It’s no secret that Twitter has been in decline for months. The will-he-won’t-he-destroy-the-network drama has spawned an entire ecosystem of wobbly social media upstarts trying to fill in the vacuum left behind. And even still, Twitter soldiered on. I found it difficult to quit entirely, because there were still valuable and important conversations and voices there I couldn’t get anywhere else.

But today was different.

Today, many Twitter users started to hit an error during their morning scrolls. “Rate limit exceeded — please wait a few moments then try again.”

The problem is that waiting a few moments won’t fix the problem.

If you’d had the “good fortune” to see Twitter CEO Elon Musk’s Tweet today announcing a major feature update (behavior which he has, in the past, promised to stop) you would have seen this:

“To address extreme levels of data scraping & system manipulation, we’ve applied the following temporary limits: — Verified accounts are limited to reading 6000 posts/day — Unverified accounts to 600 posts/day — New unverified accounts to 300/day” — @ElonMusk on Twitter, 10:01 AM July 1 2023

They have now “temporarily” limited Twitter Blue users to 8000 post reads a day and all other users 800 post reads a day, based on an update a few hours later. No word on why the change or that specific number. (UPDATE: He later changed it to 1000/10,000. Clearly a “go fast and break things” mood this week.)

Let’s dig into why this time IS different — and a framework you can use to make better decisions in your own work. In this post:

• Examining the consequences
• Business implications
• Moving forward
• Impact analysis: Avoiding the Elon Trap

Elon Musk’s updated quote Tweet showing the original post about the data scraping and the 600/6000 rate limits, plus the 800/8000 revised limits. The first reply shown underneath is from the Auschwitz Memorial, who point out that these policies will silence their posts commemorating those lost.

Examining the consequences

This is where the walls come tumbling in. But why is this change so different from all the rest? Four key elements:

• The huge mental model shift this represents for Twitter users
• A decrease in value of company-curated content
• Silencing of the most prolific Twitter creators
• Weakened social graphs caused by mass unfollows to control post views

Mental model shift

No Twitter user has ever, in the history of Twitter, had to think about throttling how much they use Twitter unless it was for their own mental health. Twitter virtually invented infinite scroll behavior. In one fell swoop, that addictive loop is cut off — and in fact, that instinctive BASIC interaction loop with Twitter is now penalized, because if you go too fast or too careless you’ll burn out your posts too soon. What if you burn out your views in the morning and need a traffic update in the afternoon?

Lost value of company-curated content

Twitter has for years been trying to make “fetch” happen, metaphorically, to borrow a meme from Mean Girls. Time and time again they have foisted the For You or Home views on us, where algorithmically curated non-chronological posts are served based on what Twitter both thinks we want to see — and what they want us to see. Lately, that includes a lot of boosting of “blue checks,” or verified users.

Let’s unpack this a little.

• If you’re limiting me to 600 posts a day, they should be posts I enjoy or get value from.
• If you’re boosting a subset of your user population because they paid for it, that is at odds with the goal of showing me content I wanted to see.
• As a result, the Blue boosts make it more likely that my 600 posts are going to include content I didn’t want to see, or wasn’t relevant to me, because someone else paid to put it there. It’s as if a large chunk of the content feed is actually advertisements masquerading as content.

In a world like that, Twitter becomes almost untenable. Ostensibly, a rate-limited environment is where a really personalized algorithm would shine — but these algorithms aren’t really personalized when they’re boosting specific users due to paid status, especially when the demographics of those users may or may nor trend in a certain direction.

Furthermore, the vastly higher post view count for subscribers means that they are more likely to be replying and posting, which creates an echo chamber effect. If the group of subscribers you have RIGHT NOW is not enough to sustain Twitter with interesting enough content to draw the rest of us in, you have a Problem.

So do I or others like me risk the For You feed? No. The few times I look at Twitter, I will likely stick to my Following feed. Users will feel the need to be more in control of those precious 600 posts.

Silencing of the most prolific Twitter creators

For creators who post frequently, they may be seen as taking up too much oxygen in this new space. This is a paradox for anyone who was relying on Super Follows or other monetization tools. Twitter is, in the end, a product that relies on free content creation — but the unintended consequences of these changes is that they will dampen content creation and consumption significantly, possibly irreparably. The screenshot embedded in this post is a great example — in response to Elon’s post about rate limits, the Auschwitz Memorial (a “Verified” Blue account) replied with:

“Support memory by excluding tweets of @AuschwitzMuseum from the limits. Most of our tweets commemorate individual victims of the Auschwitz camp on their birthdays. We remind their names, their faces and their fate. They all deserve to be seen and remembered.”

Weakened social graphs

The further effects of this kind of environment is that the fun “I’ll follow them, what’s the worst that could happen” connections actually hurt or cause harm, especially for prolific posters. Since you can’t force someone to stop posting, the best you’ll be able to do is unfollow those who are “burning” my 600 posts a day too frequently. And, of course, folks who are spamming threads with unwanted replies may be more likely to get a quick mute or block because there is additional immediate harm caused just by the presence of their post on the feed.

Everyone’s reach will be dampened. Small communities might lose critical mass. Voices will be silenced. And the ease of making new connections will be replaced by a high level of scrutiny for any new follows.

Business Implications

In the end, it’s like taking the gas out of a car and hoping it runs on fumes. Will people be enticed to subscribe to Blue for more post views? Not necessarily. None of us is used to counting views, so it’s not clear to me whether 1000/10,000 views is actually a reasonable amount for an engaged Twitter user. And it’s not clear whether this comes with a reduction in manipulative patterns like resetting to the For You tab or promoting Twitter Blue users that would artificially inflate view counts with content a customer isn’t interested in.

If you add enough friction, people will leave. And Twitter made today’s choice just a few days after upstart competitor BlueSky gave an extra invite code to every active member. Mind you, it’s far from perfect there, most of all in matters of inclusion for Black folks and other marginalized identities. Even still, Bluesky quickly began slowing down today as Twitter buckled, clearly indicating a raft of untouched invites going out quickly to new users. Mastodon may be having a similar experience today.

Moving forward

There’s plenty of speculation about this change. Why this change? Why now? Why so suddenly, on a holiday Saturday? Why in a classic Elon Tweet?

The deep irony is that folks affected by this change CAN’T GO VIEW HIS TWEET. There is no other way for the app to provide news like this, and the error message incorrectly says “please wait a few moments then try again.”

That message, by the way, is how you self-DDOS, or create the conditions for a self-directed denial of service attack. When you create an outage situation and encourage folks to refresh madly, your system will be overwhelmed by incoming requests and will never get the chance to flush and catch up. (There may also be backend configuration details that are worsening the situation.) At the BARE minimum, they should have added a new error condition letting folks know not to try again until tomorrow — but we know they don’t have enough engineers left to do that.

While one cannot say for sure, it is unlikely that the “scraping” is the true root cause. One thing has motivated most of Twitter’s decisions since Mr. Musk took over, and that is money. Eagle-eyed pundits have pointed out that Twitter had a large Google Cloud contract that lapsed recently. Is this out of fiscal and logistical necessity?

Either way, Twitter has successfully pivoted from the world’s town square to a tourist information kiosk at the mall. You can get a pamphlet or two, but you’ll have to pay for the tours. Temporary or not, the types of mental model shift this will cause are severe, and each day they are left in place will be another nail in Twitter’s premature coffin.

Impact Analysis: Avoiding the Elon Trap

What can the rest of us learn from all of this?

• Always make sure your error messages are fully descriptive to avoid unintended consequences.
• CEOs are not a substitute for a technical communications and notifications system, especially in crisis.
• Impact analysis is critical, especially at scale.

When struggling with impact analysis, I recommend folks start with PICS — or it shouldn’t happen. I describe this framework for technologists looking to make informed, ethical decisions in my book, Design Beyond Devices: Creating Multimodal, Cross-Device Experiences. (You can also check out my talk on PICS, “Should You Build It”, from the Interaction 22 conference.)

PROBLEM

What problem are you trying to solve? Is it a problem worth solving? Does solving it create other problems that must be solved in tandem? Obviously, the stated problem here is “scraping”. But the proposed solution (rate limiting) does create additional problems:

• It prevents people from enjoying Twitter.
• In worst case scenarios, it can prevent people from getting mission-critical information or world news from Twitter.
• And it can prevent businesses from connecting with customers they’re paying to connect to.

Those problems must also be solved for the proposed product or solution to be effective.

INCLUSION

Who is included by your solution? Who is left behind? What does it mean to be left behind? Twitter Blue subscribers are “included”, in as much as they have a much more forgiving rate limit, and the boosts on their posts mean that they are more likely to be seen in the 1000 posts of others. But non-subscribers (and even Blue subscribers who are power users) are left behind, and in this case leaving people behind means a very harsh walled garden that will likely drive folks to other platforms, especially in times of crisis where they need world news or community support and can’t get it in real time from Twitter.

CHANGE

What is your theory of change? This one I honestly can’t answer because I don’t know what was in Elon’s head when they teed this change up for launch. Aside from reducing “scraping.” And probably reducing server costs. But the broader perspective here is important too. HOW will this change those things? At what rate? In what way? How will the reduction in rate change the behavior of the humans involved, and not just the bots in the system?

SYSTEMS

What systems are impacted by this product or service or change? In this case, the rate limiting impacts the social graph, as folks are more likely to pare down their follows to curate their feeds. It impacts algorithmic recommendations, as people will be less likely to take a chance on content they didn’t choose. It impacts advertising, as less post views means less ad insertions. It impacts corporate subscriptions, because those expensive memberships are typically made under the assumption that brands can effectively broadcast to many people at any time.

None of us can be psychic. But that voice in the back of your head isn’t wrong — you CAN make better decisions than this one. It just takes forethought, and frameworks, and a little bit of time up front.

Poetic closure for an epic journey

A screenshot of the classic “Fail Whale” from early Twitter, which would appear when Twitter reached server capacity. Poetically, it is a white whale, just as in Moby Dick.

Those of us who have been around since the late aughts remember the original Fail Whale. Held aloft by small orange birds, the Fail Whale was an error page shown when Twitter’s servers were at capacity. Poetically, the Fail Whale was a white whale, just as in Moby Dick. In the end, after years of perceived indignities lain at the Fail Whale’s feet, Captain Elon Musk has chased this white whale of server capacity and rate limits straight to Twitter’s undoing.

Oh, Twitter. Y’all can catch me on BlueSky and Mastodon and Instagram. I WANT Twitter to rise like a Phoenix and will probably go down with the ship, but it feels more like a zombie now. Just remember — when we move to another social network, we risk creating new social bubbles when it’s invite-based. Try and use your invites to make sure that Inclusion piece of PICS isn’t leaving too many people behind, so that we can make our new social homes stronger, warmer, better places than the ones we’re fleeing.

Cheryl Platz is a world renowned user experience designer, best selling author, professional actress and public speaker, and an accomplished video game developer. Her book Design Beyond Devices: Creating Multimodal, Cross-Device Experiences is available from Rosenfeld Media or your favorite online bookseller.