Eric Goldman, Deregulating Relevancy in Internet Trademark Law

• Citation: Vol. 54 No. 1
• Publication year: 2005

DEREGULATING RELEVANCY IN INTERNET TRADEMARK LAW

Eric Goldman^*

I. A BRIEF PRIMER ON INTERNET SEARCH ............................................. 511

II. INTERNET SEARCH FROM THE SEARCHER'S PERSPECTIVE ................. 513

A. The Complex Process of Internet Search ................................... 513

1. Stage 1: Objective Formulation ........................................... 515

2. Stage 2: Search Provider Selection ...................................... 515

3. Stage 3: Keyword Selection .................................................. 515

4. Stage 4: Search ..................................................................... 516

5. Stage 5: Results Evaluation .................................................. 516

6. Stage 6: Decision .................................................................. 519

7. Stage 7: Investigation ........................................................... 520

8. Stage 8: Objective Satisfaction ............................................. 521

B. Trying To Infer Searcher Objectives from Keyword Selection ... 521

1. "Objective Opaqueness" ...................................................... 521

2. Internet Searches Differ from Physical Space Searches ...... 526

III. INTERNET SEARCH FROM THE PUBLISHER'S PERSPECTIVE ................. 529

IV. INTERNET SEARCH FROM THE SEARCH PROVIDER'S PERSPECTIVE .... 532

A. Search Providers Are Editors, Not Passive Intermediaries ....... 532

1. Aggregation .......................................................................... 533

2. Sorting ................................................................................... 534

3. Filtering Content .................................................................. 537

4. Reinterpreting Searcher Keywords ...................................... 538

5. Editing Publisher Content .................................................... 539

6. Reinterpreting Publisher Keywords ..................................... 541

7. Advanced Technologies ........................................................ 541

8. Summary on Passivity .......................................................... 542

B. Keyword Convergence ................................................................ 542

1. Domain Names ..................................................................... 543

2. Link Navigation .................................................................... 549

3. Software Agents .................................................................... 550

4. Implications of Keyword Convergence ................................. 551

V. THE DEVOLUTION OF INTERNET TRADEMARK LAW ........................... 552

A. A Brief Primer on Trademark Law ............................................. 552

B. The Breakdown of the MFLOCC Test and Goodwill

Misappropriation ........................................................................ 553

1. Referential Uses ................................................................... 554

2. Associative Uses ................................................................... 556

3. Misappropriative Uses ......................................................... 558

C. Initial Interest Confusion ............................................................ 559

1. The Development of the Doctrine ......................................... 559

2. The Failings of the Doctrine ................................................. 565

3. Deconstructing Brookfield ................................................... 570

4. Consequences of a Bad Doctrine ......................................... 573

VI. DEVELOPING AN INTEGRATED LAW OF KEYWORDS ........................... 575

A. Move Infringement Analysis Back in Time ................................. 575

1. Searchers Benefit from Filtering Content ............................ 576

2. At Early Search Stages, There Is Less Reason To Believe that Trademark Owners Suffer Any Harm ............................ 581

3. Summary on the Timing of Infringement Analysis ................ 584

B. Modernize the MFLOCC Test .................................................... 584

1. Add a "Relevancy" Factor to the MFLOCC Test ................ 584

2. Consider the Search Stage ................................................... 585

3. Consider Minority Definitions .............................................. 586

4. Require Plaintiffs to Provide More Rigorous Evidence of Confusion or Harm ........................................................... 587

C. Immunize Search Providers ........................................................ 588

1. Why Search Providers Need a Safe Harbor ......................... 588

2. Implementing a Safe Harbor ................................................ 593

CONCLUSION .................................................................................................. 595

DEREGULATING RELEVANCY IN INTERNET TRADEMARK LAW

Jane, a 10th grader, is writing a report on ancient Greek religious figures for her classical history class. She enters the word "Nike" into the Google search engine in an attempt to find source material for her report. However, she only finds product information regarding footwear and clothing from Nike, Inc., because Google has removed other content due to liability concerns. Out of frustration, Jane abandons her Internet search and turns to the next most convenient free resource, her school library. Unbeknownst to Jane, her school library's acquisition director is personally interested in Roman mythology but finds Greek mythology boring. As a result, Jane only finds a few books briefly discussing her subject.

John is a rock climbing enthusiast with a penchant for the latest gear. He has heard rumors about the new Reebok "Hilltopper 483," an advanced technology shoe for rock climbing intended to directly compete with the just-released Nike "Swooshtown 96" model. He would like to educate himself about each model to decide if he wants to purchase the newest, hottest rock climbing footwear. He searches for "Reebok Hilltopper 483" in Yahoo Search but this only leads him to the official website of Reebok, Ltd. and its authorized distributors. These sites have some information about Reebok shoes but no neutral or comparative information. Frustrated by his inability to find the information he wants, John decides not to buy new shoes from either Reebok or Nike.

These two scenarios illustrate how trademark law could jeopardize the Internet's potential as an information resource and a catalyst for competition. Emerging trademark law doctrines have allowed trademark owners to excise socially beneficial content and to take unprecedented control over their channels of distribution. Without limits, trademark law has the capacity to counterproductively destroy the Internet's utility for everyone.

It does not have to be this way. Trademark law's assault on the Internet is predicated on a series of factual myths and doctrinal errors. The clichéd story assumes that predatory marketers steal customers' attention away from trademark owners using surreptitiously dishonest means.¹Under this story, it is argued that trademark law needs to protect trademark owners and searchers by ensuring that Internet searchers see only content approved by the trademark owner.

This cliché, and its supporting academic commentary, have developed from intuition-based assumptions about searcher behavior that contradict the empirical data about how searchers actually search. This Article corrects those mistakes by basing its analysis on what information scientists know about Internet search behavior.

By examining how people search for information on the Internet, it becomes clear that Internet search providers compete fiercely with each other to help Internet searchers find the content they want. To do so, search providers must discern searchers' true objectives from opaque and poorly- chosen search keywords delivered without any surrounding context. This requires search providers to choose what content to present to searchers, putting publishers trying to reach those searchers at the mercy of the search engine.

Using the facts presented by the Article, it becomes possible to develop an Internet trademark policy from the ground up. To increase the odds of efficient and successful searches, searchers should be able to pick the search terms they want, and search providers should be able to use those search terms to deliver the most helpful content to searchers. Thus, trademark law must step aside when searchers receive relevant content they may want.

In short, the introductory vignettes above may realistically preview the current direction of Internet search. However, with a more rigorous understanding of both trademark law and Internet search, there is an opportunity to preserve the Internet as a medium that helps searchers find what they seek.

Part I of the Article provides a brief overview of the Internet search process. Parts II-IV consider Internet search from three perspectives. Part II considers Internet search from the searcher's perspective, concluding that one cannot infer searchers' objectives from the keywords they choose. Part III considers Internet search from the web publisher's perspective. Part IV considers Internet search from the search provider's perspective, explaining that search providers, not web publishers, decide what searchers see. Part IV also explains that all Internet technologies are converging to use keywords to match searchers with content. Thus, historical distinctions between the domain name system, directories, and search engines are collapsing.

Part V summarizes trademark law and then provides a theoretical framework to distinguish beneficial and misappropriative uses of trademarks. The framework helps explain why trademark law doctrines like "initial interest confusion" are misguided. Part VI offers three concrete solutions: (1) courts should consider the search process stage where searchers see a trademark being...