readers like pizza, burgers and burritos, but love coffee the most.

source: https://www.quantcast.com/gawker.com

Lunch

At lunch, waiters walked around giving out champagne in a can. I took one and then realized it was only 11:45am. YOLO.

Afternoon Sessions

Data’s Impact of Commerce and Culture: Fireside Chat

Matyszczkl moderated another fireside chat with Traci Mercer from Hootsuite on social strategy for brands and the current social landscape.

Mercer said that social is not “evolving” but “moving.” People shift which channels they’re using, but it all “comes back around.” Facebook is now your mom’s social network, so millennials are flocking to Snapchat. But eventually your mom will be on Snapchat too! Mercer warned that just because it all comes back around, that doesn’t mean brands should just stick to one channel and not try out new things – it’s important to experiment with the new and emerging social trends.

Mercer also discussed the growth of dark social – people sharing within chat/messenger apps, where what they share can’t be tracked, rather than publicly on social media, where what they share generally can be tracked. She said that this trend and the growing popularity of apps like What’sApp will lead to a merging of messenger apps into social (which we’ve already started to witness with Facebook’s acquisition of What’s App last year.) So, to everyone who copies and pastes links into Gchat, beware. Social networks are coming for you!

The Big Danger in Big Data

Peter Field and Les Binet, authors of the book The Long and Short of It spoke about about creating effective marketing campaigns and the dangers of measuring success on too short of a time scale.

Advertisement

Field and Binet looked into the performance differences between brand and activation campaigns. Brand campaigns are aimed at raising general awareness and making an emotional connection with the consumer, and activation campaigns are focused on driving sales, generally through discount codes or specific product advertisement.

Their research showed that emotional/brand campaigns actually drive more profits than activation campaigns, it just takes longer. Binet warned against the dangers of looking at short-term performance as the sole measure of success: a brand campaign may look like it failed in the short-term, and an activation campaign, with much quicker pay-offs, will look like it performed much better, causing a brand to prematurely shift strategies into activation heavy advertising. He went on to say that we don’t just need big data, we need “long data,” in order to make sure we’re not missing the longer term effects.

I worked in digital advertising for three years, and the discussion around campaign success was always along the lines of “everyone agrees that clicks are terrible measures of success!” but, in practice, advertisers would still demand them. It was refreshing to see a smart, academically rigorous study of what it means for a campaign to be successful.

The Future of Media: The Innovators of Publishing

Pat Kiernan from NY1 moderated this talk with Kate Ward from Bustle, Ashley Parrish from Today.com and Jake Horowitz from Mic. Today.com is the web companion to the Today Show and Mic and Bustle are millennial focused publishers - Mic is a news site intended for 18-35 year olds and Bustle looks to reach feminist, progressive young women. Both Ward and Horowitz said their staff is made up of their target demos, so they don’t need to wonder what millennials care about, because they are millennials (and it’s true: both of them made Forbes’ 30 under 30 in Media list – Horowitz in 2014 and Ward in 2015 - plus Horowitz looked like he might actually be 17 years old)

Advertisement

All of the publishers rely very heavily on social channels for distribution. Most of the discussion was Facebook related, but Ward mentioned the importance of being quick to adapt to new trends, citing Snapchat’s quick rise as an important player in the game to highlight the unexpected turns the social landscape can take. Publishers need to be able to adapt their strategies quickly to keep up with these changing trends.

They all spoke about their reliance on social distribution in a pretty positive light. Their attitude seemed to be: isn’t this great, publishers can now reach millions of people through social that they couldn’t before! There was no mention of the dangers of being so dependent on these distribution channels, particularly Facebook, which has a history of making and consequently breaking publishers.

73% of Today.com’s visitors are mobile exclusive, meaning they only ever view the site on mobile, which seems like a direct result of heavy social traffic (since, as Traci Mercer said earlier – mobile is “the screen for social engagement.”) Having such a large percentage of social/mobile traffic could have some downsides (how long do those users stay on/engage with the site? how does the high percentage of mobile traffic affect their ability to monetize their content?) but Parrish didn’t allude to anything negative while discussing this stat.

Advertisement

All of the publishers stressed the importance of quality content and building trust with their reader base. Today.com is careful about using superlatives in their headlines – they try not to say something is the best if it isn’t, so users are getting what they expect when they click into a story. Perusing the site, I found this gem: “Which movie tops the ‘most rewatchable’ list? You might be surprised.” You might be surprised, maybe not though - the Today show doesn’t want to mislead you. Everyone nodded aggressively when click-bait and it’s evils came up, because none of them have any clickbait on their sites, of course.

source: a sampling of posts from Mic’s facebook page

Because this was a data conference, Kiernan asked a bunch of questions about ~metrics~ and ~data driven decision making~ Mic has a set of internal metrics that they monitor closely to determine success – they look at how many of articles a user has clicked on Facebook, how many stories a user visited once they got to Mic.com, and how far down a story a user made it. They’re also going to launch surveys at the bottom of their posts, to explicitly ask users how much they liked the story. I’d be interested to see how that goes and if they are able to gather any actionable feedback in that way.

Advertisement

Parrish said that at Today.com they use data on “runaway successes” to inform their editorial strategies. One of their most visited stories was on Starbuck secret menu items – from this story, they found that their readers liked the big CPG brands and started their Restaurant Redo series, where they post recipes that recreate popular food chain meals (which honestly, look pretty good. Would make.)

Ward talked about what they learned at Bustle from experimenting with live coverage of events by covering the Oscars. The stories didn’t perform as well as they’d hoped, but live coverage of less obvious events, like the How I Met Your Mother finale, were wild successes for the site. Kiernan wondered if the Oscar failure may have been due to the saturation of media coverage surrounding the event, and Horowitz countered that maybe it was because millennials don’t watch the Oscars. Kids these days!

I’m sort of disappointed that Matyszczyk didn’t moderate this talk. I think the idea of having Kiernan as a host was an “old school media meets new media” type of thing, but the discussion could have better if someone who knew more about digital media had been asking the questions. As it was, the talk was more like a bunch of young people trying to explain what they do for a living to their dad, and all the publishers got away with pretty glossed-over answers. I think Matyszczyk could have done a better job pushing the pubs and getting a little more real-talk out of them.

Advertisement

For example, they all claimed to be about quality and to be anti-clickbait, while the only reason I know about Mic is because I literally can’t look at my Facebook newfeed without seeing at least one, if not multiple, clickbaity headlines from them. Also, I would have loved to hear a little more than just their happy-go-lucky take on their heavy reliance on Facebook. How are these publishers keeping up with Facebook’s ever-changing algorithm? What steps are they taking to ensure Facebook doesn’t completely wipe out their business?

Keynote: “Inside the MET’s Digital Transformation”

Sree Sreenivasan, Chief Digital Officer at the Metropolitan Museum of Art, gave the keynote for the conference, talking about the Met’s venture into the digital world. He was a good speaker and the talk was really interesting, but not super related to data or publishing so I’m not going to write about it in depth here. Here are some fun takaways though:

A great tweet on the Met’s real competition:

On the Met’s long road to becoming a more digital company, Sreenivasan showed this memo from a meeting in 1967. In that meeting Mr. Watson, CEO of IBM and board member of the Met, was proposing that the Met buy an IBM computer to use in the Catalogue Department. Another board member doubted “that a computer would be a time saving device.”

The Met has a Media Lab team that does research on the intersection of technology and art. This team created a great Chrome extension called Meow Met that shows you a random cat from the Met’s collection when you open a new tab, so now you can look at cats online and appreciate art at the same time.

Conclusions

I really enjoyed this conference - I would have liked it to be a bit more data-focused, but I liked most of the speakers and the talks they gave were great. Plus, you can’t beat free lunchtime champagne!