Swoopo

After having Google Ad-sense advertise this site to me approximately 100 times, I finally visited swoopo.com the other day. The site advertises that it sells electronics for absurdly cheap prices, such as 20 dollars for an ipod touch or 200 dollars for a new macbook, so I was obviously curious if this was true and how could they afford to undercharge for such items. The first thing that struck me was that the site was an auction site, so there is no notion of directly purchasing something. After I looked around a while more, I discovered the trick for why they can afford to charge very low prices.

Here is the rundown of a standard swoopo auction: The item, lets say a PS3, starts off at some very low price, say 30 dollars. Then every time a user submits a bid, the price is increased by a predetermined amount, usually 15 cents and the time until the auction closes increases by a certain amount, usually 15 seconds (at some points, the time left jumped up dramatically, like by 20 minutes, but I couldn’t figure out the pattern there). At some point, someone will have placed an unchallenged bid and the auction will end. This part is relatively simple, just the standard ascending bid English auction (that is some moderator will name the next price and bidders will agree to pay that price, as opposed to having bidders name their own prices), but this doesn’t explain how this site makes any money.

The real trick is that every time you bid, you must spend one of your “bid tickets” and you have to buy bid tickets from the site. In other words, they charge you $1 for the right to submit a bid. So while one person might get an item that is way below market price, there are many more people who have essentially spent money and have nothing to show for it. So if the site is selling an item worth $1000 dollars and the bid price starts at $0, they must get 1000 bids on the item to make it a profitable venture for them (let’s ignore overhead costs right now). Those 1000 bids will only drive the price up to $150 which still makes it an amazing price to get that product at, so it is pretty likely that people will continue to bid on the item until it hits a point where it is reaching some reasonable approximation of market value. Even though Swoopo probably lost money on selling the laptop to the winner, they have made money off of all the users who submitted bids and lost.

Although this format strikes me as kind of gimmicky, the site is very upfront about the process of buying bid tickets and that you will lose a ticket every time you place a bid, regardless of whether or not it was the winning bid. Also, it seems that representatives from the site have been very good about responding to interviews and being honest about what they do. In one of the interviews that I read with a PR rep from Swoopo, he said something like “we don’t want people to feel like it is impossible, we want people to win” which sounds very reminiscent of rhetoric from casinos or from “Thank You for Smoking.”

One thing that I do find to be questionable about the site is this automatic bidding bot that it provides free of charge. Essentially, you tell the software what item you would like to bid on, how much you are willing to pay, and how many bids you are willing to use towards that item and then it will automatically submit bids very close to the end of the auction on your behalf. After watching an auction on a PS3 that started off at 37 dollars (or rather that was when I started watching the auction), it became very clear that many people were using this bidding robot since with about 5 seconds left in the auction, four or five bids would go in automatically (it would also say if the bid was placed by their bid butler as it is called). That PS3 ended up selling for somewhere in the 200 range, so I cant even imagine how many people lost a ton of their bidding tickets because of the straightforward strategy used by this bot.

Looking at this from an analytic point of view, we can see that we only really need one user to “bite the bullet” and submit the bid that will prevent the current bidder from winning and prolong the auction, so there is no need to submit a bid if you are sure that someone else is going to submit one. In fact, not submitting the bid there is to your advantage since you will have more bids later in the auction if we assume that everyone has finite resources. One could probably do some statistical analysis to estimate the probability that someone will submit a bid given what percentage of market value the current price is. If we had access to that, we could layback and wait for all the “noise traders” (people who use no intelligent strategy) to waste their bids when it is extremely improbable that anyone will win the auction, and then enter the auction later on when its more likely the auction will end within a reasonable number of bids. For instance, if we knew that product X has a 90% chance of getting a bid at price Y, then we would submit a bid at that price only 10% of the time (determined randomly). Following such a strategy would certainly not guarantee a win, but it would definitely result in using less bid tickets (in the above case, assuming that the 90% chance held for a while, we could save 90% of our bids) than using their bid butler. Obviously, Swoopo has no incentive to improve their software since their current bid butler is probably making them tons by wasting users’ bids, but it might provide a market for a third party software company to make a bidding bot that saves users lots of bids over times. Of course, if everyone used such a process, it would come down to the random numbers generated by the software, making it more of an explicit gambling game, but would save everyone money. I would also guess that it would decrease Swoopo’s profit by a decent amount since there wouldn’t be as many traders submitting simultaneous bids and thus keep the price a little bit lower.

I highly doubt that I will be making any bids on Swoopo in the future, but it does offer an interesting application of auction theory (according to my advisor, this “pay per bid” auction has not been studied too much). And if I am ever given a bunch of money and told to start a company, I might pursue that more intelligent bidding software option as a product (would it be better to charge a subscription fee or a certain percentage of the price of an item that a client of ours has won?).