Tag: content recommendation

Posted on

Improving discovery for NFTs will amplify digital creators and marketplaces

Alexandre Robicquet Contributor
Alexandre Robicquet is co-founder and CEO of Crossing Minds, an AI-powered recommendation platform for e-commerce and content. An experienced Stanford AI scholar, he holds three Master's degrees in mathematics, machine learning and artificial intelligence.

Four years ago, my team set an ambitious goal: build the smartest recommendation platform on the market. Since then, our algorithms have empowered companies of all kinds to provide tailored product and content recommendations, all without using cookies or invasive strategies.

Our algorithms are so precise, they can leverage only a few on-site actions from a consumer to predict what else they’ll like. Ostensibly, that would seem to be the happy ending.

But NFTs are different.

The past two years have shown us that NFTs’ influence is growing, and fast. NFTs hold a tremendous amount of potential not only for buyers, but for artists as well.

The purchase of an NFT doesn’t only reflect a buyer’s taste, it also tells us what they think might be a prudent financial investment.

Yet, that potential isn’t being fully tapped. The inability to predict what NFT buyers want represents one of the main issues preventing NFTs from going mainstream and digital artists from being amplified. The challenge lies in solving the “cold start” problem for NFTs, where it’s nearly impossible for brands to provide accurate recommendations for new and anonymous users.

First and foremost, shopping for NFTs isn’t like shopping for anything else we usually buy online. Thousands of new NFTs are bought and sold every week, buyers use multiple wallets, and they tend to be completely anonymous.

Each of these attributes contradicts what we expect with online shopping. We expect a finite number of pairs of jeans, for instance, not a never-ending supply of choices. We tend to link one credit or debit card, and while most online consumers today are, in fact, anonymous, there’s still usually the option to let a site remember your preferences.

Posted on

After going public via a SPAC, Taboola acquires e-commerce marketing network Connexity for $800M

Taboola, the company that operates a popular grid-based advertising and content recommendation network across media properties, today announced an acquisition to expand its reach further into e-commerce, its first big move since going public in June by way of a SPAC: it is paying $800 million in a combination of cash and stock to buy Connexity, a marketing technology company that operates an retail- and e-commerce-focused advertising network. Connexity has been owned by Symphony Technology Partners since 2011.

The deal — coming in the form of $260 million from cash on hand, $300 million from committed debt financing and approximately $240 million through the issuance of ordinary shares to the seller — will supersize and further diversify Taboola, which currently has a market cap of about $1.9 billion and is in hot competition with another content recommendation network operator, Outbrain: the two were set to merge operations but eventually went their own ways, and Outbrain itself went public this month.

Taboola said it expects the combined company to have gross profit of over $500 million for the fiscal year ended March 31, 2021 (ex- traffic acquisition costs, or TAC), with $185 million of adjusted EBITDA for the period, with both figures growing 20% in 2021 versus 2020.

Connexity was originally called Shopzilla before rebranding, and it has over the years amassed a number of related businesses, including Become.com, Skimlinks and PriceGrabber. (Even Connexity itself was an acquisition made by Shopzilla when it was primarily a shopping search engine.) Together, it’s helped the company build out what has become a sizable network focused around the business of e-commerce.

While Taboola focuses on content recommendations and advertising that runs alongside that, and Connexity is more squarely focused on the business of e-commerce, the two have something in common. They both position themselves as viable alternatives to the big players in advertising and discovery, giving publishers and retailers another way of making revenues and finding new customers without selling out data and a cut to the Googles, Facebooks and Amazons of this world.

While keeping the landscape competitive and providing viable alternatives to beleaguered publishers sounds like a noble enough effort, there are of course potential drawbacks. Taboola’s approach has long incurred a lot of criticism for disseminating click-bait and other garbage links, and some might have an issue with the concept of now a deeper move into e-commerce and selling merchandise along side that. Some in the media industry (and within the world of journalism in particular) has long aimed to keep commercial and other vested interests at arms length from its content, and so it will be worth watching to see how and if that effort shifts as publishers continue to look for profit.

Adam Singolda, the CEO and co-founder of Taboola, very much understands the challenges that publishers face, and he sees his company as building solutions to address that. He told TechCrunch that when Taboola went public, part of its sell to investors was that it would move into newer “types” of recommendations, covering new segments, that would also further scale the bigger operation, and this is a part of that strategy.

“We believe the future of the open web is e-commerce,” he added.

Taboola today has 9,000 digital property partners, 13,000 direct advertisers and 500 million daily active users on its platform, where publishers can use the content recommendation format to recirculate their own content as well as that of other publishers and advertisers on the Taboola network.

For Connexity’s part, it covers various activities like affiliate links, influencer marketing, in-stream advertising, shopping search ads, and more. Its customers include 1,600 direct merchants, and 6,000 publishers ((Walmart, Wayfair, Skechers, Macy’s, eBay and Otto are some of the most high profile of these). And in total, it says its network has some 40,000 retail-oriented publishers that can select from a pool of 750 million product offers, and an audience of 100 million shoppers.

And in a very fragmented e-commerce world rife with challenges in keeping online consumers’ attention, it says its various activities have generated over 800 million shopping leads and in 2020 more than $2 billion in sales for its customers.

That is still a relatively small part of the pie, though. eMarketer estimates that the e-commerce media market is worth some $35 billion in the U.S. alone.

Added to that, Taboola’s bet is that the publishers it already works with are going to be getting deeper into this space as part of their own drive to maximize more revenues per visitor/reader and make their own business models more viable (alongside diversifying into paid content, paywalls, alternative advertising formats like sponsored content, events, and so on).

“62% of US publishers expect ecommerce to be one of their biggest revenue channels,” Singolda said. “I strongly believe every publisher’s leadership these days have e-Commerce as a top 3 thing they want to get into in a big way.”

Connexity is a fairly multichannel offering today — a byproduct of all the acquisitions it has made into adjacent technologies — and Taboola plans to keep it all, with “massive cross sell and upselling opportunities, bringing eCommerce to every publisher on the open web, driving higher yields, going global with e-Commerce, empowering editorial teams what to write about around e-Commerce,” Singolda said.

Connexity on its own is a substantial business, and the shift to more e-commerce in the wake of the global Covid-19 pandemic has put a focus on the different tools it has in its armory to capture attention and convert ordinary site visitors into browsers and then into shoppers. It says that in 2019 it generated $151 million of revenue, $63 million of ex-TAC gross profit and $28 million of adjusted EBITDA in 2019, with that figure increasing to $172 million of revenue, $78 million of ex-TAC gross profit and $38 million of adjusted EBITDA in 2020.

Posted on

Content discovery platform Dable closes $12 million Series C at $90 million valuation to accelerate its global expansion

Launched in South Korea five years ago, content discovery platform Dable now serves a total of six markets in Asia. Now it plans to speed up the pace of its expansion, with six new markets in the region planned for this year, before entering European countries and the United States. Dable announced today that it has raised a $12 million Series C at a valuation of $90 million, led by South Korean venture capital firm SV Investment. Other participants included KB Investment and K2 Investment, as well as returning investor Kakao Ventures, a subsidiary of Kakao Corporation, one of South Korea’s largest internet firms.

Dable (the name is a combination of “data” and “able”) currently serves more than 2,500 media outlets in South Korea, Japan, Taiwan, Indonesia, Vietnam and Malaysia. It has subsidiaries in Taiwan, which accounts for 70% of its overseas sales, and Indonesia.

The Series C brings Dable’s total funding so far to $20.5 million. So far, the company has taken a gradual approach to international expansion, co-founder and chief executive officer Chaehyun Lee told TechCrunch, first entering one or two markets and then waiting for business there to stabilize. In 2021, however, it plans to use its Series C to speed up the pace of its expansion, launching in Hong Kong, Singapore, Thailand, mainland China, Australia and Turkey before entering markets in Europe and the United States, too.

The company’s goal is to become the “most utilized personalized recommendation platform in at last 30 countries by 2024.” Lee said it also has plans to transform into a media tech company by launching a content management system (CMS) next year.

Dable currently claims an average annual sales growth rate since founding of more than 50%, and says it reached $27.5 million in sales in 2020, up from 63% the previous year. Each month, it has a total of 540 million unique users and recommends five billion pieces of content, resulting in more than 100 million clicks. Dable also says its average annual sales growth rate since founding is more than 50%, and in that 2020, it reached $27.5 million in sales, up 63% from the previous year.

Before launching Dable, Lee and three other members of its founding team worked at RecoPick, a recommendation engine developer operated by SK Telecom subsidiary SK Planet. For media outlets, Dable offers two big data and machine learning-based products: Dable News to make personalized recommendations of content, including articles, to visitors, and Dable Native Ad, which draws on ad networks including Google, MSN and Kakao.

A third product, called karamel.ai, is an ad-targeting solution for e-commerce platforms that also makes personalized product recommendations.

Taboola and Outbrain call off their $850M merger

Dable’s main rivals include Taboola and Outbrain, both of which are headquartered in New York (and recently called off a merger), but also do business in Asian markets, and Tokyo-based Popin, which also serves clients in Japan and Taiwan.

Lee said Dable proves the competitiveness of its products by running A/B tests to compare the performance of competitors against Dable’s recommendations and see which one results in the most clickthroughs. It also does A/B testing to compare the performance of articles picked by editors against ones that were recommended by Dable’s algorithms.

Dable also provides algorithms that allow clients more flexibility in what kind of personalized content they display, which is a selling point as media companies try to recover from the massive drop in ad spending precipitated by the COVID-19 pandemic. For example, Dable’s Related Articles algorithm is based on content that visitors have already viewed, while its Perused Article algorithm gauges how interested visitors are in certain articles based on metrics like how much time they spent reading them. It also has another algorithm that displays the most viewed articles based on gender and age groups.

How will digital media survive the ad crash?

Posted on

Spotify alums create Canopy content suggester that won’t steal your data

Personalization comes at a steep price. All your data gets sucked up into a company’s servers where they can do whatever they want with it. But Canopy is a new content discovery startup that’s invented impressive technology that lets it learn about you anonymously while all your data stays on your device. Built by the co-founder and CTO of Echo Nest, the music data startup Spotify acquired to power its recommendations, Canopy wants to turn privacy into a competitive advantage. It plans to equip any content app with its tech that crunches your biographical and behavior data on your phone or computer so all it sends along are clues to what you want to see or hear next.

But first, Canopy will launch its own proof of concept app early next year that suggests long-form articles and podcasts based on your taste and activity. “There hasn’t been a great solution to private discovery. We think the reason people haven’t been excited about privacy is that they haven’t seen the opportunities” says Canopy founder and CEO Brian Whitman. “We are totally changing the value exchange of the internet” adds Canopy’s head of product strategy and former Spotify Director Of Music Publishing Annika Goldman. Matrix Partners is betting on Canopy’s privacy-safe vision for the future, leading a $4.5 million seed round for the startup.

That seems wise considering Whitman built one of the world’s most beloved content recommendation engines: Spotify’s Discover Weekly. “I’ve been doing music recommendation stuff since 2000” Whitman tells me. He left in 2015, and started to become disallusioned about “how much power we had put in algorithmic decision making and personalization. All your information goes to their servers.” Facebook’s Cambridge Analytica scandal only confirmed his views. “All this data is now being used against people. You’re getting bad recommendations, bad ads, and people are being radicalized.”

A year or two ago he started discussing the idea of building a content recommendation engine that didn’t require your actual data as inputs. He came up with a solution where “Instead of sending thousands of data points to the server we can keep all that personal data on your phone” Goldman explains. “Take Spotify for example. You listen to a song. It knows where you listened to that, it know how long you listened to it, it knows what you did next — all this stuff they don’t need to know to make music recommendations. We condense and summarize all that information and send it as a single vector – a effectively a summary of things you might like and we make it impossible to reverse engineer the vector to understand the data behind it.”

She likens the system to a model of the content world on Canopy’s servers. Rather than sending it your past activity, personal info, and intentions, it just sends a set of coordinates of where you want the recommendations to go next. The 11-person Canopy team is now building out its app that will ask you questions and watch your consumption behavior to tune its suggestions. Since podcasts and longer articles aren’t owned by any one service, they’re an easy starting point for Canopy, though it eventually hopes to be content agnostic. And since it never has to suck up your data, there’s no risk of it being stolen in a breach.

That’s a big selling point for Canopy’s software-as-a-service it plans to license its tech to other apps.”Being able to build a platform that can understand your data without the liability of user data is gamechanging” Whitman declares.

Still, the biggest question facing the company is “Do people really care about privacy?” Every day we learn of a new hack attack, data exposure, or company selling our private info, but we go right on surfing. Even Facebook’s growth rate has only dipped slightly in the wake of all its privacy troubles. But Goldman believes that’s because it’s become so overwhelming that people “have a head in the sand view on privacy. ‘Hh my god, all my data is out there. I’m at risk. What do I do about it?’ Well I want to give people a way to do something about it”. Namely, trust Canopy instead of the data grabbers.

But if people can’t be tought the value of privacy, it’s hard to see partners going to the trouble of buidling in Canopy’s system. Whitman admits that services would take a modest hit to their recommendation accuracy if they adopt Canopy. He’s hoping the long-term goodwill of users will offset that. On the horizon, he predicts “there’s a great awakening of awareness.”