Private Label and Co-Branding Deals:
New Deal-Making Paradigms on the Internet
by Eric Goldman
The Web has created a new way for information providers to interact with their users. Because linking on the Web can create an architecture of sites that appear seamless to users, from the userís perspective it is often unimportant which server they are accessing when moving around the Web.
As a result, relationships are being struck where one site (the "brander") will place its branding on the content or functionality of another site (the "provider"), extending the perceived boundaries of the website. These types of relationships can take a number of forms. Sometimes the provider will have no branding on the pages it provides; this type of arrangement is called a "private label" deal. Other times the provider and brander will each have their brands on the site; this type of arrangement is called a "co-branding" deal. Although this article focuses on private label deals, much of the analysis applies to co-branding deals as well.
Identifying private label and co-branding deals on the Web today is a little tricky, in part because it is hard to know whose servers the content is on when visiting sites. A few companies, such as Vicinity (http://www.vicinity.com/) and InfoSpace (http://www.infospace.com/), are in the business of providing private label and co-branding services for their content databases. Specific examples of implementation of co-branding relationships can be seen at sites like Yahoo Maps (http://maps.yahoo.com/yahoo/) provided by Vicinity and HotBot (http://www.hotbot.com/), a co-branded site between HotWired and Inktomi. There are countless other incidents of these types of relationships, many of which are not clearly apparent to users.
While private label and the co-branding agreements raise a number of new issues on the Web, there are numerous existing analogies. For example, most supermarkets and many other retail stores will carry goods manufactured by a third party but carrying the retail outletís brands. This type of arrangement has historically been called a private label deal. (Compare the classic Original Equipment Manufacturer (OEM) relationship, where the branding party sells the goods through a chain of distribution).
Private label and co-branding deals raise interesting issues in part because they expose a dichotomy between the business practices and the legal implications. From a business perspective, the parties to the deal are often less concerned about whose servers content sits onóas long as the intended audience can reach the pages and branding is appropriate, the rest of the aspects of the relationship are just "details." However, as discussed in this Article, this decision may have a number of legal ramifications.
At the heart of the relationship, a private label deal is primarily a trademark license from the brander to the provider and an agreement for the provider to provide services and perhaps access to intellectual property to the world (or occasionally only to traffic generated by the provider). In practice, this relatively simplistic description masks a number of complexities that must be addressed for the parties to achieve their objectives.
Why Do It?
There are a number of reasons for a brander to enter into a private label deal. First, like any other outsourcing arrangement, the brander can take advantage of the providerís expertise or economies of scale. For example, the provider may be in the business of building and maintaining a database that the brander does not wish to try to replicate (i.e., it is cheaper to buy than to build). Alternatively, the provider may have superior software tools, and the brander simply cannot cost-effectively or time-effectively develop competing tools.
Second, content is critical on the Internet. A private label deal allows the brander to appear to have a larger website, or to have a more extensive set of features, than it really has.
There are also several advantages from the providerís point of view.
First, a private label relationship takes the place of a licensing arrangement. Rather than using its intellectual property in only one channelóits own websiteóthe provider can "recycle" the content. A private label arrangement allows the provider to establish parallel channels, thereby potentially getting multiple revenue streams from the same content or functionality instead of one. Of course there is always the risk of channel conflict or cannibalization, but if the provider strikes relationships with different brands with access to significantly different types of consumers, this risk may be worth bearing.
For a perfect example of multi-channel recycling, consider the business practices of Vicinity <http://www.vicinity.com/>. Vicinity provides a database of maps and related materials (such as driving instructions). Vicinity has successfully private labeled its content to search engines, travel sites, yellow page directories, newspapers, retailers and others. See <http://www.vicinity.com/vicinity/publisher_cust.html and http://www.vicinity.com/vicinity/corporate_cust.html>. Each of these customers is presumably getting the same database, but the relationship stills makes sense because each site isóin usersí mindsóadding valuable add-on content/services to its existing offerings. This is true even though a user could go directly to Vicinity or elsewhere to get access to the same content.
Second, the provider can effectuate this "license" without actually having to provide a physical copy of its content. This has numerous benefits for the provider, not the least of which is that the provider has fewer issues to worry about under intellectual property law. For example, consider a publicly accessible database of facts, such as a directory of phone numbers. These types of databases are subject to little if any protection under US intellectual property laws. Databases of facts are unlikely to be covered by copyright, which does not protect facts; at best, the database will be subject to a thin compilation copyright, which may be easy to circumvent. Because it will be made available to the public, the database cannot be treated as a trade secret. Therefore, historically the database ownerís only option was to license the database using a set of restrictions synthetically created by contract and lacking any intellectual property protection to act as a backstop or as a more powerful point of leverage in case of breach.
Further, because there are no intellectual property rights underlying the content, if the content escaped the control of a contract licensee, the licensor had no power to stop downstream recipients from further "infringement." As a result, the licensorís asset was constantly in jeopardy.
Using private label deals, the provider can offer all of the benefits of the content without circulating a copy of the content to third parties. Because the provider can provide access to a single copy of the database over the Web, the provider can much more easily and reliably use technology to control the content rather than contracts and law.
Additionally, by controlling the number of copies of the database circulating in the world, it is much easier to ensure that all copies of the database are "in sync" and current. This can be a significant logistical consideration for complex, time-sensitive databases such as reservations databases.
A final benefit of these types of arrangements occurs in the co-branding context or when the provider is able to insert some branding on the branderís pages. By being exposed to new users, the provider may see increased traffic on its own site.
Making Money from the Relationship
There are two primary flows of revenues from the brander to the provider. First, there are set-up and maintenance fees analogous to hosting fees. Second, there can be a revenue share or other royalty-like relationship.
a. ††††††† Set-up and maintenance fees
The set-up and maintenance fees represent one of the first potential pitfalls for providers. At its core, the set-up and maintenance of the pages is a hosting relationship. Hosting relationships are notoriously difficult to make a profit from on a large scale. Although frequently these set-up and maintenance costs are priced on a "package" basis, these services are effectively priced either on an hourly basis or on a cost-plus basis. The provider needs to carefully circumscribe its obligations or to price the obligations at a high enough level, or the provider may find that set-up and maintenance creates a drag on profits from the relationship.
While it may sound obvious that set-up and maintenance costs be passed through to the brander, in practice many providers trivialize these costs ("oh, that wonít take much time to do!"). As a result, a provider often offers services at no cost as an inducement to "get the upside" that presumably will flow from the royalty/revenue share. Clearly this decision can be justified at times, but generally providing services upfront in anticipation of downstream royalties can have serious implications for cash flow, accounting profitability (i.e., costs are incurred before revenue is recognized), and business risk being taken on from the deal.
The provider also needs to be careful about deriving the bulk of the revenues from the set-up and maintenance costs without upside potential. This clearly turns the provider into a web host, and as suggested earlier, web hosts have a very difficult time obtaining significant margins. While web hosts can find profitable niches, this business model needs to be carefully considered.
b. ††††††† Revenue SharingóAdvertising
There are a number of ways to structure the royalty or revenue sharing relationship. The most popular way is to sell banner advertisements on the branded pages and share revenues from those advertisements. Banner advertisements, or similar advertiser-driven relationships, raise a few difficult issues.
First, there is the issue of who controls the advertising. The brander may wish to control the advertising on the branded pages to ensure that the advertising messages are acceptableóthe brander may not want competitors advertising on the site, and the brander may not want "objectionable" advertising placed on the site because of the potential dilution effect on the branderís trademarks and branding. Often (although not exclusively), the brander may also be the party with superior access to potential advertisers and therefore be in a better position to procure advertising.
The provider may wish to control advertising to ensure that maximum dollars are achieved. This is critical, of course, because most providers expect to make their profits from the relationship from the revenue split, and as discussed above, often will make little or no money from the set-up and maintenance aspects despite incurring those upfront costs.
Although it may sound odd that a revenue split does not properly create an incentive for branders to maximize revenues from advertising, there are a number of reasons why branders may not do so. First, if the brander has unsold page impression inventory on its site (and fewóif anyósites do not), the brander may not need to obtain any advertising revenues from the branded pages. Rather, if the branded pages enhance overall traffic to the branderís site, then the brander in fact may be able to see increased revenues from the rest of its site. In this situation, unless the provider gets a share of revenues generated from the branderís site, the brander will not obtain a fair share of revenues merely by splitting the revenues from the branded pages.
Second, if the brander controls the advertising, the brander could place barter ads (i.e., advertisements that are given space freely in exchange for the branderís banner advertisements being placed freely on the advertiserís site) or its own banner advertisements in the inventory, again undercutting the providerís expectation that revenues will be maximized from the branded pages.
There are a number of alternative solutions to these problems. First, the provider can remain in control of the advertisement, subject to a rigorous set of standards provided by the brander or subject to branderís veto power, which will not be unreasonably exercised. Second, the parties can outsource the advertising to an advertising representative firm, effectively letting this "neutral" third party take some control over the problem. Third, the brander can take control but guarantee the provider minimum payments (either per page-impression or per month). Fourth, the parties can exercise "joint" control, giving each party with veto power over the other partyís actions. Finally, the provider can let the brander control and establish a business model that is not predicated on maximizing advertising revenues, such as the other revenue sharing models discussed below.
c. ††††††† Revenue SharingóTransactions
An alternative business model is for the parties to sell goods or services on the branded pages, thereby letting the parties share the revenues flowing from these transactions. Obviously, there are a limitless number of types of transactions that the parties might do. For example, a reservations database can allow users to complete the transaction on the site. Another example is <http://www.animatedgreetings.com/>, a co-branded site between American Greetings and Pineapple, where users can buy "animated greetings" for $1.95 per recipient and the parties can share the revenues from these purchases.
d. ††††††† Revenue SharingóProviding Traffic to Provider
A completely different business model can arise in a co-branding deal when the primary purpose of the branded pages is to transfer users from the branderís site to the providerís site. Typically, then, the co-branded pages will contain numerous links to the providerís site, and often there will be some sort of "sample" or freebie to get the users to consider continuing their surfing to the providerís site.
In this situation, the brander will want to share in the revenues the provider generates from these users. The provider might generate revenues from users based on page impression-based advertising, which can be accounted for by tracking the users who come from the branded pages onto providerís site and counting the number of page impressions generated by these users. Alternatively, the provider may offer transactions from its website, and the brander could share in those revenues.
In these situations, then, it may be the brander who is concerned about the providerís incentives. The brander will be letting the provider use its valuable trademarks, and perhaps more importantly, the brander will be losing its traffic to the providerís site. Therefore, the brander must ensure that adequate mechanisms are in place to compensate the brander.
In most of these cases, the brander will want to "tag" its users so that their incremental benefits to provider can be measured. There are three primary ways to do this. First, the parties can try to use IP address analysis, but this remains a crude science and is rarely helpful.
Second, the parties can use registration, whereby the users being transferred are forced to register. These users can then be tracked by requiring subsequent log-ins, by issuing the users a digital certificate, or by loading a token into a cookie (discussed below). Registration is rarely a good result because of user antipathy towards such impediments.
Third, the parties can load a token into the userís cookie without registration. This method is the least intrusive to users, but the brander should recognize that the cookie method is not foolproof. First, users might refuse the token. Second, the user might edit the cookie file. Third, the user might switch browsers. Fourth, the user may be using a browser that does not support cookies. Finally, the parties need to decide if the token should expire per session or on some other basis, depending on whether the branderís contribution is best measured by instant response or any response. If the token expires per session, of course, the brander may lose some users who return to the providerís site after the token has expired.
Tagging and transferring users creates some difficult issues about "ownership" of those users, discussed below.
e. ††††††† Defining Net Revenues
In all cases of revenue sharing, the parties will need to define whether gross or net revenue is being shared. Defining net revenue in this context is not much more difficult than in other contexts, but there are a few issues that need to be addressed.
In the case of advertising revenues, the key issue is how the advertising representativeís fee will be split. Currently, advertising representatives are taking up to 50% of the revenues from the ads they are placing, so effectively this means the parties are in a three way split. To avoid such heavy fees, the parties might let the party that directly sells the advertising (i.e., if the brander or provider can sell the advertising without relying on the advertising representative firm) keep a larger percentage of the revenues. The parties will also want to deal with the issue of how unsold advertising is allocated, and if a party to the deal is permitted to use the unsold inventory for a fee.
In the case of transactions, it is usually fair to subtract sales or use tax, shipping costs and actual returns from net revenues. Since almost all online sales will be made pursuant to a payment system, the parties should also consider how the payment system fees, such as credit card or First Virtual transaction-based fees, will be treated.
Intellectual Property Ownership and Licenses
Because the parties are in a close relationship, the parties need to carefully consider a number of issues regarding ownership of various aspects of the branded pages.
a. ††††††† Owning "Customers"
It is very common for the parties to think about one party or the other "owning" customers. Of course, customer lists are valid subject matter for trade secret protection, and in this context the parties could structure this aspect of the relationship as a trade secret license. In fact, the confidential information is not just the list of customers, but all information gleaned about the customersóincluding their demographics, their psychographics and any information provided via registration. The number of page impressions being delivered, the rate being charged advertisers and the number of transactions being consummated is also information that could harm one or both parties if released and therefore is extremely sensitive. It may be that this information should be categorized as the confidential information of both parties and subject to use and disclosure restrictions on both parties. At minimum, it is likely that this information should be the confidential information of one party.
Because the server logs will contain most of the information one would need to know to deduce proprietary customer information, the server logs must be made confidential information of at least one party. Both parties presumably will want either access to the server logs, or an analysis of the server logs.
One particular use restriction to consider is the spamming of users to the extent that email addresses are collected. Each party will most certainly not want this database of email addresses being available to competitors. Furthermore, because of the negative connotations of spamming, and the associated ill-will directed at sites that collect email addresses that are used for spamming, each party may want to restrict the other partyís rights to send email to the addresses for any reason.
It makes little sense to address copyright rights in the database of information regarding customers, although sometimes the parties will be confused about this.
b. ††††††† Ownership of Siteís "Look and Feel"
To the extent that the parties will jointly design the look and feel of the branded site (which will incorporate existing branding from the brander), the parties will need to determine the ownership of this look and feel. While there are some trademark aspects to the ownership of the look and feel, there are likely to be numerous elements of the look and feel which may receive copyright protection, and the entire look and feel could be subject to its own separate copyright.
Frequently the parties will initially intend that the parties "jointly" own all copyrightable elements. Joint ownership of copyrights is usually a bad idea, because the joint owners will have numerous duties to each other that are not clearly defined under U.S. copyright law. Once the joint ownership issues are analyzed, the parties rarely will conclude that it meets their needs.
If the parties do not want joint ownership, they should be careful to avoid their efforts being classified as "joint works of authorship" under U.S. copyright law. The status of joint works of authorship can be easily destroyed in the partiesí contract by expressly saying that no joint works of authorship are intended.
In all cases, the parties need to define what will happen to the branded pages following termination of the relationship. There are no standard resolutions to this matter, but usually a satisfactory resolution can be reached if discussed by the parties as part of the contract negotiations.
Finally, if the provider is developing some or all aspects of the look and feel, the provider usually will need a license to create a derivative work (perhaps only in the form of the compilation) of the branderís materials. If the provider will be the owner of the look and feel and if the provider needed a license to create the derivative work, such license will need to continue following termination if the provider is intended to have the right to use the derivative work thereafter.
c. ††††††† Trademarks
Trademark issues are critical to the success of the private label and co-branding deals. Close attention is warranted to all aspects of the trademark issues.
The branderís license to the provider of the branderís trademarks raises few unique issues. As in other situations, the brander must establish mechanisms to ensure quality control. If there are personality or character rights involved, these require special attention.
The brander should consider to what extent the branded pages should permit the use of providerís trademarks. This is likely to be a material element of the business relationship and therefore is rarely overlooked. However, there are a few special issues that can arise. First, if the brander does not intend to allow the provider to use its trademarks on the page, does this further include a prohibition on "credits" or other acknowledgments of effort?
Second, the domain name raises its own issues. If the brander wants to completely make the branded pages appear to be part of the branderís site, the branderís domain name must be used, and the provider should obtain a trademark license to use this domain name. If the branderís domain name is not used, the agreement has effectively become a co-branding agreement and the brander should assume that all users will realize that the provider is involved. Occasionally the parties will procure a new domain name for the branded pages; the domain name should be treated as a trademark and its ownership addressed accordingly.
If providerís trademarks are permitted on the branded pages, the brander needs to decide if combination marks can be formed. In fact, often times combination marks are intended to be formed, and the parties must address the resultant rights and obligations. In particular, the parties should consider who will have the right to register the combination marks and what rights the respective parties will have to use the combination marks following termination of the relationship.
Finally, franchise law is a theoretical but nettlesome issue in these relationships. Each state has its own set of franchise laws, and franchisors usually must follow certain procedures before offering franchises in the state. Furthermore, usually franchisees cannot be terminated except for cause, even if the agreement expires by its terms. As a result of these unexpected and often unfortunate results, the party that would be characterized as a franchisor has strong incentives to avoid the application of franchise law. The factors for determining whether a relationship is a franchise vary from state to state, but usually there are several elements, including a trademark license, an upfront fee, a marketing plan prescribed by the franchisor, and a "community of interest" in marketing the product. Frequently the brander will meet a number of these factors, so care and consideration must be given to the structure of the relationship to ensure that at least one of the requisite factors is not met. Unfortunately for branders, many aspects of franchise law cannot be waived contractually, so a statement by the provider expressly waiving the application of franchise law may not prove to be adequate protection.
The Service Aspect of the Relationship
Because the brander is effectively outsourcing a portion of the branderís website to provider, the brander will want to impose all of the duties on the provider that the brander would impose on any web host. In all cases, the goal is to provide a good experience for users so they will keep coming back. While usually both parties interests are in alignment on this, the brander may have more at stake given that the userís failure to have a good experience will be associated with branderís trademarks and the brander will suffer that loss of good will in connection with the providerís content or software.
First, the brander will want to ensure that the branded pages are available 24 hours a day, 7 days a week without interruption. In particular, the brander will want to institute requirements to keep the branded pages from going offline due to a service interruption (such as a power interruption or the providerís Internet connection failing).
Second, the brander will want some warranties for the providerís services. If the provider is providing functionality (such as a chat engine or a search engine), the brander will want to ensure that the functionality works properly. If the provider is providing access to content, the provider will want to ensure that the content is reasonably accurate and current. In both cases the brander will want some protection from third party claims, such as infringement of third party intellectual property rights or misappropriation of rights of publicity or privacy. The brander should also impose some parameters on the content added by the provider, and in particular prevent the provider from distributing viruses or other harmful code in connection with the providerís content or software.
Third, the brander will want to impose requirements on the provider to keep latency times low. This means that the provider will not only need adequate bandwidth connection to the Internet, but the provider will also need to provide upgrades to the routers and servers necessary to provide a fast connection.
Fourth, the brander will want to consider the security methods used by provider. Among the potential concerns are changes made by hackers to the branded pages, spoof sites, and the storage of sensitive materials (such as databases of customersí credit card numbers) behind adequately secured firewalls.
Finally, the brander should consider its stance towards caching and indexing. Both caching and indexing may make it difficult to destroy evidence of the branded pageís existence at the end of the relationship, and therefore the brander should consider if this is a problem and, if so, the best method of resolution.
We are evolving towards a more sophisticated level of deal-making on the Web. In 1995 and 1996, many companies eager to get on the Web launched their sites with ambitious features and functions that prove costly and burdensome to maintain. As these companies have gained more insights into their business, they have realized that outsourcing these obligations make economic sense. At the same time, the seamless nature of the Web permits an outsourcing relationship to be virtually invisible to end users. In some circumstances, large portions of the branderís site will be outsourced; in other cases, the relationship will deal with just a small aspect of a larger project. In the future, it is likely that parties will be both providers and branders in the same agreement. In any case, it is likely that private label and co-branding deals will become ubiquitous on the Web and warrant careful consideration.