Tech Mobsters Manipulate Dating Sites To Harvest Voter Data And Spy On Users
Why You Should Never Pay For Online Dating - Founder of OK Cupid Reveals OKC and March.com Are Just Scams
- IAC's Dating sites exists to harvest voters for election manipulation and spy on users per www.dates111.com
- The Clinton's have a business interest in Match.com, OK Cupid and most big dating sites.
I'd like to show why the practice of paying for dates on sites like Match.com and eHarmony (and OK CUPID and Bumble And Tinder, etc.) is fundamentally broken, and broken in ways that most people don't realize.
For one thing, their business model exacerbates a problem found on every dating site:
For another thing, as I'll explain, pay sites have a unique incentive to profit from their customers' disappointment.
As a founder of OkCupid I'm of course motivated to point out our competitors' flaws. So take what I have to say today with a grain of salt. But I intend to show, just by doing some simple calculations, that pay dating is a bad idea; actually, I won't be showing this so much as the pay sites themselves, because most of the data I'll use is from Match and eHarmony's own public statements. I'll list my sources at the bottom of the post, in case you want to check.
eHarmony claims over 20 million members on their homepage, and their CEO, Greg Waldorf, reiterates that number regularly in interviews1. If your goal is to find someone special, 20 million people is a lot of options—roughly a quarter of all singles in the U.S. This sounds awesome until you realize that most of these people can’t reply, because only paying customers are allowed to message.
So let's now ask the real question: of these 20 million people eHarmony claims you can flirt with, how many are actually able to flirt back? They closely guard their number of paid subscribers, with good reason. Nonetheless, we are able to deduce their base from known information. We'll give eHarmony the highest subscribership possible.
- We'll start with their yearly revenue: $250M in 2009 as reported by the industry analysts at Piper Jaffray and CNBC2.
- Since eHarmony charges users by the month, we'll divide that big number by 12 and, rounding up, get $21M.
- Now all we need to know is how much the average user pays per month. If we divide that into the $21M they make, we know how many subscribers they have. Their rates run this gamut:
$19.95 per month, for a 12-month subscriptionFrom those numbers, we can see that they have somewhere between about 350,000 and 1,050,000 subscribers (the lower number supposes everyone is month-to-month, the higher supposes everyone is yearly).
$29.95 per month, for a 6-month subscription
$59.95 per month, for 1 month at a time
- What's the exact number? Well, I found this helpful nugget in eHarmony's advertising materials3: The most charitable way to interpret this last sentence is to assume their average account life is 6.5 months.
- We're almost there. To get eHarmony’s total subscribers, we divide their $21 million in revenue by the average subscription price. Therefore maximizing total subscribers is just a question of minimizing the average monthly fee. First off, let's do them the favor of assuming no one pays month-to-month.
- Our remaining dilemma can be expressed mathematically like this:
- After some dickery with a legal pad we discover, in the best case for eHarmony, 1/13 of their users are on the yearly plan, and the rest subscribe 6 months at a time. Thus the minimum average monthly fee is $29.18. They have at most 719,652 subscribers.
- For the sake of argument, let's round that up to an even 750,000.
So, having given eHarmony the benefit of the doubt at every turn, let's look at where that leaves their site:
Yes, only 1/30th of the "20 million users" they advertise is someone you can actually talk to. That's the paradox: the more they pump up their membership totals to convince you to sign up, the worse they look.
And the ironic thing is that although they basically admit their sites are filled with chaff, pay sites have little interest in telling you who's paying and who isn't. In fact, it's better for them to show you people who haven't paid, even if it means they're wasting your time. We'll show that in the next section.
First I want to show you what 29 to 1, advertised people to real, feels like. Here are some single, attractive OkCupid users.
And here are those same people behind a subscriber wall. That's pay dating in a nutshell.
Match.com's numbers are just as grim. They're a public company, so we can get their exact subscriber info from the shareholder report they file each quarter. Here's what we have from Q4 20094:
Remember, sites like Match and eHarmony are in business to get you to buy a monthly subscription. There's nothing wrong with profit motive, but the particular way these sites have chosen to make money creates strange incentives for them. Let's look at how the pay sites acquire new subscribers:
As you can see from the flow chart, the only way they don't make money is to show subscribers to other subscribers. It's the worst thing they can do for their business, because there's no potential for new profit growth there. Remember: the average account length is just six months, and people join for big blocks of time at once, so getting a new customer on board is better for them than squeezing another month or two out of a current subscriber. To get sign-ups, they need to pull in new people, and they do this by getting you to message their prospects.
If you're a subscriber to a pay dating site, you are an important (though unwitting) part of that site's customer acquisition team. Of course, they don't want to show you too many ghosts, because you'll get frustrated and quit, but that doesn't change the fact that they're relying on you your messages are their marketing materials to reach out to non-payers and convince them, by way of your charming, heartfelt messages, to pull out their credit cards. If only a tiny fraction of your message gets a response, hey, that's okay, you're working for free. Wait a second…you're paying them.
Now let's look how this skewed incentive affects the dating cycle, especially on sites like Match.com, where it's possible to for users set their own search terms.
Even more so than in real life, where fluid social situations can allow either gender to take the "lead", men drive interactions in online dating. Our data suggest that men send nearly 4 times as many first messages as women and conduct about twice the match searches. Thus, to examine how the problem of ghost profiles affects the men on pay dating sites is to examine their effect on the whole system.
There are two facts in play:
- When emailing a real profile, a man can expect a reply about 30% of the time. We've conducted extensive research on this, and you can read more about it our other posts. Let's couple this 30% reply rate with the fact that only 1 in every 30 profiles on a pay site is a viable profile. We get:
3/10 × 1/30 = 1/100
That is, a man can expect a reply to 1 in every 100 messages he sends to a random profile on a pay site. The sites of course don't show you completely random profiles, but as we've seen they have an incentive to show you nonsubscribers. Even if they do heavy filtering and just 2 of 3 profiles they show you are ghosts, you're still looking at a paltry 10% reply rate.
- There is a negative correlation between the number of messages a man sends per day to the reply rate he gets. The more messages you send, the worse response rate you get. It's not hard to see why this would be so. A rushed, unfocused message is bound to get a worse response than something you spend time on. Here's a plot of 12,000 male users who've sent 10 total messages or more.
The effect of the second fact is to magnify the effect of the first. For a user trying to meet someone under such constraints, a feedback loop develops. Here's what happens to the average guy:
Basically, because the likelihood of reply to each message starts so low, the average man is driven to expand his search to women he's less suited for and to put less thought (and emotional investment) into each message. Therefore, each new batch of messages he sends brings fewer replies. So he expands his criteria, cuts, pastes, and resends.
In no time, the average woman on the same site has been bombarded with impersonal messages from a random cross-section of men. Then:
Finally, in the spirit of "don't take my word for it", here's how eHarmony and Match.com themselves show that their sites don't work.
This is from Match's 2009 presskit:
Okay, Match is double counting to get "12 couples", since a couple that gets married also gets engaged. So we have 6 couples per day getting married on the site, or 4,380 people a year. Let's round up to 5,000, to keep things simple. My first observation is that Match.com made $342,600,000 last year5. That's $137,000 in user fees per marriage.
Now here's where the demographics get really ugly for them.
It turns out you are 12.4 times more likely to get married this year if you don't subscribe to Match.com.
I figured it out like so:
Remember this is the minimum ratio, because from Match's perspective, we've made a lot of very favorable assumptions along the way. And it also doesn't matter that some portion of Match's customer base is overseas, because however you account for that in their subscriber base, you also have to adjust their marriage total accordingly.
eHarmony seems to do quite a bit better than Match, claiming in their ads to marry off 236 people a day:
Their higher rate shouldn't be too surprising, because eHarmony's entire site philosophy centers around matrimony, and furthermore that's the primary reason people go there. It's explicitly not a place for casual daters.
As they've told us, their member base of 750,000 people turns over every 6.5 months, which means that nearly 1.39 million people go through eHarmony's "doors" each year. eHarmony fails at least 93.8% of the timeFrom the ad, we can see that just 86,140 of those subscribers get married, a mere 6.2% of the people who paid the company to find them a mate. And what of the other 93.8%, the 1,298,475 people who do not get married and then leave the site? Those people paid an average of $190 each for a personality quiz.
A major selling point for the big for-pay dating sites Match and eHarmony is how many millions of members they have, and they drop massive numbers in their press releases and in talks with reporters. Of course, there's a solid rationale to wanting your dating site to seem gigantic. When people look for love, they want as many options as possible.
However, as I've shown above, the image these sites project is deceiving. So next time you hear Match or eHarmony talking about how huge they are, you should do like I do and think of Goliath—and how he probably bragged all the time about how much he could bench. Then you should go sign up for OkCupid.
- Looking for a Date? A Site Suggests You Check the Data
- The Big Business Of Online Dating
- eHarmony.com's Advertising Splash Page
- Match.com's Q4 2009 Report
http://files.shareholder.com/downloads/IACI/871220273x0x349618/6d370897-220b-409b-a86e-e02801b3eed5/Gridsand MetricsQ42009.pdf. Match.com's 20 million membership claim is here: http://www.consumer-rankings.com/Dating/#table
- Centers For Disease Control
http://www.cdc.gov/nchs/fastats/divorce.htm. Not sure why they care.
- The U.S. Census "Unmarried and Singles Week"
372 Responses to “Why You Should Never Pay For Online Dating”
2014 experimenting on users
In 2014, OkCupid revealed in a blog post that experiments were routinely conducted on OkCupid users. The site revealed that one experiment included removing users' profile pictures on January 15, 2013 ("Love is Blind Day") and analyzed user responses to messages, conversations, and contact details. When the photos were restored, users who had started "blind" conversations gradually began tapering off their conversations, leading OkCupid's CEO Christian Rudder to remark "it was like we'd turned on the bright lights at the bar at midnight". In a separate A/B test, OkCupid used a placebo number instead of users' true match percentage. The results suggested that doing this caused users, who were "bad matches" under the original algorithm, to actually like each other: "When we tell people they are a good match, they act as if they are."
The revelation that OkCupid conducted these experiments on users led to criticism. Rudder attempted to defend the company, in part by suggesting that it would be unethical not to experiment on users:
I think part of what's confusing people about this experiment is the result ... this is the only way to find this stuff out [what actually works for a dating site], if you guys have an alternative to the scientific method I'm all ears.
2016 data scraping and release
In May 2016, a team of Danish researchers have made publicly available the "OkCupid dataset" project, containing (as of May 2016) 2,620 variables describing 68,371 users on OkCupid for research purposes (e.g., for psychologists investigating the social psychology of dating). The data release spurred criticism, and an investigation by the Danish Data Protection Authority.
2017 switch to using real names from pseudonyms and subsequent backflip
In December 2017, OkCupid rolled out a change that would require users to provide their real first name, in place of a pseudonym as was previously encouraged. Although the company quickly backflipped, saying that nicknames or initials would be acceptable. The announcement was received by widespread criticism and condemnation for potentially raising the risk of harassment of individuals, especially women, and minorities to doxing. It was pointed out that, unlike other dating sites that encourage the use of first names, OkCupid "encourages long profiles full of intimate details, including candid answers to questions about sex and politics", making connecting that information with a real name more problematic to users.
In 2017 OkCupid reported on Twitter that they had removed Christopher Cantwell's user profile for being a white supremacist after a woman reported receiving a message from him. This raised questions from some users who wondered about the ease with which the company could eliminate users from its platform.
User photos for data mining
Clarifai, an A.I. start-up, built a face database with images from OkCupid, due to common founders in both companies.
2019 alleged credential stuffing incident
A February 2019 report alleged that many users reported lost access to their accounts in a manner consistent with either a data breach or a widespread "credential stuffing" incident. "Credential stuffing" describes using passwords stolen from one service (like another dating site) to attack another service, on the assumption that many people will reuse passwords across websites. OkCupid denied any data breach or system errors.