Understanding Internet Co-Branding Agreements by Eric Goldman and Candice Lee


by Eric Goldman, Esq., and Candice Lee


The Internet has spawned new business practices regarding the ways users access and obtain information and services.  Because linking can create a network of web pages that appear integrated and seamless to users, many Internet companies enter what are known as co-branding relationships.  This article addresses a common type of co-branding relationship in which a “provider” maintains a set of pages (“the co-branded site”) that looks and feels like the “brander’s” web site.[i]  The co-branded site is promoted on the brander’s web site through linking.

Co-branding relationships have become ubiquitous on the Internet, particularly in light of the emergence of “portals.”  The term “portal” is used in many contexts.  In this article, portals are branders with consumer-oriented web sites that aggregate a wide range of information and services.[ii]  Many of the services offered by portals are actually provided by a third party on a co-branded basis.[iii]  For example, when a user clicks on the “stock quotes” link on the Excite homepage, the user is taken to another Excite page, clearly labeled by the provider, Quicken.com.[iv]  This page is part of the co-branded site where Excite is the brander and Quicken is the provider.

From a documentation and drafting standpoint, co-branding agreements can range from little more than trademark licenses to full-blown outsourcing agreements.  In all cases, the heart of the co-branding relationship consists of a trademark license from the brander to the provider.  In exchange, the provider agrees to furnish services to the users generated by the brander.  In a co-branding relationship, there are a number of complexities that must be addressed if the parties are to reach their objectives.  The remainder of this article discusses the major issues involved when creating and negotiating a co-branding relationship.


A.   The Brander’s Perspective

There are a number of reasons why a brander enters into co-branding deals.  Like any other outsourcing arrangement, a co-branding deal allows the brander to take advantage of the provider’s expertise or economies of scale.  For example, the provider may have superior software tools or databases, and the brander simply may not be able to cost-effectively or time-effectively develop competing tools.  Thus, co-branding deals allow the brander to appear to have a larger web site, or to have a more extensive set of features, than it can operate on its own.  These additional resources allow a brander to offer “one-stop shopping” to its users and  help make the brander more attractive to advertisers.  Also, many providers are willing to pay the brander for the promotion the brander provides.[v]

B.   The Provider’s Perspective

A provider also has a number of reasons to enter co-branding deals.  First, a co-branding relationship takes the place of a licensing arrangement.  Rather than using its intellectual property in only one channel—its own web site—the provider can “distribute” its content and services in multiple channels, thereby potentially getting multiple revenue streams.  However, when the provider makes its services easily available on the Internet, the provider then faces the risk of channel conflict or cannibalization.[vi]  The provider can manage this risk by developing relationships with branders who have access to significantly different channels of users.

Second, the provider can “distribute” its content and services without actually having to provide a copy of the software, thereby avoiding difficult intellectual property protection issues.  For example, consider a publicly accessible database of facts, such as a directory of phone numbers.  Factual databases are currently subject to little protection under U.S. intellectual property laws.[vii]  They are unlikely to be covered by copyright law, which does not protect facts.  At best, a factual database will be subject to a thin compilation copyright,[viii] which can be easily circumvented. Further, because it will be made available to the public over the Internet, the factual database cannot be treated as a trade secret.[ix]  Therefore, historically a database owner’s sole option was to distribute the database using contract covenants as the only method of protection.  If the database escaped the control of a contract licensee, the owner had no power to stop downstream recipients from further “infringement.”

As a result of the thin or nonexistent intellectual property rights in factual databases, Internet databases are highly vulnerable to misappropriation.  However, by using co-branding deals, a database owner can use technology to control distribution of its content rather than relying upon contract and intellectual property law.

Another advantage of controlling the number of copies in circulation is that the provider can more easily ensure that all copies are “in sync” and current.  This can be a significant logistical consideration for complex, time-sensitive databases, such as reservations databases and for frequently updated software in which installation of updates would be complicated.

A final benefit of co-branding relationships is that users can be transferred from the co-branded site to the provider’s site.  By being exposed to new users, the provider may procure increased traffic for its own site.


Both parties will want to understand and ensure the accurate operation of the method used to track users referred by the brander to the co-branded site (the “referrals”).  Tracking referrals can affect such crucial issues as the calculation of the revenue stream subject to a split, the provider’s obligation to display the brander’s branding to referrals, and the parties’ rights to use data about the referrals.

There are four primary ways to track referrals.  First, the parties can establish a unique URL to identify the co-branded site, in which case the brander will direct its users to this URL.  With this method, the parties can establish rules regarding all activity occurring under the unique URL, an effective solution to the referral tracking problem.  However, many times the URL uses the trademarks of both the brander and the provider.  The parties should be aware that they may be creating a combination trademark and will need to carefully consider the rules regarding the use of the combination mark.

Second, the parties can require referrals to register with the provider at the co-branded site.  The referrals can then be tracked by requiring subsequent log-ins, by issuing the referrals a digital certificate, or by loading a token into their cookie (discussed below).[x]  Although registration is rarely the preferred approach given user antipathy towards such impediments, the parties can minimize this hurdle by having the brander “pre-populate” the provider’s registration forms with user information the brander already possesses.  This makes it easier for the referral to register with the provider, thus increasing the chances that the referral will do so.

Alternatively, it has become increasingly common for branders to allow users registering with the brander to check a box on the registration screen and thereby “co-register” with both the brander and the provider.  In these cases, the parties work out a data transfer mechanism for the brander to provide information about these users to the provider, at which point the provider automatically creates an account for the user.  In this way, a user is already registered when he or she accesses the co-branded site, and thus can be tracked as described above.

Third, the parties can place a cookie into the user’s cookie file without registration.  Although not very intrusive to users, the cookie method is not foolproof.  Users might refuse the cookie, edit their cookie file and delete the cookie, switch to a browser that cannot access the cookie, or use a browser that does not support cookies.  Further, when using cookies, the parties need to decide when the cookie should expire—the more quickly the cookie expires, the more quickly the brander will lose track of some referrals.

Fourth, the provider can track referrals by noting the URL the users were last visiting and, if that URL is one specified by the brander, treat the users coming from the designated URL as referrals. HTML protocols furnish the provider with the most-recent URL that users come from, so this type of tracking is not especially difficult technically.  With more sophisticated programming, the provider can even track these users as they travel around the co-branded site by placing a keyword or identifying symbol in the URL, which is checked as each new page of the site is accessed.  This technique is used infrequently for a number of reasons.  First, the method requires more complex programming to track users by URL as they move around the co-branded site.  Second, the brander might want to send referrals to the co-branded site using non-Web promotions, and this method cannot track these users.  Finally, most branders want to be able to track users who access the cobranded site multiple times. In order for a brander to successfully track users in this situation, the users would need to initiate each visit to the co-branded site via the brander’s site.


To generate traffic for the co-branded site, the brander must promote it.  Examples of some of the methods used for such promotion include:

Navigation Bars.  The brander can promote the co-branded site by providing links to it in the brander’s navigation bars on the brander’s site.

Co-Registration.  As described above, the co-branded site may be promoted on the brander’s registration page by offering the user the opportunity to co-register for the co-branded site by checking a box.

Editorial Content.  The provider can supply editorial content that the brander publishes on its web site.  This editorial content can act as a powerful advertising tool that can induce users to find out more by following the brander’s link to the co-branded site.

Advertising/Sponsorships.  The brander can promote the co-branded site using e-mail newsletters the brander sends to its registered users, by featuring the site or the provider’s trademarks in contests/sweepstakes, and by running promotions such as banner ads, button ads and text links.

In most cases, the provider will want the brander to ensure some minimum level of promotion.  For instance, the brander may promise to deliver a minimum number of users who “click through” to the co-branded site, or a minimum number of users who register with the provider.  However, many branders resist performance-based metrics because the number of users who click through to the co-branded site often depends on the quality of the provider’s service.  Thus, achieving the minimum standards may be out of the brander’s hands.

At the very least, however, the brander can promise to deliver a minimum number of advertising impressions for the provider or the co-branded site within a specified period of time.  From the provider’s point of view, this method may not work if branders try to limit their remedies for failure to reach the target level.  For example, to get certain accounting benefits, the brander may say that its only obligation is to continue running advertisements until the minimum number of ad impressions have been delivered.  Because this remedy effectively means that ad impressions could be delivered in the distant future, even after the provider no longer desires to operate a co-branded site, these types of remedies tend to be unsatisfactory to providers.


Based on the provider’s desire to get maximum promotion from the brander, many providers want the brander to grant the provider some form of exclusivity.  Occasionally the brander will also ask for exclusivity, although this can be analytically confusing when the provider is paying the brander for promotion or when the provider’s business model is predicated on wide “distribution” of its web site.

In any respect, exclusivity is a serious request by either party and requires careful thought and drafting.  For example, CDnow, Inc., an online music seller, recently sued Lycos, claiming that Lycos breached an exclusivity clause restricting Lycos from running advertising for CDnow’s competitors.[xi]  While the CDnow/Lycos agreement enumerated some companies that were deemed CDnow’s competitors, Lycos was further restricted from promoting “any online music store sponsored or promoted by a record label.”[xii]  CDnow asserts that Lycos has been promoting web sites in violation of this broad catch-all definition of competitors.  While the suit is pending, it is a good reminder that significant care must be invested in drafting exclusivity clauses to minimize potential disputes over the scope of the restrictions.

The standard types of exclusivity clauses currently in use include:

Identified Competitors.  A party can enumerate a list of companies with whom the other party cannot enter into specified types of relationships.  The advantage of this approach is that the restricted party can tell with a strong degree of certainty whether a subsequent relationship will or will not violate the restriction.  The disadvantage of this approach is that it is not flexible, and thus, new competitors can emerge over time who are not subject to the restriction.  Sometimes the parties will deal with this by allowing the restricting party to add new competitors unilaterally, but the restricted party cannot allow the restricting party to have unfettered discretion to add parties.

Category.  Either party can be restricted from entering into specified types of relationships with companies that provide certain services or who are in certain industries.  These types of “category” restrictions are very fuzzy and, as evidenced by the CDnow litigation, susceptible to disputes.  Further, it is virtually impossible for the parties to draft a precise but flexible definition of the prohibited functionality or industry that will avoid future disputes in this area.

An example will demonstrate the confusion that can arise from category-based exclusivity.  Imagine that a web site agrees to make the other party the “exclusive retailer of books” on the site, intending to prevent the site from accepting advertising from major online book retailers like Amazon.com and barnesandnoble.com.  Does this restriction prevent the web site from accepting advertisements from etoys.com, a toy retailer that also sells children’s books, or from cooking.com, a retailer of cooking supplies that also sells cookbooks?  As a practical matter, how many mass-market retailers do not sell books of some sort or another?  In other words, perhaps the category restriction of “books” effectively makes the restricting party the exclusive retailer on the site, period—which was probably not the parties’ true intent.  Of course, these types of broad category restrictions can be subject to exceptions, but the list of exceptions, if properly drafted, could take several pages to cover all of the various unexpected ways that book retailers could creep onto the site.  Thus, before the parties decide to use a category-based definition of exclusivity, they should carefully consider whether alternative approaches will achieve satisfactory results with substantially less confusion.

Placement.  Sometimes the parties will not try to enumerate competitors (by identity or by category) at all, but instead will ensure that the promoted party gets premium placement on the brander’s web site, such as being in the upper-left-hand corner, being given more pixels than anyone else, or being the only company promoted on certain pages (to the exclusion of all other advertisers).  While these types of clauses are not technically exclusivity clauses, they can serve as a valuable alternative to the exclusivity provisions discussed above.


Often one of the provider’s key goals in a co-branding relationship is to obtain new users.  To make things easier for users (which increases the likelihood that users will actually use the provider’s services), the brander and provider often agree to automatically exchange data about the users.  These types of data exchanges raise a number of issues.

The parties need to think about exactly what pieces of user information are being exchanged.  If the parties choose to use pre-populated forms, then they can agree to specific lines of information that will be transferred from the brander’s database to the provider’s registration form; otherwise, the brander will often transfer user data directly to the provider’s database.  In addition, the provider may be transferring back referral information to the brander.

To effectuate these exchanges, the parties need to agree on the technology used to transfer this data.  Unfortunately, there is no standard technology to implement these exchanges, meaning (1) the parties may have radically different conceptions about how to do this, and (2) often at least one party, and perhaps both, will have to do custom development work, a serious proposition for most Internet companies who often have hundreds or thousands of engineering tasks that are backlogged and awaiting the attention of strapped engineering departments.  In either case, once the parties agree on the initial procedure for completing the transfer, they may also need to have a set of procedures to deal with changes one party wants to make to their web site or their back-end systems, which would impact the data transfer mechanism. The parties often give little attention in the contract to either the initial transfer mechanism or to procedures for changing the transfer mechanism.  This oversight leaves open wide areas of potential dispute (and possible abuse).

The parties also should consider if they want to “synchronize” their databases and propagate user-submitted changes in shared user information across both databases.  This concern most often arises in the portal context, where users are notorious for giving the portal false personal information, but may be more willing to give truthful information to a co-branded service promoted by the portal if a user is required to be truthful to get the benefits of the service.  In this case, the portal may ask the providers to “synchronize” their respective databases with any changes referrals make to their information.  However, synchronization may not always be desirable, especially if it is possible that users will be changing their information to make it less truthful or if users would legitimately want to have different information on file with the different companies.

In all cases, exchanges of user data require careful attention to the applicable privacy policies and laws.  In the 1997 and 1998 mania to release consumer-friendly privacy policies, many web sites launched privacy policies that restricted their ability to share user information with third parties, even with a co-branded service provider or brander.  Some providers are minimizing this problem by creating a custom privacy policy for the co-branded site (different from the provider’s standard privacy policy for its main web site) indicating that the provider will be receiving data from, and transferring data to, a specified brander.  While this approach may solve the legal problem for the provider, it does introduce a less consumer-friendly privacy policy for the co-branded site.


There are four main types of payment streams in co-branding deals: development fees, exclusivity fees, placement and advertising fees, and fees based on user actions (such as clickthrough fees, bounties or revenue shares).

A.   Development Fees

To implement the co-branded site, the provider usually will have to do some development work.  In some cases, the provider has scripted its site so that it is relatively easy to dynamically place the brander’s branding in specified spots on the provider’s page templates.  In other cases, the provider needs to implement the co-branded site using the brander’s page templates, or the provider needs to make further custom changes to a provider’s standard functionality based on the brander’s specifications.  Additionally, the provider may need to do some work to implement the data transfer mechanism (as discussed above).

In some cases, the provider will be paid development fees by the brander to do the necessary development work.  In other cases, the provider will have to fund this expense out of its own budget as part of the “hidden” costs of entering into the co-branding deal.

Ironically, many branders are now demanding their own development fees even though their development work is often minimal compared to the work invested by the provider.  Because the development fees can be recognized when the development is complete, branders seek development fees to try to accelerate the revenue recognition of payments from providers.

B.   Exclusivity Fees

If a party is subjecting itself to exclusivity restrictions, the other party might compensate the restricted party by paying a fee tied to such restrictions, often called an “exclusivity fee.”  Exclusivity fees can be recognized regularly (monthly) irrespective of the parties’ actual performance under the agreement; therefore, sometimes the parties use exclusivity fees to smooth out the accounting treatment under the agreement.

C.  Placement Fees

Placement fees (sometimes called slotting fees or carriage fees) are used to compensate a party for guaranteed or actual promotions of the other party or the co-branded site.  Placement fees are usually recognized when the actual placements (e.g., banner ad impressions) are delivered.  This can create some uncertainty for both parties.  The brander may have variability in the quantity of ad impressions it can deliver from period to period, resulting in fluctuating revenues.  The provider usually has to recognize expenses corresponding to the delivery of the placements, meaning that a brander who concentrates placement into a single period could cause the provider to have an enormous and unexpected accounting expense during that period.  Therefore, sometimes the parties will establish minimum and maximum placement amounts during a specified period of time to avoid accounting surprises.

D.  Variable Fees Based On User Activity

1.   Clickthrough Fees

Sometimes the provider will pay the brander based on the number of users who “click through” from the brander’s site to the co-branded site.  Branders tend not to prefer receiving clickthrough fees because their earnings will be unpredictable.  Providers also need to be careful if they use clickthrough fees, because the brander may entice referrals to click through to the co-branded site in ways that do not conform to the provider’s economic expectations.

2.   Bounties

Sometimes the provider will pay the brander based on the number of users who actually sign up for the provider’s services through the co-branded site.  To the extent that a provider really seeks to increase its number of registered users, bounties can be an excellent metric for payment from the provider’s perspective.  However, occasionally branders will place such severe restrictions on how providers can use information about the referrals—such as requiring the provider to transfer the referral’s account to a new provider at the end of the agreement—that bounties can cause the provider to wildly overpay for the economic value it is able to derive from referrals.

3.   Advertising Sales

Often the parties create an “inventory” of advertising and promotional opportunities on the co-branded site.  Allocating the inventory and the resulting revenue stream raises some difficult issues.

Control of Inventory.  The parties need to determine who controls the sale of the inventory.  A party may want to control the inventory to ensure that the advertising messages are acceptable; for example, a party may not want competitors’ advertising or “objectionable” advertising placed on the co-branded site.

A party may also wish to control advertising sales to ensure that the revenue stream is maximized.  There are a number of reasons why the selling party may not have proper incentives to maximize the ad sales on the co-branded site.  First, if the selling party has unsold inventory on its own site, it may prefer to direct all ad sales to other sites it operates (where the revenues may not need to be split with others) instead of to the co-branded site.  Second, the selling party might place barter or “house” (self-promoting) ads in the inventory, again undercutting the other party’s expectation that revenues will be maximized from the co-branded pages.

There are a number of alternative solutions to the inventory control problem.  First, a party can remain in control of the inventory, but all ads accepted to be run on the co-branded site would be subject to a rigorous set of standards devised by the other party or subject to the other party’s veto power.  Second, the selling party can guarantee minimum payments (either per page-impression or per month) or minimum performance metrics (such as a minimum cost per thousand impressions (“CPM”) and a minimum percentage of inventory sold (minimum “sell-through”)).  Third, the parties can exercise “joint” control, giving both parties the right to sell ads and veto each other’s actions.  Fourth, if there are multiple advertising spots in the co-branded site’s page templates, the parties can allocate the inventory by letting each party solely control some of the spots.

Ad Serving.  The parties also need to determine who is going to serve the ads sold for the co-branded site.  Usually, ad serving is handled by the party selling ads, but sales and serving do not need to be connected.  In any case, the party serving the ads will likely bear some out-of-pocket expenses which should be reflected in the ad sales split.

4.   Transaction Fees

In many cases, the provider sells goods or services on the co-branded site and the resulting fees are split between the parties.  The parties need to carefully define the revenue stream subject to the split. The brander is concerned that the provider will encourage referrals to complete transactions in locations or through methods where the resulting revenue is not subject to the split.  The provider is concerned that the brander will try to take a share of transactions outside of the streams upon which the provider expects to pay.

In either case, the parties need to precisely define the deductions to be subtracted from the applicable revenue stream.  Usually cost of goods sold is not deducted from the split, but it is usually fair to subtract sales or use tax, shipping costs, and actual returns.  The parties should also consider how the payment system fees, such as credit card fees, will be treated.


A.   Referral Information

As part of the operation of the co-branded site, the provider, and sometimes the brander, will generate information about referrals.  Such information can range from modestly valuable aggregated demographic and psychographic information to extremely valuable personally identifiable information including, in some cases, such sensitive information as credit card numbers and social security numbers.

Properly drafting clauses governing the use and disclosure of referral information remains one of the most vexing problems in co-branding agreements.  There is no industry-standard clause for this situation, so each clause requires, but rarely receives, careful and individual consideration.

Sometimes one party will try to assert sole “ownership” over the referral information.  In almost every circumstance, this is not the optimal result, since the parties almost always need to use, and possibly disclose, the referral information as part of their normal business operation and as part of the relationship.

Sometimes the parties will consider “joint ownership” of the referral information, a problematic phrase because there is no intellectual property right in the referral information to which joint ownership could apply.  It does not make sense to consider jointly owning the copyrights of the referral information, since the referral information is almost always just facts and therefore not subject to copyright protection.[xiii]  Furthermore, under copyright law, joint owners have certain duties to each other,[xiv] such as a duty to account and possibly a duty to avoid waste, which the parties rarely intend to implicate.  The referral information certainly may be the trade secret of both parties, but the proper way to assert ownership over the trade secret is to establish a set of use and disclosure restrictions on the other party. The declaration of “joint ownership” is not sufficient to effectuate a trade secret license and often obscures the need to be explicit about specific use and disclosure restrictions.

Often, after careful consideration of the real economic and competitive risks, the parties realize that they do not need to aggressively restrict the other party’s use and disclosure of referral information.  In many cases, all the parties really need is to restrict the other party from using the referral information in a way that benefits the other party’s competitors, such as targeting referrals for competitors’ advertisements.  Otherwise, both parties may be willing to let the other party freely use and disclose referral information—subject, of course, to the party’s privacy policies and applicable law.

B.   Impressions

Third party impression auditors such as Media Metrix[xv] have become a powerful force on the Internet, and thus web sites are doing what they can to improve their Media Metrix ranking.  Because the co-branded site creates an inventory of impressions that will count towards Media Metrix rankings, the parties are keenly interested in who will get to count the impressions towards their ranking. Usually the party whose domain name is used for the co-branded site “owns” the resulting impressions for Media Metrix purposes.  Thus, the parties may negotiate over whose domain name is used in connection with the co-branded site.  However, if the parties submit the proper documentation, Media Metrix will count impressions towards a specified company’s ranking even if their domain name is not being used, so the parties can by contract assign “ownership” of these impressions.

C.  Trademarks

Because trademark issues are critical to the success of co-branding deals, close attention is warranted to all aspects of trademarks.

The brander’s license of its trademarks raises few unique issues.  As in other situations, the brander must establish mechanisms to ensure quality control and typically will want to address the other types of restrictions traditional in standard trademark licenses.  If there are personality or character rights involved, these require special attention because of the unique and difficult intellectual property rights they raise.

The parties will often want to discuss the domain name early in order to address issues such as the possibility of the combination mark and Media Metrix reach.  Occasionally, the parties will create a unique domain name or trademark for the co-branded site, in which case the provider may want to restrict the brander’s right to use this unique mark post-termination.  Of course, if the unique mark is also a combination mark, this issue will be governed by the license to the combination mark.

Finally, franchise law could create havoc if it is applied to co-branding relationships.  Each state has its own set of franchise laws, and franchisors usually must follow specific procedures before offering franchises in the state.  The factors for determining whether a relationship is a franchise vary from state to state, but usually include several elements, including a trademark license, an up front fee, a marketing plan prescribed by the franchisor, and a “community of interest” in marketing the product.[xvi]  Frequently a brander will meet a number of these factors, so care and consideration must be given to the structure of the relationship to destroy as many of the franchise elements as possible.  Unfortunately for branders, many aspects of franchise law cannot be waived contractually,[xvii] so a statement by the provider expressly waiving the application of franchise law may not provide adequate protection.  Furthermore, usually franchisees cannot be terminated except for cause, even if the agreement expires by its terms.[xviii] As a result of these unexpected and often unfortunate results, the party that would be characterized as a franchisor (typically the brander) has strong incentives to avoid the application of franchise law. Fortunately, we have yet to see any litigation asserting that a co-branding site is a franchise.[xix]


The brander wants referrals to have a good experience with the co-branded site to maintain the goodwill associated with the brander.  Therefore, the brander may want to require that the provider adhere to minimum service levels with respect to the co-branded site.  Providers usually aggressively resist being subject to service levels, considering they are also motivated to provide a good experience to referrals.  Further, to the extent that providers are paying branders for promotion, the provider can be left in the unusual situation of paying the brander for the privilege of providing minimum service levels to the referrals.  Therefore, not every co-branding agreement has service levels, and frequently providers will water down the remedies available to the brander if the service levels are not met.

Examples of the types of service levels commonly addressed in co-branding agreements include:

Uptime.  Uptime refers to the percentage of time that the co-branded site is available to users.  Ideally, the co-branded site will be available 24 hours a day, 7 days a week without interruption. However, usually the parties will agree that the site will be up some lesser percentage of time (e.g., 99%).

Server Speed.  Slow servers can be as bad as down servers, so the parties may agree on a minimum time it takes for servers to respond to referral requests.

Throughput.  The size of the data pipelines connecting a provider’s servers to the Internet can be another bottleneck, so the parties may specify a minimum size of the data pipeline.

Error Correction.  The parties may agree on a procedure or time period for provider to fix errors in the software used to operate the co-branded site.

Security.  The parties may agree on steps that the provider will take to keep the co-branded site or its associated data secure and free from unauthorized intrusion or hacking.

Browser Configuration.  Because different browsers process HTML differently, the co-branded site can look different to referrals, depending on their browser.  Also, referrals using old browser versions may be limited in accessing some of the more advanced technological features of the site.  Additionally, some sites require one or more third party plug-ins for the site to operate properly.  Thus, the parties may agree on what browsers and plug-ins the co-branded site will require or support.

Customer Support.  The brander may want customer and technical support inquiries submitted to the provider to be acknowledged or resolved within a specified period of time, or may establish a procedure to bring the issue to the brander’s attention.


In a few short years, co-branding deals have become an integral part of the Internet’s business infrastructure.  As a result, we are beginning to develop a more thorough understanding of the issues associated with co-branding deals, which is slowly improving the efficiency with which these transactions can be done.  Unfortunately, in light of the continued confusion manifest in co-branding agreements—and the all-too-frequent abuse of leverage a party may exert when negotiating a co-branding agreement—we still have much work to do before the drafting and negotiation of co-branding agreements becomes efficient.  However, with a deeper understanding of the real issues presented by co-branding agreements (and a concomitant understanding of the trivial or insignificant issues), co-branding agreements can become an even more valuable and useful tool for companies trying to build businesses on the Internet.

© 1997-1999 Cooley Godward LLP.

*  Examples of co-branding agreements discussed in this article are available online at http://www.scu.edu/techlaw.

 Eric Goldman (formerly Eric Schlachter) is an attorney practicing cyberspace law with Cooley Godward LLP in Palo Alto, CA.  He is also an adjunct professor of Cyberspace Law at Santa Clara University School of Law.  Cooley Godward’s web page is located at http://www.cooley.com, and Eric’s personal home page is located at http://members.theglobe.com/ericgoldman/.  Eric can be reached at egoldman@cooley.com. The authors thank Laura Lerner for her assistance with this article and generally.

†† Candice Lee is a J.D. candidate (expected May, 2000) at Santa Clara University School of Law. Candice was a law clerk at Cooley Godward LLP in Palo Alto, CA during the writing of this article. Candice can be reached at  c3lee@scu.edu.

      [i].   The look and feel of the brander’s web site includes a number of features that users identify with the brander, including the brander’s trademarks.

      [ii].   See AOL <http://aol.com>; Yahoo <http://www.yahoo.com>; Excite <http://www.excite.com>; Lycos  <http://www.lycos.com>; the Go Network <http://www.go.com>; and Snap! <http://www.snap.com> for some of the most visible portals on the Internet.

     [iii].   A number of providers have emerged to service the needs of web sites exclusively on a co-branded basis. For example, InfoSpace.com provides such services as maps and directions, government directories, shopping directories, business and people finders, weather and real-time stock quotes on a co-branded basis, incorporating the “look and feel” of the brander’s pages. See Infospace (visited Nov. 3, 1999) <http://in-100.infospace.com/info/pbi/cobrand.htm>.

     [iv].   See Excite, (visited Nov. 3, 1999) <http://www.excite.com>.  The co-branded page is clearly marked with the Quicken.com logo and contains the term “quicken” in the URL (<http://www.quicken.excite.com>).

     [v].   See discussion infra Part IV., entitled “The Brander’s Promotion.”

     [vi].   For example, when multiple versions of the same content are easily accessible on the Web, each version might compete with each other for users’ attention or money.

    [vii].   Various legislative efforts to provide statutory protection to databases have been proposed over the years, such as H.R. 354, 106th Cong. (1999) (The most recent incarnation of the Collections of Information Antipiracy Act, proposing an amendment to Title 17, United States Code, to include protections against misappropriation of “collections of information”).

   [viii].   See Feist Publications v. Rural Telephone Serv., 499 U.S. 340 (1991). Factual databases may be protectable under U.S. Copyright Law as compilations, 17 U.S.C. § 101 (1994), but this does not provide any protection for individual factual items.  Further, in some cases, even the selection, arrangement and coordination of the database is not copyrightable.  See Feist, 499 U.S. at 358.

     [ix].   Uniform Trade Secret Act, § 1(4)(ii), 14 U.L.A. 437, 438 (1990).

     [x].   A “cookie” is a piece of information placed on the user’s hard drive by the web site.

     [xi].   See Shannon P. Duffy, Internet Agreement Leads to $1 Million Suit, THE LEGAL INTELLIGENCER, July 20, 1999, at 3.

    [xii].   Id.

   [xiii].   See supra note viii.

   [xiv].   1 Melvin B. Nimmer, NIMMER ON COPYRIGHT, § 6.10-§ 6.12 (1999).

    [xv].   See Media Metrix (visited Nov. 3, 1999) <http://www.mediametrix.com>.

   [xvi].   See Thomas M. Pitegoff, The Inadvertent CyberFranchisor, 3 CYBER. LAW. 2, 2-3 (Apr. 1998).

  [xvii].   See, e.g., Wright-Moore Corp. v. Ricoh Corp., 908 F.2d 128 (7th Cir. 1990) (articulating Indiana’s strong policy against allowing parties to contract out of the protections provided by its franchise laws); see also Jerome-Duncan, infra note xix, at 909 (under Michigan franchise law certain provisions will be considered void and unenforceable if contained in any franchise documents).

[xviii].   See generally, Instructional Sys., Inc. v. Computer Curriculum Corp., 130 N.J. 324, 614 A.2d 124 (1992).

   [xix].   Consider, however, Jerome-Duncan, Inc. v. Auto-by-Tel, 176 F.3d 904 (6th Cir. 1999) (holding that a geographically-exclusive Internet subscription agreement between a Michigan car dealership and an online car dealership referral service was not a franchise agreement, even though the dealership was required to display the service’s mark and adhere to a number of standards promulgated by the service); and Computer Currents Publishing v. Jaye Communications, Inc., 968 F. Supp. 684 (N.D. Ga. 1997) (the court assumed the existence of a franchise where a trademark licensee had the right to publish online a local version of the licensor’s newsletter if the licensee followed the format, practices, and standards established by the licensor, and agreed to pay a license fee based in part on a percentage of the licensee’s gross receipts).