Adverse Selection! What Does "Unlimited" Mean?

Bandwagon of http://ridethebandwagon.com/ estimated that the average iTunes user had some 30 to 60GB of stuff:

What does unlimited mean? - bandwagongroupie | Google Groups: Terence Pua: Based on survey data (months before launch), our research showed that most iTunes users have about 30 to 60GB...

Bandwagon of http://ridethebandwagon.com/ thought that if their average customer had some 39 to 60GB of stuff to back up, they could make money selling dynamic backup services for $99 a year.

But if you have 5GB of stuff, will you pay $100 a year to back it up when $100 will buy you two Western Digital 80GB hard drives? No.

What kind of people will spend $99 a year for dynamic backup services? Terence Pua found out:

However, in a little more than 24 hours after we opened the doors, unlimited means people have as much as 2TB of iTunes stuff! That is an outlier for sure but many people have over 100GB. As you know, we are currently using Amazon S3*, who charges $0.15/GB/month. We charge $99/year for Bandwagon. At that price, people who have over 55GB become a loss to us (and that's not including the bandwidth costs at $0.20/GB).

It's simple arithmetic that if we don't cover costs, we can't stay in business to serve your needs.... We'd like to have a fruitful community-based discussion to help us come up with a more long term pricing/model...

But they are getting some nice publicity out of this.

Comments

[Gex]

Hmmm... Users with 30 to 60 GB of iTunes probably have it on a hard drive that is AT LEAST 100 GB, if not much, much more. And that hard drive probably contains all sorts of other data. I hope they didn't make the mistake of assuming all people would back up is their music. I mean, I didn't do a detailed study for a business plan, but it seems obvious that the numbers the present represents only a subset of what a customer would back up.

[yoyo]

I can't imagine backing up several hundred gigs of data, besides you can get 500 gigs for $150 or so...

[derek]

Forgive me for economic naivety, but until now I had a different understanding of what "adverse selection" meant, because I kept hearing the phrase in the context of what a problem it was for universal health coverage funded by federal taxes.

Now you've explained it in terms of offering the customers a service that only a part of the customer base (the part that will draw most heavily on your service) will buy, it seems clear to me that all the talk about how adverse selection is a problem for universal health care is backwards. By being funded from taxes, universal health automatically gets 100% buy-in. It's _private_ health care that has a big issue with adverse selection.

One way of dealing with this is to restrict the service, offering more to those who pay more, and denying more to those who pay less. Again, I have often heard how universal health care will suffer from "rationing", but again, this model suggests the opposite: it's _private_ health care that results in rationed treatment, and public health care that can most easily offer 100% service to 100% of the customer base.

[Jon Gallagher]

What we actually have is the consequences of low barriers to entry.

Remember when everyone and their brother was running a search site? It was easy, run a bunch of crawlers, stick them in a database and use regular expressions to suck out the answer.

Then Google comes along and actually applies solid engineering, mathematics and hard work to the problem. What's the most powerful part of Google, the foundation that they build every new product (and their coming world domination on)? Storage. An enormous operating system linking inexpensive PC's in a global hive-mind large enough to contain the entire Internet in memory.

The Bandwagon is nothing more than those early unlamented search sites. Like most consultants they are trying to stay one page in the manual of technology ahead of their customer, and arbitraging the their difference between their cost of storage from Amazon.

Here's a simple idea guys that you could easily test. How much of that TB per customer is repeated among other customers? For example, how many copies of Dark Side of the Moon are you storing? Why not store the first one, track what you have, then as the next one is uploaded merely create a tick mark in the user's account entry? At the very least it would seriously speed up the upload process.

Use some engineering skills and test some assumptions, like whether the 80/20 rule applies to common music ownership over a sufficiently large population? Hell, I would be willing to bet ($10 paypal to your favorite charity) that the "sufficiently large population" is on the order of 1,000 members (sorry don't have time to scope out whether that plugs into a statistical model. Dr. DeLong, some help here?) So instead of 1TB you are more likely looking at 200Gb for you power users and < 100Gb for everyone else.

I mean come on people, do you think GMail actually stores every byte of your email? A simple Huffman compression will yield at least a 10x compression on the normal (pathetically small) email vocabulary, and if you cheat and use tokens for common phrases you can even beat that. Which means your 2.7GB for Gmail is residing on ~270MB of disk that Google paid $.20/GB (or less) for. Sure they take advantage of their cost of storage. But the reason they pay such insanely low prices for storage is that they engineered an infrastructure that they just plug more disks into with little marginal cost to them for upgrading their storage capacity. Try adding a disk to your windows box in less than half an hour, then figure out your marginal cost of storage.

[walkingtheline]

The really interesting here is the use of the S3 service to roll out this sort of massive and scalable storage service. This is a completely disruptive technology, since Amazon is competing with budget server companies price-wise, but coming at the market with instant scalability.

I'd be very surprised if the company offering this service has more than one or two (competent) coders on team. Their best solution is probably just making their own cost structure transparent to their users, and charging an extra penny per gigabyte or something. They could always just offer Amazon's price point to heavy, heavy, heavy users.

On the other hand -- who did they expect would take them up on their offer?

[Ken Houghton]

Jon Gallagher - External 500GB HD spotted at Costco yesterday for $219, complete with "USB" for hookup.

I'll also dissent on this being "adverse selection." If you assumed your mean (not median) iTunes user has 30-60G, and given that you know that the 80GB iPod was released only within the past six months, and your b/e is 55GB, your pricing model s*cks.

Good technology cannot overcome bad management

[Jon Gallagher]

Thanks Ken about the 500 GB, I got an email from one of my suppliers offering 250 for $80. Wheras I am a nobody with a credit card, as opposed to a Google with billions, what do you think they pay?

I am in complete agreement with your last line. I have a corollary, "Never computerize a mess. It just gets messier faster".

So why exactly does Bandwagon exist? If they are true music lovers, how about helping me with the vast world of recorded music out there? What do they like? What do they recommend? I spent the end of last year obsessively copying down people's "Best Of" lists, and Thursday/Friday Random 10 to try and get some idea of what used to be at the tip of my fingers in college (pre-Internet even), what's going on in music I care (or *could* care) about?

They're just another pretty website with zilch behind it.