By Hubert Walker for CoinWeek ….
As reported last month by CoinWeek, Heritage Auctions filed a lawsuit in the United States District Court, Northern District of Texas in Dallas on December 9 against Christie’s auction house and Collectrium, the tech startup it acquired in 2015. Court filings list eight counts detailing multiple allegations of copyright infringement, computer fraud and unfair competition, among other complaints. Announced on Monday, December 12, the suit alleges that Christie’s “unlawfully appropriated almost three million auction sale listings from the Heritage website.” This information was then, according to Heritage, imported into Collectrium’s searchable database.
The following article examines the lawsuit in greater detail and seeks to provide some context outside of the world of numismatics.
The Parties Involved
Heritage is an American auction company based in Dallas, Texas. It was founded in 1982 when Steve Ivy and Jim Halperin merged their businesses to better weather the early 1980s downturn in the numismatic marketplace. Both men had begun their coin careers as teenagers, and by the mid-’70s their respective companies had independently branched out into auctions. Over the last 34 years, Heritage has grown to become a company that sells almost one billion dollars worth of items at auction, with $860 million USD sold in 2015.
Officially founded in 1766 by Scottish auctioneer James Christie, Christie’s of London is one of the world’s oldest and most successful auction houses. It is arguably the world’s leading art business, with offices in Rome, London, Geneva, Hong Kong and New York, to name a few. Over the last couple decades, Christie’s has expanded through a series of sales and acquisitions (including the ownership of Spink and Son between 1993 and 2001), with the company itself having been owned by François Pinault’s Groupe Artémis S.A. since 1998.
Collectrium is a collection management platform for a variety of collectibles, including fine art, luxury goods, antiques and coins. Its key points are high-resolution photo handling, attributes and provenance. The company was founded in 2009 by Boris Pevsner and started life as a cloud-based phone, tablet and pc application described as “Shazam for art”. In 2011, Bloomberg BusinessWeek named Collectrium one of “America’s Most Promising Start-Ups” (4/15/11). Christie’s bought the small company in 2015 for a reported $16 million (a big deal in the art world), and now provides access to the software and Collectrium’s database as a “perk” for their own customers. Collectrium does, however, remain independently managed by the Collectrium team.
According to Heritage, Christie’s used “web-crawling” software to “scrape” proprietary pictures and information from the company’s website, HA.com. This activity first came to Heritage’s notice in July of 2016. A web crawler, or spider, is a kind of internet bot, which is any software that performs an automated task or tasks over the Internet. The web crawler’s specific task is to index pages, files and information on one or more websites. After this is done, the web-scraping software or bot extracts the relevant data from the website.
Then, Heritage alleges, once Christie’s had ported the images and data onto their own computers–sometimes copying item listings verbatim–the London-based auction house rebranded and resold the data as its own as part of a subscription database. Christie’s will not comment on the allegations, but does refer to the Collectrium software as being in “beta”.
By the time the complaint was filed, Heritage had traced the illicit activity to 40 different accounts, one of which was registered to one “Jason Bourne” from Katmandu. The spider itself was associated with an account registered to one Leanne Wise; Heritage was able to connect that account to a computer that was also shared by the Director of Product Development at Collectrium, David La Cross. The scraping software, Heritage alleges, had been actively working on HA.com since March 2016, accessing some five million pages by the time Heritage suspended all user accounts involved with the web crawler and scraper. However, according to lawsuit Collectrium seemed to have found a way around the company’s ability to detect its activity on the site, since the database had continued to post listings on its own website that were posted after the account suspensions.
Heritage claims that almost three million of the self-proclaimed 11 million items in Collectrium’s database were pulled from HA.com. Therefore, Heritage Auctions is asking the court for a preliminary injunction on all data scraping and for Collectrium to close its website for as long as it takes for all proprietary material from HA.com to be removed. Heritage also seeks monetary damages, which are detailed below.
The case seems to have been spurred on by a recent email from Christie’s to its customer base, inviting them to try a searchable database called Collectrium Market Data Beta. Christie’s claimed that this proprietary database consisted of seven million auction items plus sales results.
Heritage alleges that much of the content in Christie’s recently bought Collectrium database was proprietary material stolen from Heritage. According to the Dallas-based company, approximately 2.7 million of the 11 million listings in the Collectrium database were sourced illegally from HA.com. Allegedly, Christie’s is using Collectrium to misrepresent Heritage Auctions’ copyrighted material as its own. Heritage claims that this action diverts customers from Heritage to Christie’s, and also misleads customers as to the extent of Christie’s actual expertise in various fields–including numismatics. Heritage believes their sales in coins, currency, entertainment and sports memorabilia and comics and comic art, et alia, are bigger than those achieved by Christie’s.
Original content on HA.com is the “centerpiece” of the website, according to the complaint. The images on HA.com were taken by Heritage employees and independent contractors paid by Heritage to supply them. Every item ever offered at sale or sold by Heritage since 1993 is included on the site, which features over eight million images to illustrate the various entries. Heritage is also responsible for writing original descriptions of every item.
In all, Heritage claims to have spent over $20 million building HA.com over time–in addition to the money that the company has spent to accumulate the data that Christie’s allegedly stole.
Access to and use of this copyrighted original content is delimited by the HA.com website use agreement, which Heritage alleges that Christie’s violated when it set up the accounts that Heritage has tied to web crawling and scraping activity. All registered users–including those that may or may not have been registered by Christie’s, Collectrium, or their employees and contractors–must agree to the following terms:
- Registered users must agree and acknowledge that the information and images, etc., on HA.com are owned exclusively by Heritage Auctions.
- User access is licensed for limited use, i.e. it is only for personal, not commercial, use. One individual image may be used to accompany a news article written without Heritage’s written consent. But again, this is only if the writing is not commercial, and only if the image is credited to Heritage.
- The registered user must agree that there are a few prohibited uses of the website. The user of HA.com may not republish, distribute commercially, duplicate or “exploit” any aspect of the website, from code to content.
- The user my not download, reproduce, modify, distribute, transfer, sell or create derivative works utilizing any of the exclusive content on HA.com except as presented above.
Heritage states that any unauthorized use of its website may result in civil or criminal prosecution.
Like many Terms of Service (TOS) agreements that one clicks through on a daily basis, the user agreement specifies that merely using the website acknowledges agreement with and acceptance of the following statements:
- That a user may not scrape data or use a robot or spider on the site;
- That a user may not gain unauthorized access (though this statement is vague);
- That a user may not use site content in violation of the property rights of Heritage Auctions;
- That a user may not “reverse engineer” the site;
- That a user may not inject materials into the website;
- That a user may not remove watermarks from website image assets; and finally
- That a user may not burden or slow down the site maliciously, evidently referring to a denial of service (DoS) attack.
An individual user on HA.com must register in order to see prices realized for each item listed. This by definition means that the user must agree to all of the above statements and limitations in order to see price information generated by Heritage Auctions.
The company provides two services to the registered users of its website that it feels deserve special mention and comparison in the complaint. There is a “My Collection” Service, which helps users organize their collections. It benefits Heritage by identifying future bidders and consignors and by allowing custom advertisements to be displayed for each user. It also serves to collect information useful to any market evaluation and analysis that Heritage may see fit to undertake. Customers are also more likely to use Heritage’s auction services since the company already has their information. The service also provides a public option that lets other people view and buy items from a registered user’s collection.
The second service Heritage mentions in its lawsuit is “My Wantlist”. Here, users can enter the criteria they use when they search for additions for their collections. Heritage then pulls up past data on such items from the company’s database. This allows Heritage to market more effectively to customers it knows already want the specific products it has offered or is offering currently. And like the “My Collection” service, it enables Heritage to better evaluate market demand.
Both services Heritage explicitly identified as especially harmed by the actions of the web spider and scraper.
The behavior of the web spider had two stages. First, it accessed HA.com without logging into a registered user account. When it was doing this, Heritage alleges, the spider was identifying which pages contained sales data. Once it had–the complaint is not specific as to whether the spider identified ALL such pages first or if it found one page and immediately scraped it–the spider logged into registered user accounts to scrape said information.
Heritage traced the spider back to the registered user account of one Leanne Wise of Boiling Springs, South Carolina. Heritage assigns what they refer to in the complaint as a “Guaranteed Unique Identifier”, or GUID, to each computer that accesses the site – though they are most likely referring to a “Globally” Unique Identifier. A GUID is a very large number, often 128 bits in length, that is virtually guaranteed to be unique. Please note, this brief definition does not do justice to how extremely unlikely the intentional duplication of another person’s GUID is; one article explaining how GUIDs work states that assigning a unique GUID to every object on the internet once every second for every person on earth would take over a billion years to use up all the possible GUIDs under the current system.
At any rate, Heritage claims that they were able to use the GUID associated with the Wise account (under which the spider was operating) to confirm that it was the same computer as that used by David La Cross. The IP (Internet Protocol) address was also the same, but IP addresses are far more vulnerable to spoofing.
HA.com alerts the parent company when unusual traffic is coming in from a single IP address and or registered user account. Heritage also claims that Christie’s use of IP randomization made the task of determining unusual activity more difficult – which the complaint implies was done intentionally, for the purposes of avoiding attention from Heritage’s system. While such IP randomization can be used in this manner, it is also a privacy tactic used innocently by some people online, and is not necessarily malicious.
According to Heritage, the Wise account accessed HA.com over 186,000 times in nine days. Upon discovering this unusual activity, the account was blocked. The spider, however, continued to scrape the site, and by the time of the complaint, a total of 31 accounts were confirmed by Heritage to have employed the spider. The entertainingly named “Jason Bourne” account, for example, accessed the website 131,000 in eight days. It should be noted that while using a fake name is not necessarily proof of maliciousness, it is a clear violation of the website’s terms of service as delineated above.
Approximately five million pages were accessed between March and October 2016, with proprietary information from the Heritage website posted–again, according to Heritage Auctions–on Collectrium as recently as November 19. Heritage says that it can prove that scraping occurred even after the accounts were suspended but admits that the company does not know how Collectrium is collecting the information now.
Based on what it alleges was the abusive use of its website, Heritage’s suit declares that Christie’s and Collectrium employ a “Fraudulent business Model”. Heritage accuses Collectrium of not actually being a tool for collectors but is instead really just a way to bring business to Christie’s. Heritage goes so far as to describe Collectrium as a “Trojan Horse”, which has its own connotations in the field of computer security.
Collectrium’s auction data is only available to paying account holders, so if the allegations were true, then Heritage’s proprietary information would add significant value to Collectrium in multiple ways. Not only would it serve to “inflate” Collectrium’s numbers and make the database more appealing to customers, but the Heritage data would also promote future sales for Christie’s, and potentially encourage consignors to sell through Christie’s instead of Heritage.
Collectrium claims it is independent of Christie’s and will not share customer information with Christie’s unless a customer gives his or her consent. Heritage alleges that Collectrium tells Christie’s immediately whenever a new account is made, and that this initiates the creation of a new Christie’s account as well–whether the creator of the Collectrium account desired one or not. This implies that all Collectrium passwords and security question answers, plus emails, are shared between the two sites – only one of which, Heritage alleges, was authorized to know or keep such private information.
Regardless, Christie’s appears to use Collectrium in much the same way as Heritage uses the two services mentioned previously, which indeed gets to the crux of the matter. Each company hopes to gain consignments before a customer consults another auction house. This is especially important in those fields where Christie’s is not the industry leader, such as coins.
As of August 31, 2016, about 150,000 of the seven million listings on Collectrium allegedly originated from Heritage’s website. By December, the number of listings in the Collectrium database had increased to 11 million, and Heritage alleges that 2.7 million of those were originally from HA.com. Heritage alleges that, in most instances, content was copied verbatim, though references to Heritage’s sales of similar items were removed, and Heritage’s copyright notices were replaced those from Christie’s or Collectrium.
According to Heritage, Collectrium essentially allows its customers to view Heritage’s proprietary data without visiting the Heritage website, thereby circumventing Heritage’s marketing efforts.
Among the harm to the company that Heritage alleges is the time and money spent tracking and thwarting this situation; any and all business potentially and actually lost because of it; the reduction of Heritage Auctions’ ability to distinguish itself as a brand (one auction company appearing to be as good as any other); lost licensing fees that Christie’s would have had to pay in order to do what it has allegedly done in a legal manner (Heritage claims this cost would be in the millions of dollars); and the loss of control over how its data is ultimately used.
The complaint alleges eight counts against Christie’s and Collectrium.
1.) Count one is copyright infringement as defined under Sections 106 and 501 of the Copyright Act (17 USC). Under this law, Heritage would be entitled to damages per instance of $150,000 USD (section 504), though it is up to a jury to decide actual damages. An instance is defined as an individual item listed on HA.com. Also included under this count is a request for attorney’s fees under 17 USC Section 505.
2.) Count two is of violations under the Computer Fraud and Abuse Act. Heritage computers are involved in interstate and foreign commerce and communication under 18 USC Section 1030(e)(2).
Heritage alleges that Collectrium “knowingly violated” 18 USC sections 1030(a)(2)(C) and (a)(4), which, they claim, resulted in at least $75,000 in real damages. Were Heritage to win the case, the actual damages the company would receive is yet to be determined.
3.) Count three concerns violations of the Digital Millennium Copyright Act (DMCA). This count focuses on the fake accounts as violations of the DMCA under 17 USC section 1201(a). The Digital Millennium Copyright Act was signed by President Bill Clinton in 1998. It is more commonly known as a tool to enable the music industry and recording artists to go after illegal downloaders of music and mp3s.
Per section 1203(a) of 17 USC, Heritage would be entitled to recover statutory damages of no less than $200 or more than $2,500 per act of circumvention according to 17 USC section 1201, and no less than $2,500 or more than $25,000 per section 1202 violation. It is important to note that each auction item in question is both a circumvention AND a violation. Heritage also requests attorney’s fees and injunctive relief as permissible under 17 USC section 1203(b).
4.) Count four addresses harmful access by computer as defined under the Texas Civil Practice and Remedies Code sections 143.001 and 143.002. These charges under Texas code concern the alleged access by Collectrium of Heritage computers without consent, the burdening of Heritage computers, the impairing of the integrity of Heritage’s copyrighted work, and the intentional fraud wrought by the use of multiple fake accounts. Attorney’s fees and injunctive relief is sought.
5.) Count five is of trespass, based on the allegation that Christie’s and Collectrium entered the HA.com website using fraudulent information and then interfered with Heritage’s possession and use of its own property. The parent company, Heritage Capital Corporation HCC, is pursuing this count against all defendants; injunctive relief is sought.
6.) Count six is one of unfair competition in violation of Texas law. Injunctive relief is sought. The complaint does not go into much detail here, but this count seems to be the crux of the matter as mentioned above.
7.) Count seven concerns civil conspiracy, alleging that Christie’s and Collectrium conspired together against Heritage Auctions. Punitive damages are sought, and would be determined in court.
8.) The final count, count eight, is of breach of contract as alleged by the parent company HCC against all defendants. This count concerns the violations of the Website use agreement. Damages would be determined at trial.
As for damages and awards, Heritage specifically asks for the following:
- Statutory damages of $150,000 USD per violation of the Copyright Act, $2,500 per violation under 17 USC section 1201 and $25,000 per violation under 17 USC 1202 – or as otherwise permitted by law;
- Actual damages to be determined at trial;
- A preliminary injunction for Collectrium to shut down access to their website until all proprietary content from heritage is removed;
- A preliminary and permanent injunction to prevent Christie’s and or Collectrium from accessing HA.com using false information or robots; continuing to infringe Heritage’s rights; selling, distributing and publishing content that rightfully belongs to Heritage Auctions; removing, downloading or copying content from HA.com; or otherwise committing acts of unfair competition;
- That Collectrium and Christie’s return or destroy all copies of Heritage property in their possession;
- The “disgorgement of unjust gains”;
- Pre-judgement interest;
- Punitive damages to be determined at trial;
- Attorney’s fees; and
- Anything else the court deems right.
The complaint was filed by Mark W. Bayer (Barnes and Thornburg LLP) of Dallas, Texas and Armen R. Vartian and Laura C. Tiemstra of the Law Offices of Armen R. Vartian. Vartian is a specialist in the law concerning art and other collectibles, and serves as legal counsel to the Professional Numismatists Guild (PNG) and is based in California. He is the author of Legal Guide to Buying and Selling Art and Collectibles (1997).
Attached to the lawsuit are three examples of objects sold for auction by Heritage and whose listings on HA.com the company alleges that Christie’s and Collectrium copied illicitly into their database.
Exhibit A is a 1792 P1C Birch Cent, Judd-5, Pollock-6, R.8 MS61 Brown NGC. It sold at a Heritage auction in August of 2016. Exhibit C is the same item, same information as seen on Collectrium’s website.
Exhibit B is the late rock musician Prince’s yellow cloud guitar, which Heritage sold in June of 2016. Exhibit D is a yellow cloud guitar owned by Prince as seen in the Collectrium database.
Exhibit E is a 1907 $20 UHR sans serif edge Judd-1907, Pollock-201, R.8 PR58 PCGS, which was sold by Heritage in August of 2012.
When considering Heritage’s case against Christie’s, it is important to be mindful of precedent, both legal and technical.
In non-competitive cases, other courts have found in favor of web scraping before, but what would happen in a legal battle between business competitors is not so clear. In the 2000 case eBay v. Bidder’s Edge, eBay sought and won relief from an auction aggregator website that was scraping data from eBay’s servers.
The Heritage suit also comes after the Magnus art-pricing mobile app by Artace Inc. was pulled from the Apple Store after art education website Artsy accused Magnus and Artace of doing something similar to what Heritage alleges that Christie’s has done.
Other tech companies have also gotten into the art tech game, and how they deal with the legal and ethical issues of copyright and intellectual property will provide important context. For example, search engine giant Google is actively developing its Google Art Project.
And finally, the December complaint is actually the second lawsuit Heritage has brought against Christie’s. In 2014, Heritage Auctions sued Christie’s for $40 million, alleging that Christie’s had enticed three “luxury handbag specialists” away from Heritage. The group included Matthew Rubinger, who was Heritage’s main expert on Hermès designer handbags. The 2014 case will go to trial sometime in 2017. For their part, Christie’s claims that the case is “without merit”.
As usual, CoinWeek will continue to follow the story.
* * *