Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Show HN: Verified Twitter Followers (fruji.com)
62 points by mittermayr on Oct 27, 2014 | hide | past | favorite | 46 comments


I run fruji.com, which has become somewhat of an unexpected monster and received more interest than I ever thought it would. Too much actually, so that most of the bottle necks with regards to communicating through Twitter's API have become massive issues. I've had lots of celebrity and super large brands/companies sign up with onboarding periods of weeks (imagine, you sign up, your first report shows up 3 weeks later, uggh). So, after some customer feedback rounds, I've realized that most of the big accounts care about their verified Twitter followers a lot (they are often big relays/multipliers). That data is much quicker to produce, so I've created a MVP that does just that. It seems to work fine after a few first tests with around 50 users, but I haven't announced it yet to my existing users as well as the general public broadly. Would love to get some technical feedback first from you guys (you've guided me so well through creating so many other things in the first place and avoiding mistakes).

EDIT: No need to sign up, this is basically what you'll get after signing up (this is a shared page, if you look at it from your own account, you're asked to upgrade through tweeting or paying $2 if you have more than 5 verified followers): http://verified.fruji.com/results/3f4cdab2fa81ac30fd57ef5abd...

EDIT 2: The tech behind: It's a Sinatra/Puma setup, jobs are being queued up through Sidekiq, data is being stored in Redis (just lots of ids and id lists), payments are being run through Stripe and front-end is basic JS with super simple custom HTML templating. It runs pretty slim.

EDIT 3: In case you do sign up, please use 'freefruji' to get the full report for free. No reason to pay now.


Last time I ate at Taco Bell was over six years ago, and yet they are following me!

How do they know?


You should check your Twitter history, maybe there was late night drunken "I need to talk to Taco Bell right now" involved :)


Taco Bell's Twitter account is only following 14 people. Are you Andrew W.K.?

https://twitter.com/TacoBell/following


Taco Bell Canada: https://mobile.twitter.com/TacoBellCanada

I'm almost vegan now, so I'm not their target market.


Just a heads up - you put anything into the promo code box on the upgrade back and press verify and it'll return a white page (I assume error!). :-)


that's strange, I see that my api responds with nonsense, but it shouldn't redirect you. hmm. I'll check it out! thanks!


No worries. Everything else looks great :-)


It would be nice if you could also get this report for an account that you don't own yourself.


it's doable, but the twitter limits are much stricter. I've done it on request for certain clients, but it would probably be great to implement as a self-service feature.


This is interesting, but the algorithm needs to be tweaked. It's cool that @barakobama follows me, but he also follows 200k other people, so it's not like I'm a special snowflake. I have other followers that are, imo, much more impressive. Some followers have 100k+ followers and only follow a few hundred accounts. Those are the people that are more exciting to find out about.

edit: Well, it's not so much that the algorithm needs to be tweaked, it's advertised as showing VIT followers and that's what it shows. Maybe sort your VIT followers by the ratio of followers/followees?


great, great response! my original service (fruji.com) did much deeper analysis (figuring out when someone is important, vs. just running a large marketing account, likeliness of large account holders actually reading your tweets based on the noise on their timeline, stuff like that) - but this took so much data-fetching from Twitter that I ended up with a 3-week onboarding process for the large users.

but what would totally make sense, based on your thinking here, is to apply some of the stats/filtering (a few buttons or sliders maybe) to help sorting those selected few verified accounts based on a few heuristics.

I'll keep thinking about that and see what I can come up with.


This is like the opposite of what "startups" do. They put out an MVP and scale from there. but in fact you found that Fruji did too much and scaled it back and came out with an Even More Minimally Viable Product. so cool. EMMVP!

I wonder what else you feel you could do singularly that would help twitterers. Or at least like to see in such a plain awesome usable way!


thanks for the kind words! yeah, well, you wouldn't believe how many startups go from a perfectly fine MVP to scaling like crazy in one direction thinking it's all good from here onwards only to realize the MVP was adressing a very different aspect/nuance of the market than the built-out product does. I think klout started with the right MVP, but ended up being a terrible built-out product in the end.


In case anyone is interested in the architecture behind this, job-scheduling, etc., I'd happily write it up in a blog post or something. It took a few releases/products to put together a nice, reliable and slim model that withstands HN mayhem easily (I run this on a $15-20/month server without that machine even showing a hint of sweat). I use this model/architecture for all prototypes I build and it allows to release/modify/adapt quickly once it goes live.

EDIT: Also, Stripe is just awesome. It just works, so, well. I even made $20 today, so cool, I'll make sure to re-invest this in one of the next Show HNs! Thanks guys!


I'd definitely be interested in reading a blog post on that. Good work with the release!


thanks for the kind words, Scott – i'll see what I can get done in the next few days to put together a write-up maybe.


So... what's the value in knowing this? For a lot people, most if not all of their "verified twitter followers" also follow 10s of 1000s of others. I am not special.

What's the value of being followed by a "verified" user anyway? I know I'm pretty naive and uneducated about the whole "social networking" marketing/advertising industry.

P.S. On this page, the CSS attribute "user-select: none;" is really annoying.


good questions. I guess a lot of it depends on the goals a person might have (get into certain industries, build a reputation by providing specific accounts with helpful responses, etc.) -- and to a great extend (for most of us) it's just a nice to have bit of information that is not directly actionable. this is also why I tried to implement a free 'see up to 5 verified followers' type of thing, because that's where most people (me at least) typically are. anything higher than that could mean you might already have an idea of what that's worth to you and how you can act upon that information.

the user-select is something I put in last-minute because there was something sort of annoying for some reason when I tried it on a phone, but couldn't figure it out, so I went noob. sorry about that, will get fixed.


It's not perfect, but looking at only 'verified' users cuts through about 95% of the chaff (although you do lose many excellent non-verified people, naturally).

But if you discover you're followed by a handful of your favorite famous programmers, maybe a VC or two, things like that.. it could be useful to know you might be able to take advantage (in a nice way) of those connections some day.


and, what peter said, exactly. and especially keep in mind: I didn't mean to push this out to you guys as users necessarly, initially was thinking of getting some tech/implementation feedback rather than market testing this. the incentive for creating this little product was mostly kindled by many larger accounts I host at fruji.com who asked me to get them the data faster and that's all they'd need for now, just a list of their verified accounts (as a means of pre-selecting).

i personally also fully agree that there are many, many folks out there without a verified account who provide tremendous value. this really is just one niche approach. one page in a large report basically.


Seems to work! In fact, it showed me 6 people, but I didn't sign up fully.

Sadly, my verified followers mostly are made up of companies I've worked for or companies I've complained about/used Twitter as customer service with.


same here. it's usually (sadly) also often companies trying to build their initial user base.


I found something about this very interesting.. that is, Stripe + low dollar amount + something potentially useful == me copying and pasting in my credit card details from 1Password very quickly.

I've not come across anyone else doing anything like this, yet there are probably thousands of such small services I'd be happy to pay a few bucks for here and there, especially if it's through a payment system I trust like Stripe or PayPal.

That aside, I wish it'd let me export the list of verified followers in a nice way after I've paid my $2 :-)


yes, absolutely agree. this is the power of Stripe combined with (finally!) super affordable v-host providers I was partly adressing earlier in a comment. my cost is actually next to nil. They make it possible for me to just push this out with minimal implementation work (this took about 3-4 days to write) and really sane fees all around. It's so simple, the refunding is so straight forward, it actually makes sense to keep costs low, simply because I can, at this point.

and of course, I could slap a $20 price on that and see if it still validates. and the implication of a higher price > higher quality etc. would possibly play into that as well.

but I am a very reluctant 'paying user' myself. it takes ages for me to convince me until I finally pull that credit card out. I feel many people just charge what everybody else charges, or try to get the most out of it (which all makes sense). but this is just a small tool, I want to stick to coffee-pricing if at all possible.

and you guys paid my server bills for the next 3 months now, so this is nice enough for me today.


I wouldn't have paid $20 for it as it is, BTW.

I spend hundreds of dollars a month on data streams related to my business, but this falls into the personal expense bracket. $5 would be my "just go for it" limit on this, but I'd definitely pay more if the offer was stronger (for example: $20 to do 10 accounts, have a CSV export, get a report emailed to me once per month for a year - value adds like that).


check out fruji.com then ;) but let me know once you've done (help@fruji.com) and I'll get you that pro account for free.


Haha, I had no idea you had a main product at the main domain - I'll check it out later :)


Interesting. I paid the $2, or at least tried to. The Stripe form processed and gave a "paid" checkmark, but then redirected me to a plain HTML page that simply said "Oh, too bad." When I reloaded it prompted me to pay again, although I believe it went through.

So, interesting app but needs some payment debugging.


oh wow, that's not good. I've fixed it now, you should be good to go (account is upgraded). sorry about the troubles there!


Works now, thanks!


What's the difference between you guys and SocialRank.com/Verifiedfollowers.co?


I don't try to be different, I don't even try to compare to others at this point. When I made fruji.com, everybody said: Dude, there's Klout, you are nuts, it doesn't make sense. People often forget how big the pond actually is, there's so many fish in the sea here, there's enough room for all kinds of niche ideas. A lot of my paying users on fruji.com are also paying users for thousand-dollar enterprise social monitoring tools. I usually only get curious about competitors once people start asking for certain features that keep them from sticking with my product and forces them to go back all the time.

EDIT: Just one thing that differentiates this little thing from the ones you mentioned: I don't require your e-mail address to proceed. Couldn't test the sites you mentioned for that reason, wouldn't let me go past that.


a small differentiator to people who are serious about paying, imo.


definitely. but most people find out they're willing to pay after trying out a bunch of services. and it's always tough to unlist yourself from all the e-mail spam that follows once you've found the service you want to stay with. but I get the argumentation, requiring e-mail on fruji.com for the same reasons.


Nicely presented and implemented...although the "upgrade" link on the box that appears when you mouseover the icon to the right side of the first 5 results does not seem to work. When I click it nothing happens.

Tested in IE11 and FireFox.


fixed (I think). thanks man!


I actually built something similar at a hackday a few weekends ago specifically for startup people (i.e it showed you investors, journalists, etc. who followed you):

http://startupfollowers.com

(although far more crude/hacked together than fruji)


i like it!


Boy, I sure will "be proud" that the Seamless and Roku corporate accounts follow me. I'm sure that's because of my high quality content and not a craven attempt at brand/customer building ;)


Thanks for that tool! Really great way to find the "hidden" gems among the followers and keep a closer eye on engaging with them.

/edited my spelling. Thanks tgcordell!


thanks for taking the time to try this out, really appreciate the kind words.


germs? gems?


Cool, now I get to decide if @gocomics following me actually means anything!

(I'm a cartoonist, and have a friend who is syndicated there.)


I just checked them out, they at least have a lot more followers than they follow, so this usually is a good sign. although, unfortunately, they have tweets from about 8k people on their timeline, so it's gonna be tricky to get their attention in all that noise.


Yeah, if I actually want their attention it is probably better to go schmooze them in person at a comic con or something!




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

Search: