How we came back after being shut down in four days | Hacker News Reader

How we came back after being shut down in four days

blog.leaky.com

203 pointsjasontraff86 comments

I love this story, because it involves bullying from big companies turning a marginally shady† business model into a genuinely interesting and valuable above-board offering. These stories so often end in shrill shoulda-coulda ranting and uninformed legal speculation. This one ends in a great hack and an even better company than the one that originally launched.

† I don't think the Leaky people are shady, just that scraping hypersensitive sites is a duct-tape-and-baling-wire solution.

It also pivots them from a thin affiliate model, where they add no value above the traffic stream, (and fight other people doing same with less ethics/stature) to one where they have a competitive moat and produce value independent of insurance purchases. They're essentially a publicly accessible interface to "transparent" documents mandated by the regulations which, at present, might as well not exist outside the industry.

It's almost like they're a patch to the law, really.

Let me preface this by saying that I am not trying to troll, I genuinely want to know if I am missing something. How does the pivot remove them from the affiliate model? Granted, I applaud them for rewriting it and "doing it the right way", but does it change the way they could make money?

"Thin affiliate" is a term of art, roughly meaning "a site which adds no user value, just taking one or more clicks to delay your task success." They've made a lot of people $$$ but Google hates them and they're competitively weak relative to other affiliate plays where there is a value add. (Mint is most obvious example that springs to mind at 5:30 AM.)

Don't worry, I almost always assume good faith on HN.

They basically have original knowledge they can sell, now, versus "just" scraped data. They can license their price prediction engine to anyone who doesn't want to go through the agony of doing what they just did.

And this does have real value. Myself and 2 other friends were this close to doing exactly what Leaky is doing, but I (the programmer) did not want to spend a lot of time implementing that price prediction model off of hard-to-read (and constantly changing!) legal documents.

From the previous HN post a few days ago, multiple people raised the concern that Leaky quotes them at twice what they are actually paying. Someone else said that without looking up a credit score the rates are basically worthless. So I still have my doubts as to the accuracy of Leaky's data, even though they say it's within 1%. Maybe that's 1% given some default middle-ground credit score?

I was told that the rates insurers file change frequently. So I'm also interested to see how quickly Leaky can stay on top of up-to-date results. That's my second accuracy-related concern.

If these concerns did not exist, I'd already be in this business. I'm sure the same can be said of other people. In fact, I assumed that with such an obvious need for a business like Leaky the absence of such suggested a higher barrier to entry than is visible at first blush. I think what happened in Leaky's case is that they were too naive to know they were being naive, and then kept pushing on it anyway. Isn't that a Steve Jobs quote or something?

I'd respond, but you look like a competitor! :)

Isn't that a Steve Jobs quote or something?

I believe it's Emerson. But Steve Jobs might well have adopted it, like the fake Leonardo quote ("simplicity is the ultimate sophistication" - actually William Gaddis) and the fake Picasso one ("great artists steal" - actually T.S. Eliot).

> actually T.S. Eliot

Almost! With context: "One of the surest of tests is the way in which a poet borrows. Immature poets imitate; mature poets steal; bad poets deface what they take, and good poets make it into something better, or at least something different."

I'm also a fan of the way Gilles Deleuze framed it: "Proust says: 'Great literature is written in a sort of foreign language. To each sentence we attach a meaning, or at any rate a mental image, which is often a mistranslation. But in great literature all our mistranslations result in beauty.' This is the good way to read: all mistranslations are good -- always provided that they do not consist in interpretations, but relate to the use of the book, that they multiply its use, that they create yet another language inside its language. 'Great literature is written in a sort of foreign language...' That is the definition of style."

Not almost; unequivocally. That the wording got polished over time is just what happens in popular discourse.

I don't see the connection to Deleuze. Eliot isn't talking about mistranslation or about how audiences receive art. He's talking about Dylan lifting the melody to "Don't think twice" from some other song.

Yes, almost. Eliot almost said that -- said something superficially similar to that -- but didn't say that. People use "great artists steal" as incorrectly as they use "information wants to be free." The line hasn't been polished; it's been co-opted and transformed. Which is exactly what he and Deleuze were interested in, and why they so often circle the same drain.

Of course how audiences receive art and how artists create it aren't disjoint processes, but I understand you getting hung up on Deleuze's use of the word "mistranslation" there. It helps to know that he's working out of a tradition that treats the idea of translating as a figure for "owning" (in the Heideggerian sense, also translated "enowning").

What I meant is that there is no other plausible lineage for that quote - and I've looked. If you claim otherwise, I'd love to see the citation.

It doesn't help to replace one obscure jargon with another, even more obscure jargon.

Absolutely that's the provenance of the saying. Eliot just wasn't saying what the saying says =).

I am just wrapping up a test on canned homeowners' insurance quotes on our local real estate site, LA Life. Have a look at an example:

http://lalife.com/address/15436_Mayall_St_Los_Angeles_CA_913...

We got involved with a broker who actually ran these quotes, property by property(!) So far it's been pretty dismal, just due to our typical users' interest in the product. But if anyone has some ideas, I'm all ears, email my username at lalife.com.

How would Google recognize they aren't a thin affiliate though?

I think there is also a deeper discussion here on what it means to be transparent. Insurance companies C&Ding them so hard is evidence to me that they are trying to avoid transparency, and the /20,000 pages/ says that their algorithmic pricing models aren't really all that transparent in the first place.

They're as closed as their systems. The information is out there. Now wonderfully creative folks have found a way to make the information more transparent.

I have to say, I'm surprised they managed it. I would not expect a site that was C&Ded out of existence so fast ever to reappear.

Thanks again for all of the support

Great idea for a product, but the privacy policy here is seriously inadequate.

In order to use this, I would have to give a lot of personally identifiable information. The privacy policy basically gives Leaky the right to do whatever they want with my personally identifiable information, such as providing an unlimited set of third parties for "internal business purposes," an unlimited number of present or future affiliates which they will "seek to require" to adhere to the privacy policy, to anyone who may purchase the company or its assets, or for a fairly extensive list of fraud and law enforcement reasons, to disclose it to an unspecified set of entities.

On top of that, there is no limitation on how long Leaky will hold such personally identifiable information, nor any information about how to get Leaky to remove it. So basically, once I enter information into Leaky, they may hold onto it forever and sell it to whoever they like for quite a wide variety of reasons, with no recourse for the user to protect their information.

I don't think that I would ever want to put such personal information in a service with such a weak privacy policy. And given the new model, in which this is just an implementation of the complex pricing structures used by auto insurance agencies, I wonder why this couldn't be made into an application that runs locally and never exposes my personal information to anyone else. Yes, you would probably have to go with a paid application, rather than an ad-supported web site, with all of the problems that entails, but for a product like this with such important privacy concerns and such little value (to the customer) to putting the data online or sharing it with people, it really seems like it would make more sense as an offline application, not a web app.

Sure, they could do with a better privacy policy.

However, what exactly is the sensitive data here? Your first name, your zip code, your car make/model/year, and your profession.

99.99% of people would happily give you that data for a free pen, just for the asking.

Also, while the Leaky guys (being in YC) are likely to actually be better than their privacy policy, what happens when you click over to the actual insurance company and give them your full details (SSN, credit info, vehicle accident history, license info, etc.). I'm more concerned with how AIG will use my data than how a small startup will, and the insurance company is the one collecting vastly more -- and their privacy policy allows them to use it for all sorts of things which are opposed to the user.

So, I haven't actually tried the service (due to my privacy concerns, as well as just time). I was assuming they would use more information than that (credit info, vehicle accident history, etc) if they were going to do a good job of modeling how car insurers do pricing.

The amount of information you mention is not as troubling (though you missed your email address), but I would still like to know that they will delete that information after they are done with it, not retain it forever while maintaining the right to give it to pretty much anyone they want as long as they are "affiliated" with Leaky.

And yes, a YC startup may be more trustworthy than an insurance company, but on the other hand, the insurance company needs to collect that information, and is much more heavily regulated, while Leaky, as a private company under not much regulation, can do pretty much anything it wants.

I tried it and found that it underestimated my insurance cost by exactly $500 with both Geico and Allstate but interestingly the spread between the two was correct (about $100)

I've noticed some other really oddities (and the pricing is way off for me, unfortunately -- if only I could get a policy for what they're saying!).

For example, if I list my occupation as "Consultant", Leaky says Geico will give me a $150 discount. Bizarre.

Edit: Clarifying that Leaky says Geico will give me a discount, not that Geico actually does.

Auto insurers will definitely give you a discount for certain fields of employment. "Engineer" drops it $100 or so every time I've tried. My business cards sometimes say "Software Engineer", but I dropped out of a math program, so it's pretty bogus.

Haha, yes but... Consultant. I hope somebody in Hartford, CT (or equivalent) at least finds the notion of giving a discount for having a title which is almost entirely devoid of meaning.

Leaky guys -- I'll try to ring Geico and tell them I'm a consultant so I'd like my $150 discount. Will let you know how it goes.

sakai -- We take our accuracy very seriously; email us at sayhi@leaky.com and we'll look into it.

If you're in California (the only state that we officially support), we're within 1% in about 50% of cases and within 3% in about 80% of cases - and we're always getting better. However, there are still occasional outliers that we are always trying to identify and fix.

If you're outside California, well - the prices will be different. However, the magnitude changes (eg for switching cars) will be similar to what you would see in your state.

As for the occupation discounts - those come straight from the insurers' filed documents.

I haven't read your comments (too long, no time) but I think you should fix them. They seem like they might be critiques of a product you haven't even tried.

If you haven't read them, how do you know they are critiques of a product he hasn't tried?

>"a YC startup may be more trustworthy than an insurance company"

This, and that it is probably true, makes me embarrassed for my country and how awful we allow businesses to be to people.

The fact that the insurance companies acted so swiftly and decisively shows how vulnerable they are. It's pretty obvious that insurance is an industry whose margins are bolstered by consumer ignorance.

I'm glad you guys are spending the time trying to fix this, and it is very impressive how you bounced back so soon.

Most bizarrely almost all of their advertising seems to have the same script: "switch to us and save money". Some even go so far as to mention you can call them when there is an incident, but you'd think that is standard!

The regular wording is really sneaky : "99% of customers that switch to GXXX save money". That's completely to be expected : The only people that have an incentive to switch are those that would save money.

It would be far stronger if the wording was "99% of customers would save money by switching to GXXX" - which is seductively similar, but completely different.

Maybe it is because I'm not a US citizen, but the premium price is the second most important factor to me. The most important is what will happen when there is an incident. The insurance company could quite easily deny you, drown you in paperwork, abandon you to lawyers, make your life miserable etc. Or they could be a delight. Any company can take your money, and any company can take less of your money. That is the easy bit.

I actually pre-emptively switched from Farmers to CSAA because of this. farmersreallysucks.com has lots of reasons why although it seems some industry deployed software named Colossus really helps in annoying customers.

This same thing must be why Americans are so confused over healthcare. Each individual believes they are indestructible so why should they pay for other people's health when no one will ever need to pay for theirs. Pay the least amount in car insurance since you'll never be in an accident.

I think it's something most people haven't learned yet, or at least the advertising hasn't caught up with. After getting screwed by Geico on a claim, I'm certainly more aware of this.

Awesome story. Expect to see state and federal legislators get lobbied to make non-humans reading these published documents illegal (big sad face, the insurance guys have no imagination) but you are doing your customers a great service, and that is what its all about.

What ever you do, don't say "Well if someone offered to buy us for $2B we'd just sell and shut down the site."

Actually, this is an excellent point - get a head start and see how you can get involved with legislators/regulators to pitch yourself as a neutral 3rd party that is essentially checking if insurers are playing by their own rules. This way, you are also becoming a sort of audit tool that legislators never had an easy access to.

Nice angle. Any way to hold them responsible when Leaky's quote does not match the insurer's quote? Force them to make the price 100% reproducible from public documents given the same information on the consumer.

What a great and inspiring story. Damn well written, too. Reminds me of those lost cat stories where the cat somehow jumps out of the car hundreds of miles away and shows up back home against all odds six months later.

I completely understand why you would go back to the thing you were passionate about. It's amazing how ingenuity becomes available when you do that.

By the way, I think you should blog up a storm around your startup. You have a great david-and-goliath story and the ability to tell it. Everyone is going to be cheering for you.

Edit: hmm. After reading http://askolo.com/thomas#4f72682c83a416244b000022, maybe not blog, I don't know. But definitely find a way to keep sharing your story.

Thanks, that's a really nice thing to say

I'm glad you put up the notice about California rates only. I'm in NYC and saw that I was paying nearly double what I should be paying within the same insurer.

My question... Is it possible to vary that much between the different states?

The short answer is yes.

Because insurance is regulated on a state-by-state level, each insurer will submit different models to different states. For example, in California doesn't include credit score and New York does - this alone could change your rates by 60%.

However, while the final prices may not be the same, the magnitude changes that happen to your prices as you make different selections (ie changing your vehicle or get a speeding ticket) will be largely similar across states.

Now that I remember it... I found it odd that it was asking me my profession. I entered Engineer and it dropped substantially.

Anyways congrats! It's really interesting stuff. I have very similar issues to Leaky on my personal finance/banking startup. Really like the way you incorporated the ratings at the end.

Yeah some companies do _really_ big discounts for profession -- engineer being one of the better ones.

In my test, there was almost a $200/yr. difference between Engineer and Software Engineer. I wonder what makes software engineers specifically statistically poorer drivers?

Some companies draw distinctions that sometimes seem arbitrary, which is another reason that we think Leaky is helpful.

However, there's almost always a statistical analysis for the decisions that they make and it may have nothing to do with driving record. It might be that one profession stays with the insurer longer or is a more profitable customer by allowing the insurer to insurer multiple-lines for them (house, life, etc.).

Huh. So now all of us "Software Engineers" will start claiming to be "Engineers" on our auto insurance, and then the insurance companies will have to recalculate their rates to raise them on the "Engineer" category.

Congrats guys, and the site is beautiful. One small thing: The progress bar for the survey is good to have but on the second page (Driver Info), it gets a little long and I was suddenly unsure of how much more time I needed to spend on it.

You could possibly have the progress bar hover like the pricing sidebar, so even as users scroll down to finish their form they get reinforced that it's the last step and the end-point is coming soon.

Just a quick thought -- my headaches from doing insurance shopping previously came from being lost in surveys and not knowing when the end would come. Your form is so simple, it would be good to highlight it more.

Thanks, that's really helpful feedback

I really like this real-time results. This site particularly caught my attention because I spent a couple of years working at a company that basically has been doing this exact thing for nearly 10 years called www.insurancenoodle.com (business insurance instead of car insurance, though). I can't see what Leaky's path to earning revenue will be but there's already a simple, proven business model. It involves becoming a licensed agent so that you can actually write the policies on behalf of the carriers and you earn a commission.

At least as far as business insurance goes, the carriers are happy to have you sell their insurance products as long as you're licensed in the state you're selling and sign up with them as an agent. They rely on independent agents to sell their products, so it's really not a new or unusual business model. But if you're not working with the insurance companies then I have no idea how you would make money except I suppose advertising. It would be a wasted opportunity if Leaky is not trying to become licensed agents for these carriers.

I'm not sure how much things have changed but back in 2002 we were dealing with crazy archaic data feeds and screen scraping. The carriers didn't have any APIs and we spent a lot of our time tuning our scraping and parsing code.

Now that the insurers know you can do this (legally), they should just let you back in (which would improve your accuracy slightly, and make it easier to cover other states).

Once someone's on Leaky, they're going to buy the best priced insurance they can. An insurance company trying to block you basically just guarantees that no matter how good their rates might be, they won't get Leaky customers.

(Also, if you're in California, I strongly suggest bumping your insurance up to 100/300 with high uninsured/underinsured motorist, unless you're utterly broke. It doesn't raise your premium much, and really keeps you from getting screwed. The minimums are utterly insufficient in a place with both lots of expensive cars and lots of uninsured/illegal drivers, even if you drive a $500 Honda.)

This is a great and inspiring story. Would love to hear more about the in-the-trenches technical work involved in the OCR and modeling involved for the insurance industry. Surely there were a few hard earned lessons there that others could benefit from. Keep blogging!

The truly great thing about this is it's probably going to be more reliable. They could've silently detected Leaky's scrapers and served much lower prices to it, but they can't as lie in the insurance filings without committing perjury.

Wow. That's incredible...what a combination of intelligence, passion, and hard work. I don't even have a car, but this makes me want to use Leaky.

Thanks, we really appreciate the support!

This is amazing. I'm really happy for you guys.

Thanks Kurt, we're just really glad that we'll get to see it through this time!

Good Luck!

I'm amazed that you don't have sites like this in the US. Price comparison websites are very popular in the UK. Not just for insurance (car; home; etc) but for electricity bills and cell phones and Internet service provision and etc etc.

There is a problem in the UK of "referral fees" - insurance companies selling on details to no-win no-fee solicitors.

(http://www.telegraph.co.uk/news/uknews/law-and-order/8742611...)

And there have been other dodgy practices by insurance companies.

(http://www.bbc.co.uk/programmes/b014lzz1)

there are hundreds of these...this is just a YC backed one

What I don't get is if you alienate the insurance companies so that they don't want to do business with you, where is your business model? The only potential client for such a service is an insurance company. Or am I missing something?

That's only if you want a piece of the affiliate sales / agent money. You can always go for just ad revenue.

Very nice. I just got insurance, and am impressed at how close you were to what I am paying. However, there is no option for telling you I was involved in an accident in which I was not found at-fault.

Anyways, nice work.

Great! We do a lot of testing, but it's always nice to hear that the prices we gave were accurate!

(Insurers shouldn't care if you had an accident and weren't at fault - their models say that should not affect your price, which is why we only ask for 'at-fault accidents')

Ahh.. I see, thanks for that info. This means I can bargain harder :)

How do the insurance companies that offer price comparisons do it?

They don't actually care to get the competitors' prices right. They find a "worst case" example somewhere and go with that, with tiny fine print saying something like "results not typical" or "results may vary based on location".

For example, Geico says "you could save up to $500". I don't even pay $500, but as long as they can show there was a single case where they saved someone $500 they're golden.

This

There are services out there that provide the info. I once used these guys: http://quadinfo.com/Pages/Default.aspx.

Anecdotal evidence is anecdotal and I'm a major skeptic of all things, but I'm amazed at hour accurate Leaky was for me.

My wife and I pay $119 per month for car insurance and Leaky's magic was within $3 bucks. Wow.

EDIT: Might I suggested some kind of accuracy checking feature? For example, my Leaky estimate vs. actual was really close and it would be cool if you collected data on this. Like if I selected "I use company X and my rate is $Y", then you could compare accuracy.

Thanks! We do a lot of internal testing, but I'm always really excited when I hear that we were close for an actual person!

Thanks for the feedback too, it sounds like something that we're working on to allow users to fold their information into our routine accuracy testing - which should be helpful for them and us.

Were you made aware of any legislation or caselaw that would illustrate that scraping was illegal or otherwise actually actionable?

It seemed like a moot point. Either way, it was unsustainable in the long-run and we didn't have the bankroll to battle it out in court.

Doesn't matter.

The insurance companies would be entirely within their rights to block the scraper, and they'd probably be happy to sue and drag it out for five years. It wouldn't take that long for Leaky to run out of funds (and the distraction would likely be fatal even if the legal expenses weren't).

I was just asking for information. It might not matter to them, but it might help others.

My point is that it doesn't matter to anyone. It might be entirely legal, but they'll be happy to bankrupt you while you attempt to prove it.