What are the facts?
eBay operates a large online auction platform maintained on a network of computer servers. Bidder's Edge (BE) ran an aggregation service that collected and displayed auction listings from numerous sites, including eBay, to help users compare offerings across platforms. Initially, BE and eBay discussed a data-licensing arrangement, but negotiations failed. eBay maintained measures to manage automated access—publicly posting a robots.txt file prohibiting crawling of many dynamic search functions, employing technological blocks (including IP blocking), and incorporating terms of use and a User Agreement restricting automated querying without permission. Despite eBay's policies, BE used software robots or spiders to send automated search queries to eBay's site for the purpose of harvesting listings. After eBay sent cease-and-desist letters and implemented IP filters, BE continued scraping by routing requests through third-party proxy servers to evade those blocks. Evidence showed BE's robots generated on the order of tens of thousands of requests per day (approximately 80,000–100,000), constituting roughly 1–2% of daily traffic to eBay's site. eBay alleged this load consumed system resources, risked service degradation, and threatened to invite similar conduct by others. eBay sued under several theories, including trespass to chattels, and sought a preliminary injunction to stop BE's automated access. The district court addressed eBay's likelihood of success on the trespass claim and the equitable factors governing injunctive relief.
What is the legal issue?
Does a web crawler's unauthorized, automated access to a publicly available website constitute a trespass to chattels by impairing the function and value of the site operator's computer systems, thereby justifying a preliminary injunction?
What rule applies?
Under the Restatement (Second) of Torts §§ 217–218, trespass to chattels occurs when a person intentionally interferes with another's possessory interest in personal property, and that intermeddling either dispossesses the owner or impairs the condition, quality, or value of the chattel, or deprives the possessor of its use for a substantial time. Courts, including in CompuServe Inc. v. Cyber Promotions, Inc., have applied this doctrine to computer systems, holding that unauthorized electronic signals or automated activities that burden or diminish a server's functioning may constitute a trespass. For a preliminary injunction in the Ninth Circuit, the movant must show a likelihood of success on the merits and the possibility (or likelihood) of irreparable harm, or, alternatively, serious questions going to the merits and that the balance of hardships tips sharply in the movant's favor, with consideration of the public interest.
What did the court hold?
The court granted eBay's motion for a preliminary injunction, holding that eBay demonstrated a likelihood of success on its trespass to chattels claim and a sufficient showing of irreparable harm and balance of equities. Bidder's Edge was enjoined from accessing eBay's computer systems through automated robots, spiders, or other mechanical means without authorization.
What is the reasoning?
Trespass to chattels. The court found that eBay possessed a legally cognizable property interest in the physical components of its network—its servers and associated computing resources. BE's automated queries were intentional and, after eBay's explicit revocation of consent via cease-and-desist letters, robots.txt prohibitions, and IP blocking, unauthorized. Even though BE argued its scraping imposed only a small percentage of total traffic, the court concluded that the use nonetheless impaired the condition and value of eBay's servers by consuming processing, memory, and bandwidth that eBay could otherwise allocate to customers and business operations. The court emphasized the ongoing nature of the intermeddling and the realistic risk of cumulative harm if other aggregators followed suit. Such impairment sufficed under the Restatement standard, particularly where the injury to system performance and reliability need not be catastrophic to be actionable. Public website and consent. BE contended that because eBay's site was publicly accessible, its access could not be a trespass. The court rejected this, reasoning that any implied license to access a public site is limited and revocable. eBay's explicit restrictions (terms of use and robots.txt) and targeted revocation (cease-and-desist) negated authorization for automated scraping, and BE's deliberate evasion of IP blocks by using proxy servers confirmed knowing, nonconsensual intermeddling. Harm and injunctive relief. In assessing preliminary injunctive relief, the court accepted evidence that BE's traffic consumed measurable system resources and risked interference with eBay's operations and user experience. The court also credited the prospect of a flood of similar scrapers multiplying the burden. Weighing the equities, the court found that the hardship to eBay from the ongoing impairment and the risk to system integrity outweighed BE's claimed need to continue unauthorized scraping. The public interest favored protecting the reliability of online marketplaces and upholding property and contract-like boundaries in cyberspace. The court relied on analogous precedents applying trespass to chattels to electronic communications, such as CompuServe v. Cyber Promotions, and distinguished contrary outcomes that lacked proof of impairment. Accordingly, the court concluded that eBay was likely to succeed on its trespass claim and faced irreparable harm absent an injunction.
Why is this case significant?
eBay v. Bidder's Edge is a touchstone in cyber-trespass jurisprudence, widely cited for recognizing that automated scraping of a website can constitute trespass to chattels when it is unauthorized and imposes even modest but nontrivial burdens on servers. The case illustrates how traditional tort doctrines adapt to digital property and how technical measures, policies, and cease-and-desist letters can define the scope of consent online. For law students, the decision highlights the relationship between property interests in computer systems, contract-like site terms, and equitable remedies. It also frames later debates: subsequent cases, notably Intel v. Hamidi, tightened the requirement of showing actual functional impairment under California law, and more recent litigation over scraping (e.g., hiQ v. LinkedIn) explores different statutory frameworks like the CFAA. eBay remains essential for understanding early judicial efforts to police automated access and resource consumption on the web.
Does making a website publicly accessible amount to blanket consent for automated scraping?
No. The court held that any implied license to access a public site is limited and revocable. Website operators can condition or withdraw consent through terms of use, robots.txt directives, cease-and-desist letters, and technical blocks. Continuing automated access after consent is revoked, especially by evading blocks, supports a finding of unauthorized intermeddling for trespass to chattels.
How much harm must be shown to prove trespass to chattels in the online context?
Under eBay v. Bidder's Edge, even a relatively small but nontrivial burden on server resources can satisfy the impairment element, particularly when the interference is ongoing and risks cumulative harm. However, later decisions, such as Intel v. Hamidi, emphasize that some actual functional impairment or disruption is required under California law. The precise threshold can be context-dependent, but trivial or de minimis contact alone is generally insufficient.
What role did robots.txt and eBay's technical measures play in the court's analysis?
They were key indicators of the scope of authorization. eBay's robots.txt and terms of use communicated restrictions on automated access. Its IP blocking demonstrated active revocation and enforcement. BE's circumvention of these measures by using proxy servers showed intentional, unauthorized access, strengthening eBay's trespass claim and supporting injunctive relief.
Did the court rely on the Computer Fraud and Abuse Act (CFAA) to issue the injunction?
No. While eBay asserted several claims, the court's preliminary injunction rested primarily on the common-law tort of trespass to chattels. The decision focused on eBay's possessory interest in its servers and BE's unauthorized, resource-consuming access rather than on the CFAA's statutory thresholds for unauthorized access or damage.
Would the ruling prohibit all search engines from crawling public websites?
Not necessarily. The court did not issue a blanket rule against crawling. It addressed a specific context in which the site owner revoked consent, posted restrictions, and the defendant intentionally evaded technical blocks while imposing measurable burdens. Search engines that have permission, comply with robots.txt, or operate under licenses or accepted norms are situated differently from a party that continues scraping after consent is withdrawn.