How Twitter killing off the character limit could actually be great

You may have heard that Twitter, which is practically defined by its 140-character limitation on tweets, may get rid of that character limit. This is very hard to imagine, like if Starbucks stopped selling coffee or Guy Fieri shaved his goatee.

More precisely, as CEO Jack Dorsey hinted, Twitter may allow users to embed much longer snippets of text within still-short tweets. “We’ve spent time observing what people on are doing on Twitter, and we see them taking screenshots of text and tweeting it,” he wrote in a screenshot of text that he tweeted. “Instead, what if the text… was actually text? Text that could be searched. Text that could be highlighted. That’s more utility and power.”

The reaction I saw on Twitter, which was mostly power users and other long-time tweeters, was mostly that emoji that shows someone throwing away trash.

Slate’s Will Oremus and academic Zeynep Tufekci pointed out that Twitter wants more content inside Twitter. “Twitter is not removing the character limit,” Tufekci tweeted. “It’s killing the link.”

Like Facebook or Instagram, the social network (like all other ad-supported social networks) has one overarching business goal: holding your attention as often as possible for as long as possible. Because attention is money, more or less. And to that end: “Instead of funneling traffic to blogs, news sites, and other sites around the Web,” Oremus wrote, “the ‘read more’ button will keep you playing in Twitter’s own garden.”

So, let’s stipulate that full-on publishing inside Twitter would not contradict Twitter’s core business imperative.

But I think the change will play out differently than these smart observers expect—and it could make media companies (at least) happy.

Here’s why: it all comes down to data. Engineers at Silicon Valley companies, even public ones, have a lot of power. Facebook’s Instant Articles, in which media companies directly post their work to the social network instead of their/our own sites, began as an engineering quest to reduce mobile load times.

Think about that: Facebook engineers wanted stories to load a second or two faster, and now the entire digital media business has reconfigured itself around that desire.

It’s sometimes easy to assume that the business case drives the decisionmaking at these companies. But in practice, these companies care at least as much about keeping engineers happy as they do people who work in “monetization.”

And what engineers want is data. Data data data data.

Look at Jack Dorsey’s explanation of the change: “What if the text… was actually text? Text that could be searched. Text that could be highlighted. That’s more utility and power.”

Translation: What if human-readable text became machine-readable, searchable, editable data?

This is really the history of Twitter’s product evolution. They take some thing that people come up with and that humans understand like hashtags or retweets, and then they make it machine-readable. Culture becomes data. I would argue that Twitter has become a better product because of it. (Twitter may also be a worse community, but that’s not the kind of thing that it’s easy to measure.)

Now, screenshots of text are lost data. Like the load-time obsessed Facebook engineers, Twitter’s people have a point, too. Screenshots of text are dumb (and bad for the blind).

But people screenshot because it increases the engagement on tweets. People do it because other people don’t click on links very often (anymore). People do it because Twitter doesn’t let us do anything else.

Sure, Twitter could run OCR software and simply record the text in screenshots as plain text. That’s easy. But there’d be a few errors. It’d be a bit messy. And you know what else is lost data: the metadata about where the screenshotted text originally came from.

So, let me float an intriguing possibility for how the text service could run, one that might make everyone happy. (This is conjecture: I’ve contacted the company about my theory, but have yet to hear back.)

What if Twitter’s text-embed worked like a Pinterest for text? You’d have a little browser bookmarklet/action within the app that would let you highlight text and embed it into a tweet. That’d get Twitter more data: the text and the metadata about where it came. This wouldn’t encroach on the media’s reasons for Tweeting any more than a screenshort, but it would be better for users.

Would such a tool allow for direct posting? Probably, but as a secondary mode. Again the model could be Pinterest, which lets users upload images, even if that’s not how most people use it. (This would also absolve Twitter of having to build a real authoring tool, which would be a hassle.)

Intriguingly, what this tool would generate for Twitter would be a dataset of the most interesting and engaging (and data-minable) text on the Internet. That’s gotta be valuable, no? But more importantly: the engineers would love it. So much new data! So much power and utility unlocked!

There you go. Maybe Twitter effectively cutting its defining feature will go just fine.

Just don’t forget: your text is their data. The letters mean something different to humans than they do to machines.