When it comes to auctions on the web, users rarely have any kind of control beyond setting the maximum price they’re willing to pay. If, for example, you wanted to buy a power supply on eBay, but only if you won a separate auction for the computer it would work with, you’d have to manually check in on each auction – you couldn’t conditionally say that you only wanted one if you could have the other.
Auctionomics, a new startup launching tonight, is looking to add another layer of control to auctions on the web. The company has built a powerful backend SaaS solution that allows websites to conduct complex auctions, where the criteria involved are more complicated than simply determining the highest bidder for an item. It’s a bit confusing, but fortunately end users won’t have to deal with the intricacies of how the auctions run – they’ll just find that they have more of a choice when they’re bidding on items on the web.
Imagine trying to choose which hotel you’d like to stay in for an upcoming trip. Instead of having to simply sort by price, the engine would allow the user to enter more complex requests using conditional statements, submitting multiple bids at the same time. For example, if I was heading to Lake Tahoe, I could say I would be willing to pay $200 a night to stay at Harrah’s, but only if I could stay for three nights in a row at that price. In the same query, I could say I’d be willing to pay $150 a night to stay Caesars, but only if I could get that price for a whole week. And so on.
The service is powered by an engine called MaaX (Milgrom Assignment Auction), which was developed by Stanford professor Paul Milgrom, who designed the original auction system used for the FCC spectrum and has been involved with auction design and strategy for major companies and the government for decades. The system is built to maximize the number of bids from each participant, as well as gains for all parties involved. It does this in part by allowing users to submit queries with multiple variables, like setting a set spending limit across an entire multi-part bid, and allowing for substitutions between various items.
This all sounds pretty complicated (and it is), but co-founder Dr. Silvia Console-Battilana says the services that eventually implement the technology will be able to simplify the choices that users are exposed to. For example, a travel site could just show a few options at a time in a multi-step wizard, or it could simply reduce the number of available variables.
Beyond consumer sites, Auctionomics will appeal to a variety of other markets, like the energy industry. Whereas previous energy bids from energy companies would have to be placed separately, using MaaX energy companies could conditionally bid on energy from multiple sources (I could say I wanted up to 100 Watts from Northern California at $X and up to 200 Watts from Southern California for $Y, but that I did not want more than 150 Watts total).
While Auctionomics is not currently powering any public sites, the company is in discussions with a number of possible partners to implement its technology (though it sounds like it may be a while before any of them go live). This added control certainly sounds promising, I just hope that sites can find a way to give it to us without becoming overwhelming, otherwise users will shy away from it.
At the end of the day, Auctionomics is just another eBay’s competitor.
What? Because it has auction in the name?
Electric companies do not bid on electricity sources using ebay.
Oh wait this is spam isn’t it? (I’m actually not joking, I just realized… I’m a turkey)
you say that like theres something wrong with that.. and clayton is correct
One to watch, sounds interesting and looks like a good team behind it.
Cheers
Sahar Sarid
Co founder, Bido.com , Social Auction Platform
Very cool. It’s always good to have new tech like this coming out.
I’d imagine the filter/search tech behind it is pretty amazing. Even simple search filters can really tax servers if not done carefully.
Sweet
The following research papers might be useful for Auctionomics developers, which they both use Neuro-Fuzzy algorithms (a hybrids of ANN :- artificial neural network and FIS :- fuzzy-logic inference system ).
1) Abstract:
———
This paper presents the design, implementation and evaluation of a novel bidding strategy for obtaining goods in multiple overlapping English auctions. The strategy uses fuzzy sets to express trade-offs between multi-attribute goods and exploits neuro-fuzzy techniques to predict the expected closing prices of the auctions and to adapt the agent’s bidding strategy to reflect the type of environment in which it is situated. We show, through empirical evaluation against a number of methods proposed in the multiple auction literature, that our strategy performs effectively and robustly in a wide range of scenarios.
PDF Download:
————
An Adaptive Bidding Agent for Multiple English Auctions: A Neuro-Fuzzy Approach
2) Abstract:
———
Markov Chain Model provides a concise mathematical model to describe the online English auction process, converting the complicated interaction between the bidders and
auctioneer into a tractable mathematical problem,
which is a milestone for researches involved in
this area. However, the assumptions about the
parameters are not consistent with the actual
phenomena, for example, the distribution of the
private values and the arrival rates. Furthermore it is hard to obtain the values of these parameters. In this paper, a hybrid method, neuro fuzzy, is proposed to predict the final price in addition to exploring the complicated, possibly nonlinear, relationship among the auction mechanisms and final price. The empirical results show that neuro fuzzy system can predict the final price accurately
much better than the others, which is of great
help for the buyers to avoid overpricing and for
the sellers to facilitate the auction. Besides, the
knowledge base obtained from neuro fuzzy
provides the elaborative relationship among the
variables, which can be further tested for theory
building.
PDF Download:
————
A Final Price Prediction Model for online English Auctions—A Neuro Fuzzy Approach
I have no experience in the development of auction system, however I have developed & used neuro-fuzzy computing techniques for applications in control systems over the past several years.
There are many neuro-fuzzy algorithms available today and the one I have been using is called ANFIS (adaptive neuro-fuzzy inference system), which is very popular today especially in feed-back control system design and electronics. There are others such as NefCon, POPFNN , TWNFI and many.
I read an article in the IEEE Transaction on Control some years ago, that the module for automatic docking of the Space Shuttle aircraft uses ANFIS. When the Shuttle approaches the space station, its ANFIS guided system is switched on, which guides the Shuttle to the docking port of the space station. The ANFIS guided system is more reliable, since it can guide the aircraft to acceptable tolerance of mis-alignment within a few centimeters, and any deviation of misalignment much more than the acceptable tolerance and doesn’t lock properly, it may potentially lead to disaster.
Anyway I can point you out to more info (appropriate literatures) on neuro-fuzzy if you wish.
You mention how complicated it is twice, but that the end users don’t need to understand it in order to use it. But we’re here to understand it!
Thanks to Falafulu Fisi for not assuming we’re stupid.
Someone should be contacting the lady below.
Patent # 7,272,575 Method and system for facilitating service transactions
Patent App # 2008,0021,794 Method and system for facilitating service transactions
Patent App # 2002,0069,079 Method and system for facilitating service transactions
Patent App # 2002,0120,554 Auction, imagery and retaining engine systems for services and service providers
Interesting article.
nice
Thanks to everyone for the comments. ‘Falafulu’ and ‘Clayton’, since you seem to be quite technical and into the theory part, here is my co-founder’s paper that explains the source of the concept:
http://www.milg...20Exchanges.pdf
Users however would just be faced with very simple additional choices that current auctions cannot support, at least not while finding the best possible allocation. Users would get a better deal, sellers would receive more bids, and who ever runs the site would get more profits from higher traffic and happier returning costumers.
Windowslogy, it is true that we could compete with ebay in some spaces, however our team members have run/advised very complicated auctions such as radio spectrum auctions and energy auctions. Ebay could not even go close to that level of sophistication (and they have no interest in doing so).
Silvia,
Thanks for pointer to the paper, and its interesting. You’re using optimization as shown in your paper.
Cplex from ILOG (recently acquired by IBM) is the most advanced and popular commercial industrial-strength optimization API out there. I had used it when I was at University years ago. I believe that it has advanced with more functionalities since then. Integer programming wasn’t available then, but I believe that it has now been included. I both use LP (linear programming) & QP (quadratic programming) for some of my developments in financial asset allocations (mean-variance/Markowitz-type).
I am interested in Integer Programming for further application in financial asset analysis/modeling, but I guess that I will get to it at some stage. My interest was prompted when I came across the following paper. It has direct application to MBS (mortgage-backed-securities), which I have developed some evaluation algorithm for MBS (just the standard one).
A two-stage stochastic integer programming approach