Regulating Relevancy: Initial Interest Confusion and the Internet byEric Goldman

Regulating Relevancy: Initial Interest Confusion and the Internet
Eric Goldman
Marquette University Law School

How Search Engines Work

  • Search robots copy web pages
  • Every word on every page is indexed into large database
  • People search database using keywords
    • Keywords encapsulate consumer interests
  • Search engine displays search results
    • Results often list title, URL, initial text
    • Results ordered by proprietary algorithms
      • Google uses a combination of link weighting and keywords (location matters)
      • Some results based on pay-for-placement
    • Some search results may be cut off


  • Trademarks are words/symbols that identify/distinguish the source of goods in the marketplace
  • Trademark infringement occurs when there is likelihood of consumer confusion
  • Each circuit has its own multi-factor test to determine likelihood of confusion
    • Sleekcraft in 9th Circuit; Polaroid in 2nd Circuit
    • 7th Circuit “digits”: mark similarity; product similarity; area and manner of mark’s use; consumer care; mark strength; actual confusion; and intent to palm-off (Forum, Schwinn, AHP)

Initial Interest Confusion

  • IIC is “the use of another’s trademark in a manner reasonably calculated to capture initial consumer attention, even though no actual sale is finally completed as a result of the confusion” (Brookfield)
  • Historically, IIC was analyzed in connection with “purchaser care,” “actual confusion” or “competitive proximity” factors in multi-factor likelihood of confusion test
  • Starting with Brookfield, some courts treat IIC as a bypass to the multi-factor test

Brookfield v. West Coast

  • Video rental store launches website at “” and uses “moviebuff” in metatags
  • High-end entertainment publisher has senior TM rights in “moviebuff”
  • Parties have some competitive proximity
    • Some searchers might accept defendant’s database instead of continuing to seek plaintiff’s
  • Using standard multi-factor test, court concludes the domain name infringes

Brookfield and Metatags

  • West Coast used “moviebuff” in metatags
    • Metatags are in hidden portions of web page
    • Historically, some search engines gave extra relevancy credit to metatags
  • Court says standard multi-factor test doesn’t apply to metatag analysis
  • The billboard analogy
  • Metatags created initial interest confusion
    • Misappropriation of search engine traffic because of goodwill association
    • But West Coast can say “Why pay for MovieBuff when you can get the same thing here for FREE?”

Post-Brookfield Abuses

  • Website criticizing/parodying newspaper (OBH)
    • Momentary confusion trumps disclaimer and negates parody defense
  • Website criticizing tax negotiator (JK Harris)
    • No fair use when website tried to improve ranking in relevancy algorithms
  • Website listing dealers of used equipment (Caterpillar v. TeleScan)
  • Website marketing travel services (hotels, airfare) to conference attendees (Key3Media)

Why IIC Doesn’t Work

  • Doctrine lacks a test or good definition
    • Possibility v. likelihood of confusion
  • Every word on a web page acts as marketing
    • It’s not just about metatags or domain names
  • Questionable assumptions about search behavior
    • Consumers searching on TM expect to find only TM owner
    • Consumers expect perfect relevancy in search results
    • Hitting the back button is a “harm”
    • Users reviewing search results are confused about what’s at the destination
    • Consumers stop their searches mid-stream
    • Users guess at domain names
    • Metatags make a difference in relevancy algorithms
    • Search engines don’t change their relevancy algorithms
    • Consumers tolerate search engines with irrelevant results
  • Permits TM owner to “own” a word

What Should We Do?

  • Eliminate IIC as a standalone way to establish likelihood of confusion
    • Consider the totality of circumstances to assess likelihood of confusion
    • Don’t allow possibility of confusion to suffice
  • Go back to using IIC as non-dispositive support for specific factors of multi-factor likelihood of confusion test
    • Should apply only when marketer is trying to arbitrage switching costs
  • Rely on search engines to continue improving relevancy algorithms