Gary Stock: UnBlinking

Bad Chemistry


Over the past ten months, I've wasted dozens of hours and several hundred dollars discovering the many ways online dating site is broken. With design so poor, algorithms so lame, matches so weak, and support so unhelpful, I'm astonished they remain in business.

As a paid member, I wrote to technical support repeatedly, explaining the most egregious errors, and recommending corrections. No response. So, I documented a few of the problems here, hoping might be inspired to pursue some quality control.

I've tested other online "dating" and "personals" sites, and plan to review several of them soon. But, understands neither how to make technology function, nor how to let people function. Writing this standalone review is rather like ending a clinical trial early because one regime leaves most patients much more ill: might work for you, but for most people there are far more effective options. No other major site is as flaky as Chemistry.

Please use QuickTopic to share your comments. [Usability note: All links open into the same target window.]


What is wrong with these people?!?

You'll be asking yourself that question as you slog through the profiles at many online dating sites. You may ask it about the members, or about the site designers -- likely both. Generally speaking, there is little good news in this arena. Let me offer some perspective, if that's possible in such a dreary environment.

First, everything here is my own opinion, based on my own experience (although most of the design challenges apply uniformly to all users). It seems that other people may have met great partners at I strongly suspect they'd have been just as successful by throwing a dart at phone book! However, no one is testing and documenting that strategy on thousands of people every day -- and the methodology would be touchy -- so we may never know.

Second, I agree that a funky website is a relatively trivial issue in the scheme of things. There are millions of poorly designed sites! (And, there are many things worse than poorly designed websites.) But, few web services advertise nationally throughout the media, as has done. Few have the potential to waste so many people's time and money. Few touch their users with such potential unpleasantness, by intentionally creating hope in poorly implemented technology. Chemistry seeks to be a major public venue, and claims to help fulfill fundamental human needs -- thus the service needs to live up to its hype.

Finally, yes, I did meet a couple of interesting women through However, I had to slog through nearly 3,000 profiles to find them. Generally speaking, about ten percent of those profiles prompted me to show interest. Among those 300, about ten percent completed the excruciating process of questions and answers to permit open communication between us. Among those 30, about ten percent were still interesting enough to me that I wished to meet them. People who have thought clearly about what they want in a relationship likely will see a similar annihilation at each stage. In my case, it cost about 100 hours of online effort, and about $100 in Chemistry fees, to discover each woman worth meeting. could warn users about that up front -- and admit they'll stretch the process out unnecessarily (though profitably for them) over many months. I could have accomplished review of 3,000 profiles easily in a few days on many other dating sites. However, Chemistry refuses to let a user search profiles. Since the maximum possible number of matches is ten per day (and it is easy to fall well below that volume) users are captive for many weeks, in a scheme that must profit immensely. My irritation with the service is not merely that it is poorly designed and expensive. What makes the service unbearably bad is their insistence on stretching that unpleasant, cryptic, and awkward experience out over months.

Here are dating sites I could recommend -- only because they are somehow better than -- damning each with faint praise in its own peculiar way.

  • E-Harmony is every bit as expensive and inexplicably random, but less broken. Like, you'll have met the vast majority of your best matches in the first 30 days. (Both sites apparently reveal your best matches among their existing membership in dribs and drabs over the first several weeks. After those, you'll match only among their few new members each day, or the dregs of their database.) E-Harmony rarely delivers more than seven potential matches per day, but, it's not quite as poorly designed or implemented as Chemistry. If you have no idea what sort of partner you want to meet, or you aren't very dedicated to the quest, e-Harmony is worth a try.

  • seems slightly more competent and attracts more varied women. It's from the same company as -- but it's also considerably less expensive. You'll be able to search some criteria, across larger areas geographically, and even keep track (slightly) of who interests you! Compared to, Match reveals so much more information about potential matches that it may belong in an entirely different category. Unfortunately, the proportion of abandoned or poorly attended profiles is equally high.
    I was regularly astonished to see solid evidence that steals from the database to fulfill promises. Frequently, match profiles ask you to "Send e-mail" -- a concept which is available at Match, but not at Chemistry. Several profiles delivered to me at stated explicitly, "I'm using because...". So, save some money, and go directly to Match.
  • OK Cupid is the rare quality site that does what it is designed to do, and allows you to search for who you want, based on what the members declare about themselves. I suspect OK Cupid's service functions so well because it was designed by people who know how technology functions, and have observed how people function. (The absurdity of the interface tempts me to think no one there has ever watched a human being try to use their service.) By the way, not only is OK Cupid better than the big names, it's entirely free! Oddly, and unfortunately, it is not as highly populated as it deserves to be.

  • Several other sites deserve honorable mention on certain elements, such as Plenty of Fish, for quality design and communities, or Yahoo! Personals, for quality profile data and members. Please note which two elements merit that adjective -- "quality" -- for each site. Those honors do not cross over!

Chemistry is not the donut. It is the hole.

The most obvious system design failure at you believe you know the geographical area being searched on your behalf -- but, you don't. offers no documented option for finding a match more than 250 miles away. Your profile choices include "Preferred distance (miles)?"

  • 25 miles
  • 50 miles
  • 100 miles
  • 250 miles

Apparently, the perfect match for you already lives within 250 miles of you! Lucky, huh?

Of course, if you can find the "My Account" page, and change your "Zip/Postal Code" to some other location, you can search within 250 miles of that other (entirely bogus) location. You'll just have to remember what zip code you claimed to live in -- especially when you find yourself matched with a person who is already skeptical of your identity. In general, female users of are far more skittish over that issue than users at any of a dozen other sites I tested. And, man, do they love their freakin' golf!

Whether you discover how to spoof your zip code or not, the underlying flaws in design and implementation are inexcusable. If you select a search radius of 250 miles, you will NOT receive matches 100, 50, or 25 miles distant. Let me emphasize this, because at first reading it seems unbelievable:

When you select one search radius, you will not receive any matches for a smaller radius.

That is, if you are willing to look as far as 100 miles from home -- and commit to that willingness -- you will never see matches 50 or 25 miles from home. I am absolutely confident that 90% of all users do not understand this. I strongly suspect no one at Chemistry even knows it is true. If they did, I suspect they would pull their own heads off in humiliation. (Once they realize it, I believe they'll try to fix it -- fast. But, it was undeniably true from mid-March to late-December 2008, and thus likely has been true forever. That kind of bug is built in at the beginning, not in later versions.

Experienced programmers will recognize the error as one they laughed at in Programming 101 -- likely in elementary school. The query for 250-mile matches should include all matches less than 250 miles distant (but at Chemistry, will not). The 100-mile query should include all matches less than 100 miles away (but will not)... and so on. I've demonstrated this again and again, every week since I began using the service and immediately noticed the flaw.

The demonstration is simple, because the database seems so small, thin, and relatively low in quality. By setting any meaningful criteria (such as an age range of only a decade, non-smoking, and athletic body type), you can very quickly exhaust all valid matches in any radius. For example, query the system for several days at a 100-mile radius. Very soon, you'll have seen all possible matches, and begin to receive only one or two matches each day, representing only new members.

In that exhausted environment, when you click "Refresh Matches," the system duly reports that there are none. To demonstrate the fatal flaw, immediately change your search radius to 50 miles, and [Save] that selection. Refresh matches again; more likely than not, several will appear which were invisible to you when searching 100 miles. If the 50-mile "donut" also is vacant, change your radius to 25 miles, and refresh again. Those matches found at 50 or 25 miles were absolutely not visible to a search of a 100-mile radius. Welcome to Bizarro World!

I did conjure up potental interpretations of these incomprehensible constraints. So, if the folks at won't admit the error, they could consider posting these as "Features." (In the intelligence community, this might be called "plausible denial.")
  • If you choose to search a 50-mile ring around home (thus, nothing closer), other members who know you from work, or around town, won't discover that you have a profile. Workplace embarrassment averted!
  • If you choose 100 miles, the spouse you're cheating on is unlikely to wander in on your liaisons so far from home. Divorce court averted!
  • If you go for 250 miles, most matches you actually meet will be unlikely to become a stalker -- too much travel involved! Boiled rabbit averted!

Good thinking, Dr. Fisher!

Mommy! Make the bad lady go away!

The dead-distance-donut-debacle alone makes almost irredeemably bad. However, the interface design enjoyed the same loving care and attention to detail as the underlying code and algorithms.

The most relentlessly painful design failure: to refuse a proposed match, a user is required to click six or seven different commands. There simply is no other commercial site, anywhere on the web, so poorly designed. Why? Perhaps competent people designed those sites, or their support staff will pay attention to users' suggestions -- neither of which is true for

Because it is such a powerful indicator of the laissez faire, nonsensical approach takes to most functions, let's walk through what's required to get rid of a really, really bad match -- several of which Chemistry will try to foist on you daily.

First, though, as proof of concept, let's see how "competitor" handles this function. Yes -- another service from the same company that owns Apparently there was one competent interface designer in the firm (at some point).

Does she spark your interest?  (Click Yes, No, or Maybe to move on to your next match.) At, when you view your "5 matches" for the day, you see a single menu directly adjacent to the user's photo and headline, saying, "Does she spark your interest?" (Click Yes, No, or Maybe to move on to your next match.)" Immediately below are green [+ YES], red [X NO], and yellow [? MAYBE] buttons. Simple. Functional. Quick. That single click records your decision, updates your account, and displays the next match, if any. One click. Take note, Chemistry: it can be done.

In contrast to such basic utility, at you can not and will not receive another match until you take all six of the obtuse steps listed below to get rid of the current matches in your account! I implore Jakob Nielsen and devotees to verify that your health insurance will cover work related strokes -- or else, stop reading now!

Thus warned, I present the astonishing process of ejecting a bad match from your "New Matches" folder. I will wander freely, in italics, a sample of the myriad threads that comprise the majestic tapestry we call

STEP 1. You must drag a slider between "FIZZLING" and "SIZZLING." [gagging sound] Please, let's ignore the gutteral notions of "Fizzling" and "Sizzling." Don't get me started! Just accept that you need to drag hard left to start the grueling process of removing a single match. Action Needed
Tell us your interest level for her to keep receiving new matches.
Move the slider in either direction and click continue.

If you do not drag this slider -- if you leave it in the default middle position, which would seem to indicate neutrality -- your only subsequent choice will be to accept the match. That is, the default setting (slider in the middle), does not permit you to decide later whether you are interested: the presumption is, "I'm interested." How peculiar! The matches are so random, you will almost never be interested.

STEP 2. You must click [CONTINUE] to register your fizzling slider action.

Action Needed

You have shown LOW interest in Debbie. How would you like to proceed?
Move Debbie to your Archives [ARCHIVE].
Move Debbie to your Active Matches [MAKE ACTIVE].

STEP 3. On a new page, you must remind Chemistry that you just ranked this match LOW.

You soon will learn to rank every less-than-ideal match very low on the Fizzle to Sizzle scale. If you bury the slider at the Fizzle end, Chemistry is able to figure out, "You have shown LOW interest" (though it requires at least five more actions to confirm). (Conversely, if you peg the slider hotly into Sizzle, Chemistry notes, "You have shown HIGH interest" -- though it requires two more clicks to confirm.) However, if you drag the slider anywhere else, Chemistry presumes you have shown "MODERATE" interest, and refuses to offer you the option to archive the match. That is to say, if you rank the match only slightly toward the left, fizzling end of the slider, Chemistry will not offer you the option of archiving that match. Your only option will be to make them active, and show interest in them. Hmm... is this where the absurd pop trope, "Love me or Hate Me!" got started?

Severe digression: here's one horrific behavior I'd forgotten about from the early days, and just stumbled upon anew. If a match has previously shown interest in you, they appear on a separate menu, as "Interested" -- not among your "New Matches." If you make the mistake of clicking "Continue" without sliding the interested party to "Fizzle," you have instantly shown interest in that match, and can not remove them from your active matches for a full 24 hours! Yes, it takes six or seven clicks to take a match off your active list, but you need not even confirm putting them on your active list!
Worse still, there is a third menu (beyond "New Matches," and "Interested") for when "You've been noticed" by someone. Hmmm... how does "You've been noticed" differ from their being "Interested?" There's no explanation! If "You've been noticed" by someone, you have no way to delete that reminder. There are two buttons next to their photo: [BACK] and [SHOW INTEREST]. Once "You've been noticed" that profile remains in your account indefinitely, and you can do nothing to remove it! It appears they may remain "Interested" in you until their account is closed.

So, what do you think of New Member?
SHOW INTEREST. As long as we're in this downward spiral of digression, here's a stunner: it is actually possible to leave a profile name as "New Member." There is no requirement to select any screen name! Thus, every day or two, "New Member" will appear in your daily catch.

In the unlikely event that you left the Fizzle Stick untouched, or moved the slider higher, you still have to confirm. Otherwise, you must confirm that you already told Chemistry to Archive this match. (I'll save everyone the indignity of explaining the convoluted gyrations if you decide to "Go Back" for a do-over.) I am sorry to report that the implementation is at least as convoluted and difficult to comprehend as my articulation of it here in text.

STEP 4. On a new page, Chemistry asks you to report why you're closing the match. Your choices: Action Needed
Are you sure you want to archive Debbie? This will stop all communication with her.

Do not archive and return to the Previous Page
Move Debbie to your Archives.

  • No photo
  • Not a good fit
  • Different interests
  • Not enough information in profile
  • Educational differences
  • Professional differences
  • Religious differences
  • Marital Status
  • Doesn't Want Kids
  • Pursuing another relationship
  • Taking a break from dating
  • Did not respond
  • Difference in age is too great
  • Communicating outside of Chemistry
  • Physical distance is too great
  • Prefer a non-smoker
  • Prefer a non-drinker
  • Different relationship essentials
  • No chemistry based on short answers

Consider the brokenness of the system revealed by this list.

First, it appears to be in order of most-likely to least-likely responses, perhaps from data gathered from users. For example, approximately half of all matches delivered to me had no photo. That Chemistry offers no option to exclude profiles of matches unwilling to expose their appearance (as would be an option at most other dating sites) seems incomprehensible. Unless... knowing how small their database is already, Chemistry may include the useless profiles to get credit even for matches they know I'll never select.

I am so amused by Chemistry's explanation of this particular exercise in futility:

Our system is designed to continually learn from your feedback: who you like and who you don't like. Your input is automatically integrated into our matching system so that over time we get a more complete picture of who you are and the kind of person you're looking for.

Can you hear me laughing? For three months, I automatically and intentionally archived EVERY profile self-identifying as "Conservative" or "Ultra Conservative." The volume of those matches never fell, remaining two or three of every group of five. I even changed my own identity to Very Liberal, hoping that would raise the bar. Still no benefit. Would you employ a matchmaker who would pair a self-identified "Very Liberal" with a self-identified "Ultra Conservative"? Does that sound like the work of a shrink who understands people, or relationships?

Here's another brain-buster. Many users refuse to post a photo, and some get downright agitated if you make the mistake of asking about the issue. In many people, a few specific personality traits must be behind that decision -- extreme shyness, low self-esteem, conflicted privacy beliefs, and so on. sings the praises of their own profiling questionnaire, so those personality traits must be known to the database -- right? So, I tried ranking photo-free profiles "Fizzle" for a month, but their volume remained just as high. So, whatever else the personality profile discovers, it doesn't discover profound shyness, poor self-esteem, general paranoia, or related issues. Hmmm... those couldn't ever become problems in a relationship, right? Right?!?

Consider other options for explaining your decision: "Prefer a non-smoker" and "Prefer a non-drinker." These options are already set within your profile. If you've asked for non-smoking, non-drinking profiles -- why are smokin' drinkin' fools appearing among your matches?

Luckily, you are not required to choose a reason for closing the match -- you can simply ignore this step. That's lucky for you, but not for anyone else. At e-Harmony, for example, you must choose a reason (from a list that is just as incomplete), and your reason will be revealed to the person whose match you closed. In that setting, at least everyone can learn over time why they're being closed -- perhaps they're searching the wrong age range, appear too religious, need to add a photo, or whatever other reason may be cited frequently. At Chemistry, 100 men could close a woman's profile for the same reason -- something easy to correct, for example, like "Not enough information in profile" -- yet that woman would never know why.

I still have to laugh, because most of the real reasons you close a match are not on the list, anyway. A few suggestions for the missing reasons I could have used at least daily:

  • This person can't spell, but I can.
  • No one could possibly be that boring.
  • I prefer not to meet more scam artists.
  • Are you serious with all the NASCAR photos?
  • Please forward this member to
  • Stop sending matches who think golf is public service.
  • Arrgghh! My eyes! Medusa in the house! Make it stop!

See, now, if offered those options, I might understand why they wouldn't reveal them to the match I closed!

We're sorry there wasn't a spark!

Blessed has been moved to your Archives.

If you decide later that you would like to continue getting to know her, click on 'Unarchive' next to her name in your Archives. STEP 5. Click [ARCHIVE] again. True, this step does nothing. You just need to confirm what you already confirmed. For the second time. Or the third. I forget.

After four different buttons, three different pages, two confirmations, what's left? Is there a partridge involved?

STEP 6. You must click [BACK TO NEW MATCHES] to start the tedious effort again.

Keep in mind, the system knows you have more matches, but doesn't deliver you there. You actually have to ask specifically to see them.

I must admit, not delivering you back to your new matches is a defensible interface behavior somewhere between 20% and 80% of the time. Why? Because the system doesn't bother to check to discover whether any exist. So, you'll be offered the chance to go [BACK TO NEW MATCHES] whether any new matches exist, or not.

Synergistic effects of chronic exposure to hazardous chemistry

The list of errors and inappropriate functionality at is vast. Here are a few others that may shock and awe the seasoned netizen.

My photo is not offers no option to exclude matches who won't share a photo. Thus, you have no idea what 30% to 50% of all your matches look like -- the matches you're paying to review. Fine, those folks should have the option not to post a photo. But, other users should have the option not to waste time and money rejecting such profiles day after day. It is possible to process hundreds of photo-free profiles in a single month! Most sites make it possible to exclude the too-nervous, too-married, or too-unpleasant-looking among us from searches. Chemistry does not, perhaps so they can exploit a larger database, even though the profiles are rarely of interest. The site offers no explanation.

It is possible to store some sort of "null" primary photo -- so that your profile appears at first to have no photo, even though several may exist. Here's an example, in which thumbnail views of the member's photos reveals that the "Primary" photo is an image that says simply "My photo is not available." My photo is not available.  And even if it were, you can't see me in it.

Even if a user offers a photo, they have no obligation to provide a photo that means anything. Many sites reject photos that do not show the user's face, or refuse to permit a photo with other content to be your primary photo. Note what passes for a "primary" photo at! This user's only photo is of no value -- not even for evaluating their golf swing -- but at least it suggests they know how to upload a photo. Bear in mind, this is the only photo this match provided to help others evaluate them as a potential partner. Thus, it is better than some one-third of the profiles at, which have no photo at all. My photo is not available.  And even if it were, you can't see me in it.

• Users may set their profile to prevent viewing of what few photos do exist. I can see the possible value of the more common setting, "Not just yet... see my photo after the next step." This might be useful, since the "hidden" partner could decide whether to respond to the "seeking" partner, and determine whether to take the next step, thus revealing their photos. However, consider the exceedingly more paranoid setting, "See my photo after Step Four, First Meeting," which I observed several times. Seriously. Think about it. Not just yet... see my photo after the next step

• If you ever do achieve direct communication with another member, messages are stored in your "Inbox" -- which is hidden two menus deep from your home page. If you find your Inbox, each older message is headed by a small flag icon, yellow or red, indicating that the message "Will be deleted" in 7 or 14 days, respectively. As of December 20, all my messages back to April were still present, all happily waving their tiny yellow and red flags.

• If you use the default credit card payment option, will silently extend your subscription until you phone and tell them to stop. Terms of Service apparently permit them to continue charging you indefinitely without even notifying you that a charge is upcoming, or was just made. This abuse is common at gutter-level sites, but it's astonishing Chemistry would stoop so low. Expect to be billed until you ask them to stop. Frankly, if Chemistry had not pulled this stunt with me, twice before I noticed it -- and refused to reverse the billing -- I might not have bothered to write this review.

I recently phoned to express my irritation about those silent, recurring credit card charges. A bored young woman drearily recited their TOS authority. I said it was a despicable practice, and that her rote defense proved mine to be a familiar complaint. She was unmoved when I warned, "If you won't correct this, I'll write about your practices online." This is the first time in 14 years of online life that I've made such a promise. That alone should provide some perspective on how frustrating can be, compared to other poor experiences available online.

Perhaps I've made my point, and will have the sense to fix their algorithms, interface, database, policies, and business practices -- they're welcome to ask for my help with many of these issues. On the other hand, if gripes about my review, maybe I'll set up a site dedicated to advertising their errors, with a special forum just for disgruntled users! It'd be a spot where we could organize our class action effort -- to get all our money back. I have a feeling people would pay good money to join! Unless Chemistry is foolish enough to want me to set that in motion, let's just chat for fun... using Steve Yost's cool gizmo, Quicktopic.


Privacy Policy   -   Site Copyright © 1995-2009 by Gary Stock. Commercial use prohibited.
May be excerpted, mailed, posted, or linked for non-commercial purposes.