Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Twitter Gave Meerkat Two Hours’ Notice before Cutting Access (fastcompany.com)
177 points by byoogle on March 15, 2015 | hide | past | favorite | 75 comments


In case you are wondering what Meerkat actually did via Twitter's API:

- Auto-follow @appmeerkat on twitter when you sign up

- Auto-enable push notifications for @appmeerkat's tweets when you sign up

- Auto-tweet when you schedule a stream

- Auto-tweet when you start streaming

- Auto-tweet when you comment on a stream

This is the entire disclosure they offered: "Everything that happens on Meerkat happens on Twitter"


At first I thought this was a bad move and would make developers wary. But Meerkat's use of Twitter was pretty bad. Not only all of the above but their tweets are completely useless within seconds or minutes and just litter tweet streams. I don't even see why all the tweeting is that necessary. I get all my Meerkat cues via notifications (which I'm about to turn off for useless-ness).


Yeah the real question is why they got a phone call with any heads-up, instead of immediate account revocation.


I don't know, maybe because warning a company / person that they're doing something you don't like and giving them the chance to fix it is decent behavior?

Or did we remove the concept of decency from our business vocabularies already?


Warning spammers or deliberate ToS violators is now a matter of decency? The company admitted they knew they were in the wrong and that Twitter would eventually shut it down. I'm not sure how it's remotely a matter of decency when it comes to intentionally malicious actors.

On reflection, it's almost certainly because the app was getting some press and Twitter wanted to make it clear what the situation was. Versus allowing the offender to pretend it was another issue or spin it against Twitter (though they seem to be trying!)


How did they manage to do this? Aren't there proper permissions requests like Facebook?


Twitter has two primary methods of authenticating API access:

1. Web version - typical Facebook style, describes specific permissions but there are only 2 options that devs can use: read or read/write (plus read/write + direct messages but that will be gone this year)

2. iOS version - menu slides in with all of your system authorized twitter accounts, you tap the one you want

I'm pretty positive #2 does not give you any real authorization messages or screens. I just logged out of Meerkat to find out but that screen doesn't work anymore because their keys are broken


I think they prompt you follow, not auto-follow. At least I don't see @appmeerkat in my followed.


Pretty sure they set it to autofollow by default.


100% sure it's auto follow - I went searching for their twitter to see if they had any news, and was surprised to find out I'm already following them.


I guess I was wrong. I must have immediately looked at my settings after signup and turned that off.


The backstory here is that Meerkat downloaded users Twitter friend lists (or "the Twitter social graph") and used it directly to send unsolicited messages to them?

If so: burn them with fire.


Meerkat legitimately broke Twitter rules with their autotweeting of replies to in-app comments without consent: https://support.twitter.com/articles/76915-automation-rules-...

> "The @reply and Mention functions are intended to make communication between users easier, and automating these processes in order to reach many users is considered an abuse of the feature. If you are automatically sending @reply messages or Mentions to many users, the recipients must request or approve this action in advance. For example, sending automated @replies based on keyword searches is not permitted. Users should also have an easy way to opt-out of your service (in addition to the requirement that all users must opt-in before receiving the messages)"

Although this behavior is still in the app I believe. Regardless, this, along with the [LIVE NOW] spam, has resulted in me unfollowing/muting everyone who uses Meerkat on Twitter, which has significantly reduced the effectiveness of Twitter for me, and which is why I think Twitter is in the right for wanting to restrict content.


I agree. Something like a "Streams" tab in the app would be awesome. Once Twitter (most likely) rolls out this feature natively with its Persicope acquisition, it would make it a lot less spammy than the way Meerkat handles it. I want to be able to differentiate between tweets and live streams.


Meerkat makes it fairly clear that "everything that happens on Meerkat also happens on Twitter" – and the comments on a stream only go to Twitter as replies to the original invite tweet.

That's not 'automatically… to many users" but specifically to those viewing a thread, and only when individual users compose individual replies. It's not a clear violation of the quoted policy… if at all. And as you note, that's not the behavior Twitter has started.


> "everything that happens on Meerkat also happens on Twitter"

That doesn't /mean/ anything.


Seems to me that the most obvious interpretation of that statement – that each action on Meerkat generates a Tweet – is also the correct one.


> That's not 'automatically… to many users" but specifically to those viewing a thread, and only when individual users compose individual replies.

Correct, but due to the Twitter mechanic where replies are shown in your feed if you are following both people (which was the case for me as the initial users are techies), it can still be a mess.


But this is growth hacking, don't you know?


Pretty sure you only get Meerkat messages if you install the app or enter/comment-on the streams. That might reasonably be an 'opt-in', depending on the descriptions offered at each step.

I'm following a lot of people who've announced Meerkat streams, and I've watched a Meerkat stream via the web – but I've not received any unsolicited messages.


"Pretty sure you only get Meerkat messages if you install the app or enter/comment-on the streams"

Nope, my twitter stream is buried with |LIVE NOW| tweets from people who are using Meerkat. I'm hoping that will stop now.


It will not. Twitter only restricted access to the social graph.

If the people you follow on twitter want to send out meerkat tweets that's their own business.


It's the meerkat app that sends out the tweets. So, you are saying that the meerkat app still has the ability to let all the followers of an individual know that a live stream is on? Seems like meerkat still has "broadcast access" to the twitter social graph then.


By "access to the social graph" I meant "access to contents of the social graph."

Meerkat can send a tweet. It can no longer examine the social graph to determine who will see it.


Now they will be Periscope tweets :-)


Fully agreed. That sounds like very bad behavior.


AFAIK they did not, in fact, do this. Here is what they did do:

1) Say that you follow me on twitter.

2) You download and run meerkat

3) I download and run meerkat. At this point you get a push notification from meerkat indicating that I've done this.

4) I start streaming on meerkat. At this point you receive a push notification from meerkat indicating that I'm doing this.

At no time do I get anything about you (unless I'm also following you on twitter).

I don't think meerkat did anything user hostile. But they did do something that was hostile to twitter (given that twitter is apparently launching a competitor) so it's no shock that they are getting cut off from the social graph.


5 minutes ago I opened Twitter to find my stream had items that Meerkat was retweeting, of people I don't follow. Good strategy to increase usage, but I unfollowed and have now deleted the app.


Disagree. I think they abused Twitter and made twitter feeds noticeably worse. All their tweets are essentially useless after a few moments and just litter feds. I may be naive but I think if Meerkat was closer to following the rules and respecting user feeds it wouldn't have been cutoff (i.e., Twitter's supposed competitor was not the reason).


Given that twitter is a stream of a whole bunch of tweets each one of which you can read really really fast, I've always been a bit confused by complaints about useless tweets. Some fairly big % of your stream is always going to be useless tweets. Has Meerkat really increased that percentage all that much?

I can see why people might disagree with me here though.


> Some fairly big % of your stream is always going to be useless tweets. Has Meerkat really increased that percentage all that much?

Sounds like it. I unfollow people sending too large a % of useless tweets, even if I find some of their feed interesting. And those were hand-written tweets written without the benefit of spammy automation!


They have, by far, the shortest shelf-life of all the tweets in my feed. It would be one thing if they were published by Meerkat, quite another that they are "published" be people I follow.

And actually, I find the signal:noise ratio in my stream to be very high.


I still haven't managed to click on a stream that was still going on. (somehow even clicking on the streams that are happening in X minutes but didn't)


If feels like Meerkat could be an order of magnitude more powerful if streams were saved for at least some amount of time. That's gotta be on their roadmap somewhere.


Yeah. Could still be "ephemeral" but give us 5 or 10 or 30 minutes or 24 hours. I don't think that would compromise the live-ness" of it, whatever that really connotes.


I really really don't want a push notification from meerkat telling me about everyone i follow on twitter doing something. That would be really really awful. I'm glad I didn't in fact sign up for meerkat.


That's usually where you choose to disable push notifications for the app within settings. It's pretty simple.


A counter argument: Uber breaks rules, a large number of people have no issue with it and consider it fighting an unfair regulatory system. These guys break the rules on Twitter and everyone sides with Twitter.

Now, I'm on the side of Twitter in this one, but I think it's an interesting thought experiment. To me the difference is that they ended up bothering Twitter's users (as seen by the comments on this thread), whereas Uber just bothered cab drivers, cab company owners and medallion holders - a relatively small audience that has pissed off a lot people.

There's also the issue of whether Twitter has a monopoly in the same sense that the cab companies do.


Twitter = our kind of people. Cab drivers = other people.

Also, the fact that Twitter can protect itself and cab drivers need the government to protect their rights goes against the libertarian ideology of HN's majority.

And don't forget that outside our bubble, Uber pissed off a lot of people that favor labor and and consumer protection and aren't big fans of American corporations hailing the return to early 20th century exploitation as "innovation".


One of the key differences is public vs. private. These are requirements from the government (town, city, municipality, county) and should reflect the will of the people. People have more rights on a public level and fairly assume that people should be able to transact without much scrutiny. Drivers want to make extra money, riders what transparent, available cheap transportation. A small number of rent-collectors are worse off in this arrangement, but broadly both sides of the market reap reward.

In the above case, Twitter is a private company, and has a right to do whatever it wants with it's service. It can get away with as much as it's users will tolerate. As a public company it has shareholder requirements, sure, but in this case it was a technology/UX problem. Meerkat was doing something that most seem to think devalued peoples user experience, and abused the purposes of implemented features. Meerkat gained a lot of exposure, and I am sure some people did benefit by it, but by and large the beneficiary was Meerkat gaining exposure by broadcasting pre-made content (spamming).


Because Meerkat made Twitter noticeably worse. All its notifications are useless in a matter of seconds or minutes (vs revolutionizing transportation to tremendous citizen delight).


Yes, but imagine if you owned a cab company in NYC or SF - how would you be feeling?


Or, you buy the Uber Driver pitch and buy a Uber approved 4 door car, and find you are working for less than minimum wage, or worse yet; you end up in Bankruptcy court?

I don't care that Uber is providing competition--I just don't like their business plan of hiring Drivers without any risk on their part. "Oh, but you're an Independent Contractor, but you need to drive a certain year four door car from our List.(A list that doesn't make a lot of sense? Oh yea, I have found selling a four door car on the secondary market is harder than selling a two door car; if your Uber job is not what they claimed?"

My point is I guess Uber provides great service for their customers, but I question the great experience for the "Independent Contractors"?

Sorry, but I saw Uber and had to chime in. Back on track--Twitter--that verification email I keep forgetting to fill out.


Interesting points - isn't that similar to how Twitter developers often feel? I often see comments here about "Don't build a business around Twitter" - seems actually somewhat similar.


Continuing the thought experiment: What about the case of AirBnB which definitely had some questionable behavior relative to Craigslist? Does that change the dynamic for you? AirBnB didn't degrade the experience of Craigslist, but it did leverage their network to build their own business. I hear this story all the time as one of the best examples of "growth hacking" (along with Dropbox).


> the Meerkat app will no longer be able to automatically push notifications that announce the live event to all of a Meerkater's Twitter followers.

Good. If "people I interact with twitter autospamming me every time they use an unrelated service" becomes A Thing then I might as well not use twitter.


But how is this different to that person firing off a Tweet?


There would be some content in the tweet, instead of an almost-assuredly dead link.


Anyone who builds applications built on Twitter's or Facebook's APIs deserves whatever treatment they get.

I've never seen companies that treats their API partners with as much disdain as those two companies.


"I get it that when you own the house, you own the rules," Rubin says. "You can say, I’m about to launch my own app, and I don’t want you to have the graph. But I think the two hours was a little aggressive and not working toward building a community." A little bit spot on. This is some kind of bullying.


"We are not naïve, we knew it was coming," Rubin tells Fast Company. "We thought that we would at least get a week notice—a fair game"

He knew it was coming, he knew it was wrong. Why would he need more warning than two hours? Why even a warning?


Nothing indicates that he knew it was "wrong", only that Twitter won't like it.


Among many problems: when applications exploit Twitter metadata to send unwanted unsolicited messages, that makes users far less willing to authenticate other apps to Twitter. I'm already paranoid about doing that (and Facebook authentication), because who knows how the app will choose to embarrass me?


Yet somehow I think promising not to spam again would not provide a path towards regaining access, post Periscope.


The bigger issue is that if there was no punishment, everyone else would now be trying to use the same tactics Meerkat did in order to spam people and obtain growth, which does result in a negative Twitter experience in the long run.

Startups will probably still try to do it anyways, though.


potato potato


You were also upvoted.


Really, at this point, I think you're certifiably retarded if you build a business even tangentially relying on Twitter or Facebook. This is exactly why people stopped doing it years ago. Why is this surprising to anyone?


They aren't locked in to Twitter, they've just used it to jump start their growth.

Can't argue it hasn't worked well for them with a lot of influential voice talking about the app.

Not to say it will succeed in the long run, but in this case they could have done a lot worse.


I wonder if they strategically waited to do this at SXSW, or if they just saw abuse reach a critical point.


Only if they strategically wanted to market Meerkat :) I had never heard of it before and doubt I would have if it wasn't for this debacle


I believe they did it _prior_ to SXSW. That's the type of event that could 10x awareness for Merkat and thus complaints when they cut it off. I have no inside knowledge but have heard it said by others they jumped on it due to SXSW.


I don't quite get why it matters if he got 2 hours or a week. Was he going to build a Twitter clone in that time?


It's a cool app, but there's been a huge degradation of my stream now that it's littered with |LIVE NOW| posts and out-of-context replies. Neither of which is fixed by this unfortunately.


As much as Meerkat's spam tactics downgrade my level of sympathy, I still think this is pretty awful behavior on Twitter's part. The list of people I follow is, essentially, personal information about me and them, contributed by me. I shouldn't be prevented from accessing my own information with the app or service of my choice; it shouldn't be part of a corporate fiefdom.

Still waiting for the first attempt to test the legal waters on scraping Twitter.


I don't think Twitter has anything against scraping lists of followers of public accounts -- after all, that's available to everyone. (One exception I think is geolocation data which can't be stored.) It's then using that information to spam Twitter itself that's the problem.

Facebook, OTOH, has pretty much turned off access to the friends graph for apps other than games.


In fairness, Twitter's API endpoints for accessing followers/following have fair rate limits. (75k follower Twitter IDs every 15 minutes)

https://dev.twitter.com/rest/reference/get/friends/ids

This is probably the only fair endpoint that Twitter has in terms of rate limiting. Getting that scale of information via conventional scraping would be infeasable.


No one likes spam, but invitation channels are certainly drying up. I sometimes wonder how fast Facebook would have grown if email spam filtering in 2004 was as aggressive as it is today, especially considering that its early growth was among .edu addresses.

(Remember that this was back before Gmail and often your email client did the bulk of spam detection, if not all .. though my memory is a little rusty)


Not that I care for what Meerkat was doing, but isn't this sort of a selling point of Fabric?


The original author included a misplaced modifier "only."

Despite their poor grammar it should have been included here.


So Now the Meerkat's awesome feature soon to get replicate by twitter itself?


I'm surprised they even got a notice


And it has nothing to do with Periscope.


If you are going to use Twitter to grow your business, at least be a little smarter about it. There are plenty of other ways to get yourself out there over Twitter without spamming users.


Well in all fairness it seems like they're growing their business like crazy but pre and post access.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: