The Increasingly Prevalent Auto-Follow Trick | Hacker News Reader

The Increasingly Prevalent Auto-Follow Trick

nickoneill.com

71 pointsbiznickman29 comments

Given two competing startups, one which starts users with a blank slate and one which a) gives them instant content when they sign up and b) manipulates them into continuing to use the application via using social proof and peer pressure, it is highly likely that only one of them will survive long enough to get geeks complaining about it. I know which one I'm betting on.

Consider it from this angle: you can undo this default for exactly the same amount of work you're proposing to subject everyone to: for each friend, manually select your desired state of followed or unfollowed. You're not going to actually do that, because it's a stupid amount of work and despite being motivated enough to write a blog post you're just not that into this app. To a close approximation, only freaks and geeks care enough about software to go through a multi-hour setting tweaking session prior to actually receiving value.

You can test this if you have an application you don't care about destroying user growth for: have them do lots of pointless work on the first user experience. See what it does to activation rates. A/B test "Do lots of gruntwork" versus "App does the pointless makework for you" if you want some statistical rigor.

P.S. If having one's monkey brain hacked on disconcerts you, stay far away from social software, because that is the whole game plan.

You seem to mix up value with spam.

I have a crazy idea. What if you were only notified by what you actually wanted to know about? Also, if I like your service/app I'm going to be insulted if you spam others with my actions without my explicit consent.

Good thing about crap like this is that people hopefully think twice before giving away their facebook/whatever account and forcing users to log in with facebook will be a thing in the past (keep telling yourself that it is because you want to make it easier for your users - you can not argue that you care about the user the slightest until you offer alternative ways to log in).

Is the concept of giving the user a choice that hard to grasp? Give the user a way to uncheck the "spam my friends" checkbox. Oh, I'm sorry, do your whole idea base itself around spamming users? Well, then, too bad. Please get off the internet.

You seem to be quite passionate about the ethical or aesthetically pleasing way to implementing following in a web application. Passion is cool, but I'd encourage you to think of what HN would be like if two ways to think about user following routinely resulted in "Please get off the Internet."

I do not necessarily disagree with your normative beliefs here. I'm just saying, as a positive statement of empirical reality, an application which is designed to work like the one you have described will perform worse than one which works like the emerging standard of autoinvites. Measurably worse. It's kind of like Girl Wearing A Headset photos: without saying they're either pleasing or tacky, one can make the observation that they've been tested to death hundreds of times and they generally increase conversion rates.

This kind of "positive statement of empirical reality" tends to (1) be true, and (2) make the world a worse place as everyone acts accordingly, until (3) it stops being true and (4) everything goes horribly wrong.

Banking, circa 2007: As a positive statement of empirical reality, selling unrealistic mortgages to unsuitable customers and then packaging up the default risk into CDSes and selling them on will perform better than lending responsibly. Banking, circa 2008: Oops.

Any pyramid scheme, in its earliest stages: As a positive statement of empirical reality, signing up will perform better than ignoring it. The same pyramid scheme, after a couple of iterations have saturated the market of available fools: Oops.

The Dutch tulip market in 1735: As a positive statement of empirical reality, selling everything you've got to buy tulips as the market climbs inexorably higher will perform better than holding on to old-fashioned goods like gold and houses and food. The Dutch tulip market in 1740: Oops.

The personal computer market, circa 2000: As a positive statement of empirical reality, what people want to buy is the cheapest possible computers (or, for a lucrative subset of the market, the cheapest possible computers with impressive-looking specs) running Microsoft Windows. Selling cheap generic boxes will perform better than trying to make something of really high quality, and anything other than Windows is a loser's game. The personal computer market, in 2011: Apple is bigger than Microsoft, many makers of generic beige boxes are frantically trying to get out of the business, and it's a serious question whether "personal computers" in general will get their market almost completely eaten by tablets (meaning, for the present, mostly iPads) and smartphones.

Spamming users' social networks is like selling crappy mortgages, tulips, or mediocre cookie-cutter computers. It may, for the time being, be an easier way to make money than the obvious alternatives, but every time someone does it the world becomes a slightly worse place, and sooner or later it's likely that people will notice and stop playing along.

Of course you're welcome to point out the "positive statement of empirical reality" that spamming works. You're probably right. And you're under no obligation to acknowledge the ethical issues that those of us with tenderer consciences may have. But when someone responds to that with "please get off the internet", I think they've got a point, just as they would have if you'd suggested that startups should lie to their customers, or sell addictive drugs to children, or dump poisonous byproducts into the local water supply, in order to increase their profits. It might "perform better". It might happen not to be illegal. But that still doesn't mean that a reasonable person should be recommending it.

For me personally, I don't mind a lot of autoconfig and getting signed up for stuff by default. It demonstrates the value of a new service in a tangible way, If I don't like it, I can un-sign-up, and if that doesn't work I can filter spam.

What I do mind is services that automatically communicate in any way whatsoever with other people on my behalf. I don't mind having a social graph, but I want to be able to place and maintain graph cuts. I think as experiences like Take This Lollipop become more popular, more people are going to think like me.

given two competing anythings the one that plays most aggressively wins in the short run (see ww2) but these tactics usually turn out to be short lived crutches. the path of user choice and openness (which starts with emptiness) is not going to win an a/b test but in the long run is likely to win the war. and its going to force you to come up with something genuinely compelling.

At the end of the day a web app can do one of three things. It can delight it's users, it can exist and nobody cares or it can annoy it's users.

It's helpful for companies to think before they act. Is this action going to delight, annoy or have no impact? The companies that choose to delight their users will win over time. Delighted users tell their friends, which is the biggest form of social proof. Too many companies choose perceived short-term gains at the cost of annoying their users. That's not a long-term winning strategy.

This comment appears to be orthogonal to the discussion, which is rooted in a comment that says aggressive auto-follow (EAFP) is likely to outperform asking permission (LBYL).

Facebook is not an LBYL company, nor is Google, and Facebook and Google are crushing the social space.

What may be the case is that normal people are in fact delighted by the EAFP approach, and that the opinions of nerds are irrelevant.

I feel like the startup community has skipped the conversation about using facebook or twitter for login and gone straight to this. There are a lot of apps now that aren't particularly social outside of interactions within the app[0] and require me to use facebook or twitter to log in. I don't want either of those to be my identity on other sites.

[0] And those interactions don't depend on an outside social graph at all. I'm thinking of Anyasq in particular here.

Using Facebook/Twitter/Any-Other-Existing-Login is a good way to speed up the experience for new users, however, this should then create a new profile and always give the users a choice on how they login/maintain their profile with you.

I fully understand providing the option. "Create an account, or just log in with Facebook" is good for the user. My objection is when the "create an account" option isn't there and I can only use the site by logging in through a third party.

Anyasq is an example of this; you can only log in with facebook or twitter.

I'm not really convinced its in startups' best interests, either.

Being so limited, that is. Letting users get started without first creating yet another account is certainly awesome, and getting rid of that friction is undoubtedly a huge component of the success of sites like Stack Overflow.

But going with only one or two of the social behemoths seems unwise. It means you've decided to hitch your applecart to a business that, by virtue of being in the social space, could very easily become a competitor. Or decide to bless one of your competitors in some way that disadvantages you. Now, that's one thing if it's another startup with which you're forming a strategic partnership. But when it's one of the Facebooks or Twitters of Googles of the world, you're dreaming if you think you aren't doing business at the King's pleasure.

And with so many OpenID providers out there, it's not like it really takes any effort to give yourself some flexibility. So it's hard to fathom why any thoughtful, ambitious entrepreneur wouldn't want to do that.

I really wonder if this falls under the category of "things hacker dislike/complain about but regular people don't care/mind"

I think Pinterest does this too. I'd guess most users think its a byproduct of Facebook connect and don't care. If I had a mass-market/consumer site, I'd consider doing it.

Heck, I'm a hacker and even I like the feature. Why else would I even connect my Facebook account in the first place, if not to use my established social circle there? You can join without Twitter or Facebook.

If it's to automatically follow people in your Facebook (or Twitter) account when you link your account - and, big and - if they are already using the app, then I have nothing against it. I'm personally sick of having to do the grunt work of mirroring manually and micromanaging my social networks for each new app. In these apps, if I find later on athat a few connections don't publish stuff I like, I unfollow them.

Giving you the option to allow the app to do the grunt work is fine.

But just doing it, without asking, (and without an option to turn it off?) is just slimey and stupid.

Compare this to Batch. Batch uses Facebook exclusively and doesn't communicate that people are "following" you when they sign up (or to use Facebook terms, they don't indicate that people have become your friend either, they simply say they've joined Batch). They just assume that you have a relationship because you authenticate with a platform that already says you do. Very straight forward.

This is very similar to Oink's behavior, but far far less presumptuous in the way it communicates. Personally, I was thankful that Batch saved me the work of curating my network. But if you use Oink's language/behavior you have an app that says I followed someone when all I did was sign up? They almost immediately lose my trust. Not a good way to start the relationship with me.

This happened when I joined Quora recently. I honestly felt embarrassed that I had "followed" 100+ at once, especially since I haven't been in touch with many of those people recently.

We bounced around on this for a long time with eggtweeter and came up with this compromise:

When a user creates an account (via Twitter), the KyMaLabs account is automatically followed. If they choose to unfollow after that, we have nothing to say about it. Other than that, we're not manipulating their follow lists in any particular way because we felt it would be either undesirable or slimy.

We decided to do an auto-follow for a few reasons:

a) It connects us with our userbase so if we Tweet information or updates, they'll see them.

b) It helps us quickly gauge new user signup at a glance. Sure we do this on the back-side with some reporting tools, but its nice for us to see our follow count go up as well.

c) We're providing a free and valuable service (while in openbeta) without ads or other junk, we think of this as the "cost" of the service.

d) We don't ask for unnecessary information from the user, we get their Twitter id and we ask them for an email address on signup and that's pretty much it. (Once we move to some paid services it'll have to change somewhat, but we chose to keep it simple during the beta period).

And that's it. So far most people don't unfollow (our metric for "didn't like it"), and the ones that do haven't complained directly yet (our metric for "hated it") so we think we've kept successfully on the side of the line of ethical use of an auto-follow.

That's not what this is about.

His beef is:

1. Signup to a new service using facebook connect

2. One of your facebook friends also joins service

3. The service automatically adds them as your friend/follower/highlighteduser/whatever inside this new service without your intervention

It's downloaded your existing social network from another application to create its internal one even though the use cases might be wildly different.

Wait, when someone signs up, you silently have their Twitter account follow your Twitter account? That's pretty egregious in my book and would merit an instant account deletion.

By logging in with or connecting my Twitter account to any service, I want some of the value I've put into curating my Twitter account to carry over to my new account on your website/app/whatever (as Oink appears to be doing here).

I NEVER want you to go back and mess with my Twitter experience. It's a violation of the implied trust I'm giving you by creating an account in the first place.

What's wrong with giving people an option to follow you on Twitter? A simple checkbox when registering would suffice. Sure, your follower count won't be as big, but it's also the only ethical option of the two, in my book.

PS. I went to your website and clicked on "What is eggtweeter?" on the homepage. The resulting page left me utterly confused about what you do. It seems like a list of what the site can do, but not what it actually does. Then I went to your FAQ and the first question immediately and succinctly explained it. You should think about moving that description to the about page, or even better yet, the home page (right now, neither page has the words "scheduled tweets" anywhere).

I thought I saw on Oink's signup a toggle that let you opt-in to auto-following other users. I was surprised to see that as usually it's opt-out (if at all). What this seems like to me is a glitch on Oink's end - not following their own preference options they expose to their users.

edit: I definitely didn't mark down to auto follow my friends. I just got the pig sound notification and it scared me half to death. Oooh boy my heart is racing.

There's an option to disable auto-syncing as they call it, and it's on by default. Interestingly enough, when I went to check out the settings area just now, I saw several Twitter friends who I wasn't friends with yet, despite having auto-syncing turned on, so this looks like a bug on their end. I got an error when I tried to check my Facebook friends' status as well.

There's also an option to toggle being notified when a friend joins (presumably this refers to auto-syncing), which is on by default too.

I get how some people might feel like it's annoying to auto follow, but the reality is that most people don't care / notice that they're now following all of their friends. And at the end of the day, having friends using the same app as you does provide a better user experience which is why this trend has emerged.

You have a valid point about having friends in the app. However I'd imagine that there's a more elegant way of accomplishing this. I realize I'm a rare use case but I have thousands of Facebook friends many of which I don't want to know everything they're up to. I'd imagine many other people don't want to know everything all of their Facebook friends are doing. Perhaps someone should complete a more detailed study on this :)

Unfortunately rude and aggressive "social" apps build the largest audiences by those same tactics. Remember how MySpace took off: the sign up process included "Enter your Gmail/Yahoo/Hotmail/Outlook login info here" which it then used to spam all of your email contacts, a blatant privacy violation which millions of users gladly participated in.

I've noticed this new iPhone app "The Eatery" doing this. It's fairly annoying, but as long as I keep my facebook friends list trimmed to "actual friends" then I don't mind so much.

Frankly, it seems "dumb". Foursquare suggests friends to me that it thinks it has connected via Facebook and Twitter (which I opted into it doing). The suggestions are easy to ignore and dismiss. It seems obvious to me. I guess this just feels like a cheap slimy way to be connections in the social graph, but I think in the process of doing so, they devalue the value of connections that people put effort into establishing.

If I were following all of my Facebook friends on Twitter, then it would just be Facebook. Do you really want to rush to build the next generic, mass-appeal social network where the content is low quality drivel?

Couldn't have said it better.

Reply on news.ycombinator.com