Thursday
Nov132008
Plenty of Fish Says Scaling for Free Doesn't Pay

Plenty of Fish CEO Markus Frind, famous nerd hero for making over $10 million a year from Google ads on a free dating site he made and ran all by himself, now sees a problem with the free model:
The problem with free is that every time you double the size of your database the cost of maintaining the site grows 6 fold. I really underestimated how much resources it would take, I have one database table now that exceeds 3 billion records. The bigger you get as a free site the less money you make per visit and the more it costs to service a visit...There is really no money in being free and we have to start experimenting with other models now or we won’t be able to compete in 3 or 4 years.
As one commenter succinctly put it: the “golden time” of AdSense is over. Time to look at costs. The POF architecture is to run scarily huge tables on single machines. They also buy and maintain their own SAN. So it seems scaling up is what is increasing costs and decreasing profits. I wonder if the economics of cloud storage and cloud architectures might have a more linear cost curve?
Reader Comments (22)
He should shard his database, by user. We run a site with 5 million members, on PostgreSQL, and just did that, which allows us to scale out horizontally pretty easily as we need. Each user's data (profile, friends lists, etc etc) is kept together on the same shard. When the existing shards reach capacity, we just add a new one.
Hang on... that sounds like broken logic.
Sure, if you've managed to go down a scaling dead-end you might find it expensive to scale further, but doesn't translate to your revenue model.
If you get a static revenue per user from a free model, you're still only going to get a static revenue per user on a paid model... unless he's suggesting a system whereby membership gets more expensive as the number of users grows.
Hello,
Other commenters point out the flaw in the argument: whatever the revenue strategy what does not scale is reliance on a single gigantic database. This is Scalability 101.
My concern here is that you wrapped your excerpt in a PRE tag instead of BLOCKQUOTE, making it hell to try and read on a small screen. Please fix your markup, unless you would prefer not to be read by folks using mobile devices. Thanks!
Sincerely,
-daniel
Since the key use case for a dating site is cross user queries, trivial sharding by user will not give linear scalability. If it's permissible to limit searches to a geographic area, sharding by zip code prefix could work acceptably. Otherwise some form of distributed index server will be needed.
Why dont you just cut down some of the countries.. like that ones with alot of scammers like russia etc.... then you will have less fakes taken up alot of room.. I dont know .. lol just a suggestion.
Just prune your database.
Delete all the users that didn't login in the past 12 months.
That would cut million users that are not interested anymore.
I don't know what to beleive when I hear this guy talk about POF. This particular comment doesn't make any sense to me. I mean he's running this thing on a handful of boxes without any staff isn't he? How expensive can maintenance be getting compared to those huge AdSense revenues (even if he's had to buy very expensive boxes and SANs)?
Unless he's not really bringing in $10M of course....
I am a co-founder of a free dating site (LiveDateLove.com) but since it's much smaller than PoF I don't experience these problems yet. I did read that Marcus spends 10K/month on the server but compared to his income it's nothing. So I don't know what he means by the increasing cost... Did he provide any figures? Perhaps this article is just another way of him advertising his business. I noticed how much he likes to brag either about his income or his server or how genius he is. And by the way, he DOES NOT run this site by himself, he gets a lot of help from the web hosting company. They're the ones looking after his servers and his site. And his girlfriend helps him out with customer support.
Hey Markus, just a thought... If your volunteer mods weren't such dicks, you wouldn't have so many people making multiple profiles. The profiles don't get deleted - you just can't post with them (great for your numbers BTW). As it is, I've currently got 8 profiles on your database (Thanks Ticketoride!)... gonna make a few more soon for back-ups. Sure, you can boast you have lots of members, but really, you don't. Most are fake profiles to get around the draconian tactics of your forum moderators. I don't know ANYONE on POF who doesn't have at least two profiles... so THAT might be your problem. You saved money by getting the village idiots to moderate your site, and instead you created a database of fake profiles and angry users. Hey, don't take my word for it. Read ANY of the reviews of POF.
(Pssst... how can somebody be the CEO of a one-man operation?)
Hello! How much users do you store on one PostgreSQL server?
I was thinking the same; heck how much increase in operation are we talking about here; give me a percentage of total revenue. Considering the single guy business model; even if he spends half of his revenue on maintaining the site, he is still pulling in few mils.
His goal isn't to run a great problem free site though, it's the beater of cars for an online website. It works for him so why would he care. As for his costs, maybe he's just annoyed that instead of $10K a month he'll have to spend 60K, you're thinking well who cares if he makes seven (or eight) figures, but going by that logic wealthy people who all be generous tippers (ha!).
This is a quote from his most recent article in Inc., I don't use POF, never have, but I love this line. "I don't listen to the users," he says. "The people who suggest things are the vocal minority who have stupid ideas that only apply to their little niches."
Here's the full article
http://www.inc.com/magazine/20090101/and-the-money-comes-rolling-in.html
In this article he is still 100% pro free site, so the OP contradicts this article.
It's definately a broken logic. Maintainance costs really increase with the number of visitors, but they increase slower - it's less then linear dependance. In the same time revenue increases as linear or higher.
Sure, keeping 5 bln records on a single machine is not the right approach. Think about distributed systems like Google does - I'e been told that they keep all data in RAM and they have tens of thousands of machines for that purpose. If any machine does not response, they just take it off and replace with the new one.
I have tried to implement the same memory-based model on http://www.livedatesearch.com/ engine but then I came to conclusion that for middle-scaled projects it's not the best choise as well. Now I vote for distributed php/mysql (InnoDB) platform based on SOAP. This what really works the best from my opinion and is scalable also.
Anatoliy
Owner and creator of http://www.livedatesearch.com/
Markus is also the only one (?) that seems to have an architecture based on nothing cached. When I read that here, I thought that was odd.
The point I'm getting at is maybe he's a master of misinformation?
The guy's brilliant. He deserves a lot of credit. You gotta admit there's either a sense of jealousy or irony in his ability to pull off that site with such little hardware (CDNs aside). In the back of our minds, even the most skeptical among us is probably wondering..."what if Markus really is running that thing on a single web server?"
Dude, you website is so slow!! I don't think the technology you are using his helping? Maybe you should consult with Marcus for some tips!
The FREE has nothing to do with the increased cost of expansion. Charging money will only partially offset the cost of scaling and delay the onset of the inevitable. The exponential increase in disk space and processor usage is due to the number of interconnections between the number of members. When there are just a few members then adding a new member will add relatively few connections, when there are millions of members just adding one new member can require millions of new connections. The thing to do here is to streamline the code, or a suggestion: get members to contribute power and space thru a grid computing method such as boinc.
Sounds to me like they are just making an excuse to start charging for something now that they have lots of members. Yahoo did the same thing with their personals. It was free for a long time then suddenly they started to charge. That's a bad, corrupt way to do business.
Hate to say this, but you're wrong re: the two profile issue. Are there exceptions? Of course! I know of one overly agressive man on pof who was obsessed with sex, didn't want to date, period....His first id disappeared, and he turned up under a new id, still clueless and obnoxious. But lots of people just have one profile/id on POF...Guess some of us are just too busy or whatever other reason to post two profiles/ id's....Just surviving in this economy is a job in itself...Yes, really....
If it wasn't paying off for him, he wouldn't still be doing it. It would take an hour maybe to make it into a pay site like adult friend finder. He's putting this info out to cut down on the amount of potential competitors he might have that are eyeing his income now and thinking about copying his business model. If you want to make money like plentyoffish, you are going to watch what they do and do as they do. When the CEO comes out and says that its not profitable to do it that way, you may change course, and that means less competition for them.
Could scaling horizontally and employing a strong distributed caching layer work to lower costs and increase efficiency?
I have to agree with the comment that Marcus is probably trying to scare off competition with this "insight". If he's making 10 million a year, so what if his meager (by comparisson) costs double or quadruple? Hell, if the poor guy only ends up clearing a million a year it might still be worth his while... It is ironic that the man who waved the flag about "free is the way to go" is now trying to backpeddle by saying too many users will take away that profitability. As others have pointed out here, there are a lot of efficient ways to cut down on the number of users.
Throwing out just enough information to the public is a tactic from a smart business man. This allows all of YOU to give him all or just enough assistance to grow an even more successful business without paying out for that help. Now he knows who to talk to and what programs/systems to use. I want to thank you all as well. Communication is such a powerful gift that most if not all of us have. To bad it goes to waste by many. DennisWageman.com
CLICKHERE