Hacker News Reader

Comment by echelon | Hacker News Reader

I wonder what the real reason for account deletions is.

Could it be possible they have overflowed their primary / foreign key allocation necessary to support further user growth? Depending on their internal architecture, a migration to a wider key might not be feasible within the time horizon they have. Maybe it's easier to delete accounts?

If they have weird sharding and a lot of microservices, this might be their only way out without shutting off registrations - and they'd never do that.

Or maybe they have metrics that show new users skip signing up when their desired name is unavailable.

Can a Twitter engineer speak more to this?

View on HN

It could be a variety of reasons.

If you forced me to guess though, I think the most straightforward, simplest one is that threatening account deletion is a good way to force older users to become active again, which will increase engagement on their platform and allow them to deliver more ads. In other words, the same reason that Facebook will start spamming your email if you stop logging in.

I can't imagine any manager at Twitter would be upset about the idea that a substantial portion of their dormant userbase would be effectively "forced" to log in every 4-5 months or so.

Wow, I'm usually pretty cynical and even I didn't think of that one. Makes sense when you put it like that though!

Also as someone said below, it forces users to accept any new changes to the ToS.

>>In other words, the same reason that Facebook will start spamming your email if you stop logging in.

Are you sure that's what is actually going on? patio11 calls it lifecycle email marketing, and Nir Eyal thinks it is just a harmless way to keep your users "hooked". Why would you even say such things about such a brilliant technique which provides such great joy to the "growth hackers"?

Oh, I just got a "notification" from Facebook that someone I care about said something I probably don't care about, but there isn't any way to know for sure so I have to log back in to Facebook again. I will be back in 3 hours.

For what it’s worth, I haven’t heard Nir Eyal call it “harmless”.

It is to cull high-volume bot programs, which are managed through the API. Note that every account will require a log-in to count as active, not just be posting.

If a single bot account was created by a person for fun, they can easily log into it once to prove activity.

But if a person (or, say, a national intelligence service) has created 20,000 accounts programmatically, they’re going to have a much harder time manually logging into every one of them to preserve them.

I doubt it’s something that interesting. I’ve read it’s mostly a legal thing regarding the original terms and conditions accepted by the user.

What's with the rash of accounts with hex/base64 names today? They've been popping up all over.

Im not sure what the other anonymous cowards are doing but I’m just not very creative and did:

    head /dev/urandom | shasum

When creating a throwaway.

Having an account name like that just makes me completely skip any opinion a person has on social media.It's probably just some convoluted logic for a throwaway shitposting account

I would be pleasantly surprised if one of the big tech companies was deciding policy based on actual technical problems rather than politics/'business' reasons/whatever junk science is trending today. That would make a nice change.

No I highly doubt it. This is most likely as they say a GDPR issue. Per the GDPR they'll have to keep track of which users have agreed to which version of the ToS and Privacy Policy and make sure that they only process their data in accordance. This is a PITA, much easier to deem accounts inactive and delete their data. They've probably been contacted by the data commissioner of some EU country and asked nicely to comply with the GDPR.

A second, somewhat related issue, is that these inactive accounts might also be used by all the russian state trolls. It's much harder to detect when old accounts start doing suspicious things than brand new ones, especially if you don't have a robust process in place for flagging inactive users.

We're talking about users that last logged in more than 6 month ago here. It has obviously nothing to do with government sponsored shills or the GDPR. It's just that Twitter wants to eliminate users that are not "monetizable", they are also looking into users I quote "log in but don't do anything" on Twitter.

Actually, under the GDPR, as a general rule, you can't just keep personal data on your servers forever so GDPR might be involved but for personal data cleanup reasons/requirements.

> Or maybe they have metrics that show new users skip signing up when their desired name is unavailable.

Well I know I did.

I am not a Twitter engineer, but I guess that this is nothing else but a great marketing campaign - suddenly many people remember all their accounts and login and that will help to produce very nice reports about the "active user base".

But why is this needed?

My guess: many people are sick of Twitter and since one president of one nation made it his personal propaganda vehicle it is becoming a mental wasteland - new ecosystems are growing and people inside these new communities are quite happy that all the zombies are contained in that trash dump, like facebook.

Actually we should all be happy that these first big "social" networks decontaminated the internet for us - these are now the Tschernobyls of social media where all the toxic waste is collected, hopefully for many years.

Interesting things now can happen somewhere else.

Reply on news.ycombinator.com