Get Ready For The Great Embeddening

Embedding isn't just for videos anymore. How the future of the internet is about a different kind of copying and pasting.

It used to be hard work to get a video to play in a browser. A site could link directly to a video file for download, but that takes readers away from the site and out of their browsers. You could, as most sites did in the 90s and early 2000s, require one of a handful of browser plugins, such as RealPlayer or Windows Media Player, to stream inline video. This worked well enough, but had its limits, too — in particular, the plugins sucked and embedding wasn't really an option.

View this video on YouTube

Few pieces of software are more maligned than Flash, but it deserves credit for two hugely important things: popularizing online video, and pioneering the embed. Unlike images or text, an "embedded" Flash video can retain its title, its metadata, and most importantly its ads. It links back to the site that it came from and therefore, in theory, to its creator. Flash video is dying, but HTML5 video players are keeping these conventions, largely because video embedding makes sense and seems good, even when it's not strictly necessary.

The text equivalent of an embed is a quote and a link; a system that in many ways favors the quoters over the creators, allowing for mistakes, poor attribution, and a near-full loss of context. Photos are trickier. Hotlinking — HTML-level embedding on an image from another site's servers — is both a terrible faux pas and now pointless in our age of super-cheap bandwidth; it also does little to maintain a visible chain of credit. There are peer-policed rules for crediting reposted photos, but all it takes is an email or a tweet or a post to one of the internet's remarkably efficient destroyers of attribution for a photo's credit and context to become irretrievable (Tumblr and Reddit are good at many, many things, and this is unfortunately one of them).

It makes sense, then, that some content companies are trying to do with text and images what YouTube did with video. The highest profile experiment in text and photo embedding is Twitter. Embedded tweets, as common as they are now, didn't exist until late 2011. Before that the best way to repost a tweet was to take a screenshot or to use a blogging platform with its own Twitter plugin, like Wordpress. Twitter wanted something more consistent and controllable, and that always led back to the main site, which is how we ended up with these:

Twitter's rationale, at the time, was simple:

Every Tweet has a story that's more than just 140 characters. It has an author, mentions @people and #topics, contains media, and has actions you can use to share or join the conversation... we believe that everyone should be able to view and interact with Tweets on the Web in the same ways you would from any Twitter client.

This didn't just presage Twitter's recent decision to essentially require people to use its embeds over their own; it laid the groundwork for a new generation of embedded text and image content.

Josh Miller, cofounder of free-floating discussion platform Branch, explained to me why his service allows and encourages direct embedding: "We don't like how conversations elsewhere online are constrained to the platforms which they take place on. How would you share an email conversation on Facebook? Or Facebook thread on Twitter? You can't." He explained that "maintaining the interactivity" was important, too. An embedded conversation that you can participate in and trace back to its home is more appealing than a quoted conversation that you can't, especially if you're the owner of the platform on which that conversation is taking place.

Twitter and Branch and Storify operate in snippets — Quora, on the other hand, allows for much longer posts. Some posts are, for all intents and purposes, news stories or features. This week, the site introduced a new tool, which lets people select any or all text and images from a Quora post and republish them in a frame — an embed, basically. Quora's official motivation here is that "many publishers have asked us for a way to [embed Quora posts] more easily," which is probably true but distracts from the real play: Quora embeds link back to Quora in a consistent and prominent way. They don't just put Quora content on other sites, they put Quora on other sites.

Embedding shortens and strengthens the attribution chain. It's fundamentally about encouraging — or in the case of Twitter, forcing — people to give credit. To a media company, credit equals eyeballs and eyeballs equal money. To a creator, credit is often all there is.

Calling Twitter and Branch and Quora "media companies" is a new and novel thing, but it's accurate. These companies *have* learned from traditional media companies. And, I suspect, media companies are going to start learning from them.

Quora is the bridge between Twitter and old media, here. Which part of the company's rationale for encouraging people to embed posts doesn't apply to the New York Times? Why, if the posts are going to be quoted anyway, wouldn't the New Yorker or the Atlantic or CNN or, I don't know, BuzzFeed, benefit from providing an embed code for its text, which keeps both the publication's name and logo, the author's byline, and perhaps even an ad? Traditional media companies might start to ask: Why is Quora doing more to to ensure credit for its unpaid contributors than we are for our expensive content? Why are we being less assertive about ads, which are our lifeblood, than Quora, which is sitting on a cloud of VC money. Copy-pasted text doesn't make ad money, and simple link doesn't always capture people. An embed does. "I think you'll see traditional media companies start to look more like Twitter, and Twitter start to look like traditional media companies," says Miller. And I think he's right.

As ridiculous as it might look now, don't be surprised if you start seeing stuff like this:

Unlike Twitter, the NYT doesn't have all-important API access to lord over potential embedders (bloggers and reporters just copy and paste from a browser anyway). But they have other tools! Like lawyers.

This could present a host of technical issues that raw copying and pasting tend to mitigate. Imagine, for example, if Imgur disappeared tomorrow. It's a small company, no funding, just a couple guys running a massive image site — it could happen! Reddit would become an index of dead links overnight. (This actually happened years ago, circa about 2005, when Imageshack nuked its database, leaving countless thousands of forums posts across the internet, which hotlinked from their servers, without images). Or – far less likely, here — if Google shut down YouTube. Chaos! The way Quora's embeds work, they, or at least parts of them, would disappear if the site went down, just like a YouTube video would.

When sites that embed die, it's messy. But what do they care? They're dead.

More imporantly, the Great Embeddening would (will?) represent a huge shift in power from the quoters to the creators, which today means less than it once did —  a typical internet user now spend time on both ends of the embedding transaction. So, too, do the companies that advocate embedding: Soon there will be no mere links on Twitter, just Cards (also known as expanded tweets). Facebook, which doesn't offer outward-Facing embeds only because it operates under the assumption that it is the One True Platform, has automatically embedded outside content this way for well over a year now. (You can't embed a Facebook post anywhere else, but you can embed almost anything on Facebook).

This symmetry helps reconcile the movement toward embedding with what Anil Dash has described as a shift from building pages to producing feeds. He commands publishers:

Start moving your content management system towards a future where it outputs content to simple APIs, which are consumed by stream-based apps that are either HTML5 in the browser and/or native clients on mobile devices. Insert your advertising into those streams using the same formats and considerations that you use for your own content. Trust your readers to know how to scroll down and skim across a simple stream, since that's what they're already doing all day on the web.

Dash imagines raw feeds of content streaming through apps; a sort of RSS for the iPhone era. This is troubling to some publishers, who, as Choire Sicha points out, would be forced to adopt types of advertising that they might not be comfortable with (hi boss!) or equipped to pull off.

Mutual embedding is creators and users meeting in the middle: apps and platforms interlocking with one another in deeper ways than they have in the past; "media companies" (a term we're probably going to need to replace soon) finally communicating in a language that, for better or for worse, has more words than just "a href="

Skip to footer