Shill Bidding on eBay: Case Study #2

Shining some light on the more sophisticated and therefore harder to detect shill bidding activity by some “professional” sellers on eBay auctions

Oh no, not another case study on shill bidding on eBay auctions? Yes, sorry, another one. This time a spreadsheet analysis of multiple auctions, from some "professional" sellers from the US and Australia. Needless to say the analysis demonstrates, once again, that, contrary to eBay's claims, shill bidding by many “professional” sellers is rampant on eBay auctions. The full comment and spreadsheet download links at:

http://www.auctionbytes.com/forum/phpBB/viewtopic.php?t=24296

      “eBay does more than any other Internet company to make sure that online consumers have a safe and well-lit place to buy and sell goods. We have the most proactive and comprehensive policies and tools in place to combat fraud and illegal activity. eBay has zero tolerance for those who abuse our platform and we work to ensure criminals are arrested and prosecuted.”

This statement, made by eBay’s Department of Spin, on 26 August 2009, is, of course, just that: pure, unadulterated, “spin”. And if you believe this statement, you may as well believe that there are fairies at the bottom of the garden, too. But, seriously, this is the reason why eBay has made so many poor decisions in recent years, they are so busy spinning fairy stories for their consumers and for the financial market and its dopey analysts, in an effort to maintain the share price, that they are continually too giddy to make a logical decision about anything else.

Or, to paraphrase what a correspondent recently told me, when you attempt to resell at auction on eBay, desirable collectibles that you had some years earlier bought at auction on eBay, and you cannot achieve half what you originally paid for them, you know there is something amiss on the “Road to San Jose”.

And so, further to my earlier detailed case study of naive shill bidding on eBay, posted at
http://www.auctionbytes.com/forum/phpBB/viewtopic.php?t=24033
I offer the following study of the more sophisticated and therefore slightly harder to detect “professional” shill bidders who are now running even more rampant on eBay; such activity made harder to detect and even harder to keep track of over time by eBay’s masking of bidders’ unique IDs; as a consequence the bidding process is even less transparent, and buying at an eBay auction is much more problematic.

Synopsis

 very little of the auction system security, that eBay claims to offer buyers, exists in fact;

 eBay has no effective matter-of-course verification of users: unscrupulous users can apparently have as many user IDs as they may have email addresses, and clearly there is no effective matter-of-course cross checking of such IDs;

 by the lack of any effectual system to proactively detect shill bidding, eBay has always effectively—and knowingly—“aided and abetted” unscrupulous shill-bidding sellers to defraud naïve (and some not so naïve) buyers, and by so doing, eBay benefits from the resulting higher “final valuation fee”;

 the masking of bidding IDs with non-unique, absolutely anonymous, periodically changing aliases serves no purpose other than to further obscure all but the most naive of shill bidding, making it difficult for even experienced buyers to protect themselves and track suspicious bidding activity over time, and defeats utterly any attempt at programmatic analysis of bidding patterns to expose such criminal activity;

 In the normal course of events, only those user-reported sellers who, by their utterly naïve and blatant shill bidding activities, so embarrassing expose eBay’s lack of any effective control of this criminal activity, will incur a nominal penalty from eBay, a penalty so ineffective that it is simply a charade;

 eBay has no interest in doing anything truly effective about shill bidding for to do so may well decimate the “professional” sellers that use the auction process on the eBay marketplace and on whom eBay relies for much of its revenue. To actually do something truly effective about such criminal activity would undoubtedly detrimentally affect eBay’s bottom line, and we could not have that now, could we?

 the linked spreadsheet analysis of some eBay auction data, once again, demonstrates that eBay—contrary to their claims—has no effective “proactive” nor “sophisticated” systems for the detection of undisclosed vendor (“shill”) bidding and that such criminal activity is rampant on eBay.

The detail of the matter

eBay is not a local auction, frequented by those who understand the dangers of the auction process, and where the auctioneer may well take bids from insects climbing up the wall; eBay is ubiquitous, remote and every detail is digitally recorded; eBay therefore has the ability to do much more with that data to better detect and control shill bidding. With respect to sales, on the eBay marketplace, by the auction process, eBay is “the auctioneer” (who else can be?) and in most jurisdictions has a statutory obligation to not facilitate shill bidding. My contention is that they do not do anything truly effective because it is not in their financial interests to do so, and by their lack of any effective proactive action they, in effect and knowingly, aid and abet unscrupulous sellers to defraud buyers.

Since eBay has masked all bidding IDs with non-unique aliases (comprised of two characters: about 1600 possible combinations), it is no longer possible for genuine bidders to be certain about anything. And that is, undoubtedly, the principal purpose for such masking.

Some people naively believe eBay’s fairy story that this masking of the unique bidder IDs was introduced to “protect” bidders from scam emails; this is nonsense; that “problem,” if it ever was a real problem, was solved by the blocking of access to bidders’ direct email addresses.

Some others actually believe that such masking was introduced to protect users who had apparently used their email address “name” as their eBay user ID and were receiving scam emails on that basis: emails were supposedly being spammed to eBay IDs with alternative ISP provider names of every common ISP provider (ie, eBayID@aol/google/hotmail/etc). I consider this reason, if it was ever put by eBay, to be simply too absurd to contemplate.

And to any other eBay apologists I would simply say that if eBay had—as they claim to have—any truly “proactive” and “sophisticated” systems in place for the detection of shill bidding, there would not be any shill bidding for users to detect, and my abovementioned earlier case study of naïve and blatant shill bidding clearly indicates that they indeed have no such effective system. The only shill bidding eBay does anything about is that which is reported by users, and for the confirmation thereof or otherwise by eBay, I have no doubt eBay relies on some very unsophisticated system of data matching. That which goes undetected by users …

It should be clear to any thinking person that the non-unique masking of bidder IDs (particularly that form of masking suffered by users in the UK) never had anything to do with the “security” of bidders—indeed, just the very opposite is the case—it has only ever been about the obscuring of all but the most naïve of shill bidding so that this criminal activity would be less of a problem—for eBay, that is.

Such non-unique masking also stops third-party providers, such as “GoofBay”, from applying truly sophisticated algorithms to eBay’s bidding data to expose the type of sophisticated shill bidding that would otherwise be difficult for genuine bidders to manually detect—and certainly not possible to detect on an individual-auction basis from eBay’s Bid History pages. One can only assume that the regular exposure, by third parties, of such criminal activity never suited eBay.

The Bid History Details pages, provided supposedly as a counterbalance to the masking of bidders’ IDs, contains a “30-Day Summary”. Why is this summary only 30 days and not 90 days, or infinite? I would submit that eBay’s “nod, nod, wink, wink”, attitude to shill bidding requires that “professional” shill bidders be able to “refresh” their shill IDs by rotating them over as short a period as possible. If the summary was over 90 days the shills would need 60 shill IDs instead of 20, if the summary was infinite, heaven forbid! And that is were an ongoing spreadsheet analysis comes in handy.

eBay continues to, disingenuously, claim that shill bidding is a minor problem that they have under control. I would suggest that, following the introduction of “hidden bidders”, shill bidding has blossomed and is now running rampant on eBay auctions, and the fact that I have previously demonstrated that eBay has no truly effective system to control this criminal activity is an indictment of their total lack of principles in providing for their “buying” users.

In anticipation of the “spin” that the eBafia’s “made men” will undoubtedly again put on this matter, I will again reiterate that the examples of obvious and probable shill bidding in the linked spreadsheet are not simply a (further) “few examples” from the “100 million” eBay listings, but are from only some “hundreds” of listings mostly from my areas of personal interest.

Unlike eBay, I don’t have the capacity to check every auction, I only look at those auctions for items that interest me or for items from those sellers (usually “professional” traders) that are regularly listing items for sale, some of whom I have suspected of shill bidding. Not all the sellers listed in the spreadsheet have any appearance of shill bidding and a number of apparently scrupulous sellers are left in the list for comparison purposes.

Again, unlike eBay, I have to work with the non-unique aliases that are not only dynamic in that the accompanying feedback values, that supposedly serve to differentiate bidders at a moment in time, may increment over time. Also, the two-letter alias is apparently also changed on a quarterly basis (what possible purpose could that serve other than to make such analysis problematic—which, undoubtedly, is eBay’s intention).

As the masked bidding aliases are not unique it is also now possible to have more than one bidder on the same auction with the same alias—as is demonstrated in this spreadsheet list. Even though eBay appends the bidder’s feedback count, which may help to differentiate like aliases at a given moment in time, this can never be definitive; however unlikely, it is possible to have multiple bidders with the same alias and the same or similar feedback count; also, in the normal course of events, feedback counts may increment, making it difficult to even manually analyse such data over any time, and certainly impossible to programmatically do any such analysis.

Then, auction data remains available on line for only 90 days and there is no way to identify an underbidder uniquely to see if that bidder has a longer-term relationship with the seller that may suggest shill activity (shills that win auctions rarely leave or receive feedback, but that in itself is an indicator that it is a shill) and, of course, the quarterly changing of even these masked aliases effectively stops that analysis too.

It should, however, always be borne in mind that where a particular bidder (or bidders) does appear on more than one auction from the same seller, it is possible (to a varying degree, depending on the category of goods, eg, unique collectibles) that it could be a genuine bidder (or bidders) interested in more than one item from that one seller. Conversely, it could be some other seller’s shill merely diluting the proportion of his own-auction bids—with never any intention to buy—so that his own-auction activity does not become obvious on the bidder’s Bid History Details page: these could be some of those bidders in the list who habitually place early “no win” value bids on many nominal-start items from various sellers.

You can imagine how much more positive any such analysis could be if we still had the unique (anonymous) bidders’ IDs to work with: then we would be able to further investigate a bidder’s bona fides. Clearly, eBay has decided that that is not in eBay’s interests.

Of course, eBay has access to those unique IDs, and eBay could do such a much more effective analysis. After all, such a simple analysis is not that difficult, even I can do it with a spreadsheet, and some effort.

Why then won’t eBay do such an analysis? Undoubtedly they have already done such an analysis, and because such analysis will have shown just how much suspected shill bidding is going on, they prefer to not talk about it, and they won’t do anything effective about the problem lest their “bottom line” be detrimentally affected. The non-unique masking of the already anonymous bidding aliases to further obscure such activity was a much simpler and less revenue-threatening, albeit disingenuous, way of dealing with the matter.

Then there is the use of the seller-selected “private” auction format. What I call eBay’s “classic shill bidders’ tool”. This auction format has always made it absolutely impossible for genuine bidders to protect themselves from such criminal activity. Notwithstanding that such “private” auctions will invariably have a greater number of bids thereon than a standard auction may have (I wonder why?), there is simply no way of analysing the bidding on such auctions. A few are listed in the spreadsheet because the usually higher number of bids on such auctions is a telling phenomenon and the reader can come to his own conclusion as to why that would be so. If it is because a shill bidder is active, only eBay can know, because they do have access to the underlying bidder IDs—but they aren’t going to tell us.

The fact is, if you bid other than at the latest practicable moment (preferably only on the starting price, if you think that is a fair price) on such a “private” auction, on the balance of probability, you are going to be “taken to the cleaners”—and eBay knows that too. Only a fool could possibly think, or a most disingenuous person suggest, that shill bidders have not always made much more use of this classic shill bidders’ device than ever did any seller concerned about the “privacy” of his buyer—and eBay knows that too.

Regardless, even without the aid of a truly sophisticated shill-detection software algorithm, it still comes down to the “pattern” of bidding by the suspected shill bidder with the seller: how many auctions, what types of items, the proportion of auctions therefrom, etc.

But, bear in mind that, alone, a high percentage of “Bid activity with this seller” may or may not be an indication of shill bidding: 100% of bid activity on only one auction with a seller, may not be significant; 100% of bids on a number of auctions with one seller only is obviously very significant. Certainly, as the number of auctions of the seller that a bidder is bidding on increases—regardless of the number of bids on each auction—the probability that it is shill bidding increases exponentially.

However, the more sophisticated “professional” shill will have multiple shill bidding IDs (as is demonstrated in this spreadsheet), and will have prepared same by continually placing bids on many other low-start auctions from multiple other sellers (and possibly even “buying” feedback from other like criminals), so that the percentage of “bids” with the particular seller will never appear abnormal, indeed may be thereby kept so low as to appear beyond any suspicion, and some examples of such possible “own-auction diluting” activity can be observed in this spreadsheet.

So, when looking at a competing bidder’s Bid History Details page—particularly that of a nibble-bidding competitor—you should also watch for any great deal of losing bid activity on multiple other sellers’ auctions: this type of activity may be an indicator that the bidder is thereby diluting the record of his bidding on his own auctions. Of course, eBay’s masking of bidders’ IDs means that such analysis can now only be done with an extensive collection of data as is compiled in this spreadsheet.

Let’s face it, genuine bidders either (naively) “nibble” bid or they eventually wise up and “snipe” a strategic single bid near the end. The appearance of any bidding patterns in between these two extremes, particularly any nibble bidding by a high feedback ID, should be looked upon as a potential “red flag.” Also, as demonstrated in the abovementioned linked earlier AuctionBytes post, eBay’s “proxy” bidding system is so open to abuse by unscrupulous sellers that to set a maximum proxy bid early on in an auction can be an invitation to indeed pay the maximum that you have indicated you are prepared to pay.

Patterns of bidding worthy of further analysis are not hard to find; simply search any auction category, skip over all those very, very, many items with zero bids until you come to one (or a group) with an out-of-the-ordinary number of bids thereon, then check that seller’s other items; you may find that that seller has abnormal activity on all, or many, of his items, and the clincher, of course, will be his use of the “private” auction format (can you guess why?).

As this spreadsheet analysis only examines the information on the primary Bid History page (not the Bid History Details page) and looks for the appearance of common bidders on a particular seller’s auctions, most of the suspected shill bidders in this collection will fall into this “less obvious” category, and will not stand out even on the Bid History Details pages. It therefore is never sufficient to examine the bid history details for only the auction in which you are interested; you must also examine, at least, the primary Bid History pages of some other auctions from the same seller for the appearance of any common bidders that may indicate that shill bidders are active, such as the examples exposed in this spreadsheet analysis.

The spreadsheet summarises the data from the primary Bid History page. This data can be sorted by a number of pre-defined sorts designed to expose bidding patterns that indicate levels of likelihood of shill bidding. Some of those “levels of likelihood” in this list go well beyond simple “likelihood”. And eBay would be aware of that too.

The bidders’ feedback counts are those extant on the noted date that this “dynamic” data was collected and may increment over time. The, always superfluous, three asterisks included in the masked bidding aliases have been removed and those aliases that appear more than once on a seller’s auctions have been flagged with an asterisk.

(You have to ask yourself, what sort of child-like logic was used when eBay made the decision to include three superfluous asterisks in every alias? And, let’s not bother to question why eBay did not choose to create a unique masking …)

The somewhat consistent “rolling” nature of the appearance of these flagged bidding IDs suggests that many of these sellers went to the same “How to Shill Bid” seminar or, possibly, they are subscribing to a shill bidding service of some sort.

You may also notice, if you sort the spreadsheet by “Bidders” only, that even when one of these flagged aliases appears on another seller’s auction (sans flag), they are almost never the winning bidder and may well be one of those bidders making very early nominal-type bids on others sellers’ actions for the purpose of diluting the proportion of bidding on their own auctions. Each bidder’s highest bid only is recorded to help make such activity more noticeable.

A very low feedback count is another potential “red flag”: every new user starts with zero feedback, but there seems to be a plethora of such “newbie” bidders in recent times. But do they actually buy anything? If every shill bidding seller now has about twenty bidding IDs, and they all start with zero feedback ... (Maybe this is from whence eBay is getting its recently claimed increase in users.)

Another potential “red flag” is a seller’s listings switching from auctions to fixed-price “Buy It Now” type listings; this could be part of the “nominal” penalty imposed by eBay when a seller is reported for, and is found to have been, (blatantly) shill bidding. Unscrupulous sellers don’t like “fixed price” listings because, if you can’t get what you want for an item at auction with a nominal start, without shill bidding it, it’s not going to sell for the price you want at a fixed price.

At the extreme, a number of the listed sellers actually have more shill IDs bidding (and sometimes winning) than genuine bidders, for example, the seller “art.warren” has multiple shill IDs outnumbering genuine bidders on his auctions by about 10:1. One of his shills more often than not wins his auctions. Whether eBay gets any FVF out of such “sales” would be a good question; certainly this vendor appears to be selling very little to genuine bidders. Then not many, even scrupulous, sellers in this category (“Paintings”) appear to be selling very much these days.

eBay is in the process of introducing a new “View Item” page. You may notice that the “High bidder” is no longer recorded on this initial page: you now have to look to the Bid History page for any information on bidders, which makes it just that bit harder for genuine bidders to spot a known common bidder placing an early (surreptitious “reserve”) bid on a seller’s auctions. Now, why would eBay do that? And, the “starting details”: price and date have also disappeared. Now, why would they do that? (Shill bidding is much more likely to occur on items with a very low starting price relative to their market value.) Presuming that these people at eBay are not all simply stupid people, the only reason for these further changes to the View Item page can be to further obscure matters. And I think we all understand why anyone ever wants to make any matter less transparent, don’t we?

But, if you are one of those buyers that are still dabbling on eBay, you don’t need to worry, as eBay’s “proactive” and “sophisticated” shill detection systems will root out these fraudsters before they get a chance to cheat you ...

Needless to say, the above comments don’t apply to the UK site where the style of masked alias (“Bidder x”) makes it simply impossible for genuine bidders to do any such spreadsheet analysis; users of the UK site are now able to protect themselves only from the most blatantly obvious and naïve of shill bidders. The sophisticated shills (such as those recorded in this spreadsheet) now have free reign in the UK with absolutely no risk of exposure: they now don’t even need to resort to the “private” auction format to defraud buyers—and eBay knows that, too.

This spreadsheet analysis—once again—demonstrates that, contrary to their claims, eBay does not have any effective system for the detection and control of the criminal activity of shill bidding, and that such activity is now running rampant on their auction system; nor does eBay have any reason to do anything about it while consumer affairs regulators, apparently, see no problem with eBay, in effect, aiding and abetting such criminal activity.

With the exception of a couple of the sellers who were reported to me (and who have previously also been reported to eBay), the rest are a random selection of sellers who habitually use a low starting price relative to the value of the goods. Not all the sellers so analyzed over a number of auctions have any appearance of shill bidding but those are left in the list for comparison purposes. The ratio of scrupulous to apparently unscrupulous sellers in this group is simply frightening—and eBay would be aware of that too.

Does eBay—as they claim—actually do anything, in any “proactive” and/or truly “sophisticated” manner, to control shill bidding? The answer has to be an emphatic “No, beyond any doubt, they do not”.

Does eBay even care about such fraud? “No, they do not”: undetected shill bidding simply increases eBay’s revenue from the resulting higher final valuation fee.

eBay is its own worst enemy. Doing nothing but attempting to further obscure shill bidding has sapped the confidence of many experienced buyers: it’s simply too much effort to have to be constantly on the look out for the unscrupulous “professional” sellers that infest eBay auctions.

I can well understand why eBay might prefer all listings to be “fixed price”; but eBay would not then be the same and, if and when the last auction is run, someone will have to remember to turn off the lights.

It would appear to me that the eBay marketplace is little more than a “house of cards”; if so many of the “professional” dealers trading thereon cannot make a living except by bidding on their own auctions, then that suggests that their removal from the system would greatly reduce eBay’s revenues and devastate their profits, and probably reduce eBay’s capital value by seventy-five percent. (What say you about that scenario, John?)

Regrettably, eBay has become (if it was not always) simply a corporate structure that surrounds a group of most greedy, devious, unscrupulous, disingenuous, incompetent, managing executives whose only interest is the recovery of their “lost” —I suspect forever¬—performance bonuses, and whose total lack of ethics represents the ugliest face of the free enterprise system, ie, profits (read, bonuses) before principles—which, deservedly, is ultimately the road to oblivion for such organizations.

A most unscrupulous, disingenuous, dare I say, criminal organisation.

Good onya John, you really know how to grow a business!

The fact that eBay has been able to get away with effectively—and knowingly—aiding and abetting unscrupulous sellers to defraud consumers for so long is a damning measure of the effectiveness of governmental consumer affairs regulators the world over; for what is their purpose, if not to protect we simple peasants from the machinations of such unscrupulous organisations as eBay?

If somehow you missed the embedded link in the opening synopsis, the latest spreadsheet, and notes on the use thereof, can be downloaded from:
http://home.exetel.com.au/philipcohen/eBay/

Anyone wishing to offer comment on this matter can contact the author, Philip Cohen, at formset@exemail.com.au

PS: If anyone is interested in obtaining a single-page analysis of the bidders on an individual eBay auction—the way the competing bidder information should have been (more effectively) presented by eBay—try the link: http://www.ebuster.co.uk/bidhistory.aspx
_________________
Clearly, the lunatics at eBay have taken over the asylum and are bent on burning it down.
“The difference between genius and stupidity is that genius has its limits.” ~ Albert Einstein.