Thirty dollars a day bought the Harvard College Democrats advertising on thefacebook.com in spring 2004. Their message promoted a trip to New Hampshire. This small booking provides a useful starting point for Facebook’s marketing history: a particular advertiser wanted to reach an identifiable audience for a specific purpose. The sophisticated global advertising system associated with the company in later years was still to develop (Grynbaum, 2004).
The year 2004 marks the network’s beginning. Facebook Ads, combining business Pages with socially connected advertising, was publicly introduced in November 2007. Early advertisements, broader membership and changes in information distribution came between these dates. Treating them as separate developments makes it possible to explain how an audience became the foundation of a commercial intermediary.
The central question is therefore how an environment for recognising people and relationships became an infrastructure for marketing. Reach was part of the answer, but so were the provenance of information, the visibility of connections and the conditions governing distribution. A commercial message could appear alongside personal updates or carry a reference to someone the recipient knew. These arrangements changed the circumstances of advertising without making every observable social reaction evidence of business success.
February 2004: A directory inside an existing community
Facebook launched on 4 February 2004. Alan Tabak’s contemporary report in The Harvard Crimson described profiles, searches for fellow students and registration using Harvard email addresses. More than 650 registrations were reported at that early point, a figure attributed to Zuckerberg rather than a later measure of the company’s overall audience (Tabak, 2004; Meta, n.d.).
This setting matters economically. University affiliation initially limited participation while helping people interpret the information they encountered. References to classes and residential houses could mean something within everyday campus life. The digital directory drew value from an environment that already existed beyond its interface.
The resulting audience differed from one assembled solely around a shared media product. Participants could know each other elsewhere, encounter each other later or recognise common affiliations. Online information extended existing opportunities to locate people and understand relationships. Its significance depended partly on connections that the website had not itself created.
For subsequent advertising, this social density was a resource rather than a finished product. Running a useful directory and operating a scalable advertising service require different arrangements. Booking, delivery, evaluation and participant expectations must work together. The history of Facebook’s audience consequently begins with one problem, while the history of its commercial offering begins with another: how to make that audience accessible to advertisers on workable terms.
Early advertising: Test appearances and paid local bookings
Michael Grynbaum reported in May 2004 that temporary advertisements for AT&T Wireless, America Online and Monster.com had appeared during April. Facebook spokesperson Chris Hughes said these tests were unpaid. Harvard organisations, including the Harvard Bartending Course and Veritas Records, also advertised on the website (Grynbaum, 2004).
The distinction prevents an attractive but misleading interpretation. A familiar brand appearing on a young service establishes that an advertisement existed. It does not necessarily establish a paying client, recurring revenue or a lasting commercial partnership. Tests can explore possible demand or working relationships before either side commits to a continuing arrangement.
The local bookings reveal a more immediate match between audience and purpose. A student organisation could reach people within a relevant environment without requiring the complete machinery of later social advertising. Concentrating an appropriate group in one accessible place already created a potential commercial opportunity. The value of that opportunity depended on the advertiser’s objective and the circumstances of use.
Different scales of activity therefore coexisted. Large brands could examine an unfamiliar medium while organisations close to its original community used it for practical communication. An account built entirely around famous corporate names would miss those smaller applications. They show how advertising entered the ordinary life of the service, through particular messages and bookings rather than a single comprehensive invention.
Targeting: An advertised capability was not proof of implementation
According to Grynbaum, the advertising material offered selection using sixteen characteristics, including college and age. Hughes simultaneously explained that targeted advertisements had not yet run (Grynbaum, 2004). The contemporary evidence thus distinguishes a capability being offered from a capability demonstrably used.
That distinction is essential when reconstructing the history of data-driven marketing. Information may be available without being applied in every commercially imaginable way. A supplier may describe what it hopes to sell before advertisers begin using the proposed service. Retrospective accounts should preserve these stages instead of collapsing them into a fully developed system.
Nor does the ability to select a characteristic demonstrate the quality of a target audience. A shared attribute can be relevant to one offer and almost irrelevant to another. Advertisers still need a reason to expect a particular message to matter to the people selected. Technical selection makes a decision executable; it does not supply the commercial reasoning behind it.
An older marketing task was therefore being translated into a different operating environment. Identifying suitable customers and occasions remained necessary. The potential novelty lay in expressing aspects of that reasoning through platform characteristics and delivery rules. Claiming that Facebook already possessed its complete later targeting system in 2004 would enlarge the evidence beyond what the contemporary report supports.
Earlier networks and parallel possibilities
The historical account by danah boyd and Nicole Ellison identifies SixDegrees in 1997, Friendster in 2002 and MySpace in 2003. Profiles, articulated relationships and access to other people’s connections were already part of the development of social network sites before Facebook launched (boyd and Ellison, 2007).
This comparison changes how innovation should be assessed. Being the first to introduce an isolated feature is only one possible contribution. Combining familiar elements in a particular institutional setting, attracting sustained participation and building a commercial service around them can also be consequential. Later prominence cannot substitute for examining those processes.
Networks can organise participation around professional roles, musical interests, personal acquaintances or institutional membership. These settings create different expectations and commercial occasions. Similar interface features do not make the environments interchangeable, because the reasons people participate and the meanings they attach to connections can differ.
Facebook therefore belongs within a history of competing and parallel approaches. Explaining its contribution does not require treating every other service as a preliminary step towards an inevitable winner. An international history of marketing is better served by asking what different organisations made possible, how people used those possibilities and which commercial arrangements developed around them.
South Korea: Cyworld and the commercial significance of self-presentation
Kyung-Hee Kim and Haejin Yun trace Cyworld’s beginnings to 1999 and its development into a social network to 2001. Personalised minihomepages, virtual rooms and purchasable decorations formed part of its environment. Their study, published in 2007, used detailed interviews with forty-nine users to examine negotiations over intimacy, autonomy and disclosure (Kim and Yun, 2007).
Cyworld broadens the comparison beyond American platforms. Digital self-presentation could support commercial activity through arrangements other than advertising alone. Designing a personal environment and giving it meaning within relationships created opportunities for paid participation. The commercial value of social life did not have to take the same institutional form everywhere.
This is not a claim that all South Korean participants shared identical expectations. A qualitative study is valuable precisely because it can reveal tensions and differing interpretations. Design, cultural conventions and particular relationships interact. A connection represented through a technical interface may be understood differently depending on the surrounding social circumstances.
For Facebook’s history, the implication is specific. Neither digital identity nor the commercial significance of personal relationships began at Harvard in 2004. Facebook developed a particular combination of these elements. Its subsequent scale should not obscure other models that preceded it or developed alongside it. Comparison explains its position more accurately than a claim of universal origin.
News Feed in 2006: Distribution became a separate service
Meta’s corporate chronology dates the introduction of News Feed to 5 September 2006 (Meta, n.d.). The important historical distinction is between information being available somewhere and updates being gathered into a continuing overview. A profile that someone might visit and an update delivered within an overview organise access differently.
Visiting individual destinations requires people to make choices before seeing their contents. A feed performs part of that organising work. Its arrangement helps determine which developments enter attention in the first place. Distribution consequently becomes a service in its own right, alongside the publication and storage of information.
For marketing, this creates an important analytical distinction. A business may publish material without receiving meaningful attention. Conversely, an update may attract notice even when the recipient was not deliberately looking for the company. The presence of content and its actual circulation must be examined together, because neither alone describes the communication that occurred.
It would nevertheless be inaccurate to assign every subsequent selection procedure to the first version of News Feed. The initial historical change concerned an alternative way of encountering information. Later ranking practices and commercial uses had their own development. Preserving that chronology explains more than treating the first feed as an already complete version of everything that followed.
September 2006: Registration extended beyond institutional entry points
On 26 September 2006, Facebook announced wider registration using a valid email address. Its statement distinguished this opening from the verification still required for certain college and workplace networks. It also referred to more than five hundred regional networks (Facebook, 2006).
The change mattered to the development of an advertising medium. A service restricted to particular institutions has a different potential composition from one with broader entry. Additional participants could bring other interests, circumstances and information needs. Growing membership therefore involved changes in the possible audience as well as an increase in its size.
Access should not be confused with evenly distributed social coverage. Being able to register does not reveal which groups actually join, how frequently they participate or what they do there. For advertisers, the relevant issue remains whether people important to a particular offer can be reached in circumstances where the communication makes sense.
The milestone was consequently an expansion of entry rather than the immediate disappearance of every boundary. A network rooted in campus affiliations was becoming a broader environment. Its commercial development had to accommodate changes in composition and use. Reach emerged through participation, while the economic significance of that participation still depended on the fit between an advertiser, an audience and an occasion.
The 2007 Platform: Applications widened the available actions
The Facebook Platform launched on 24 May 2007, according to Meta’s chronology (Meta, n.d.). The later advertising partner announcement illustrates how brands intended to connect applications with their Facebook presence (Facebook, 2007b). The environment could support activities extending beyond displaying profiles or presenting an advertisement.
For marketing history, the change concerns the form of participation being offered. A message can be read, while an application can make additional choices or interactions possible. Design and activity may become part of an encounter with a brand. Offering those actions does not, however, determine whether people will find them useful or worthwhile.
The arrangement also introduced additional organisational relationships. Developers could contribute work alongside the platform and the advertiser. A commercial idea had to become a functioning application within a particular technical setting. Marketing decisions therefore became connected with product design, maintenance and the practical operation of an experience.
The existence of an application remains evidence of an available activity rather than evidence of success. Whether that activity produced meaningful contact, repeated use or purchases requires further investigation. Platform history explains what could be attempted. The effects of particular attempts need their own evidence, especially when launch announcements describe ambitions rather than independently measured outcomes.
November 2007: Facebook Ads combined presence, distribution and feedback
Facebook introduced its Ads system on 6 November 2007. The announcement distinguished business Pages, Social Ads and an interface for evaluating activity. It identified more than sixty brand partners and over one hundred thousand Pages established for the launch (Facebook, 2007a).
This is the more precise milestone for the expanded social advertising offering. An ongoing presence, paid distribution and feedback were presented as connected services. The arrangement differed from individual bookings within an early campus directory. Both the audience and the commercial products available around it had changed during the intervening years.
The numbers and statements of benefit are the supplier’s claims. They establish the scale and positioning of an announcement, rather than demonstrating profitability for every participant. Historically, the central question is which services were brought together and how they organised the relationships between a brand, its potential audience and the intermediary.
A company could maintain a destination while also purchasing distribution for selected messages. Those activities remained distinct. Having a Page did not guarantee attention, and receiving a paid impression did not automatically create an enduring relationship. Their combination nevertheless required new choices about content, expenditure, response and the interpretation of observed activity.
Sprite Sips: Connecting packaging with digital interaction
The partner announcement described a Sprite Page and the Sprite Sips application. Participants were to configure an animated character and interact with it. Codes beneath the caps of selected US Sprite bottles were intended to unlock additional features (Facebook, 2007b).
The case is historically useful because it linked a purchased product, its packaging and a digital activity. A bottle could provide more than the display of a brand identity. It could contain a means of entering a further encounter within the platform. The planned experience therefore crossed between a physical consumer product and a digital environment.
The announcement documents the proposed arrangement and its introduction. It does not independently demonstrate additional sales. Its value as evidence lies in the design of the connection: a consumer goods brand sought to associate its physical product with a digital representation that people could use in a social setting.
This is more informative than saying only that brands had entered a conversation. The source identifies actions and transitions that were actually proposed. Whether a character attracted interest or gave people a reason to interact depended on use. Social circulation was an intended result, not an effect established merely by making the application available.
A Page did not automatically create a brand community
Albert Muniz and Thomas O’Guinn had examined brand communities in 2001 through Ford Bronco, Macintosh and Saab. They described relationships among brand admirers and features including shared consciousness, rituals and responsibility towards other members (Muniz and O’Guinn, 2001).
A scholarly account of brand community thus preceded Facebook. A technical Page could support such relationships without being sufficient to establish them. A collection of people connected to a brand differs from a community whose members develop meanings, knowledge or obligations involving one another. These are related possibilities, but they require different evidence.
For companies, the practical task was more demanding than simply creating a presence. Useful information, intelligible responses and recurring occasions for participation could contribute to relationships. A large connection count alone could not explain what participants shared or whether the environment mattered to them beyond an occasional encounter.
Community marketing should consequently be treated neither as a Facebook invention nor as an automatic result of a button. The platform supplied an organisational setting in which older social processes could become visible and accessible. Whether a particular brand community actually formed remained an empirical question about relationships and practices, rather than a conclusion available from the interface alone.
Beacon: Activities outside Facebook entered the social information space
Also on 6 November 2007, Facebook announced Beacon with forty-four participating websites. Activities beyond Facebook could become visible to friends through the service (Facebook, 2007c). For Blockbuster, the partner announcement described sharing film titles added to an online queue (Facebook, 2007b).
The proposed arrangement moved a boundary. An action on another website could become the starting point for information circulating inside the network. It connected situations that a person might not have regarded as a single communicative setting. The technical capacity to transfer information did not settle how that transfer would be understood.
Commercially, the attraction was apparent: an activity involving a product or service might gain additional visibility through personal relationships. Its meaning nevertheless depended on how the information appeared and what the person involved intended. An observed action and a deliberately expressed recommendation are different communicative events, even when both refer to the same product.
Beacon therefore exposes a basic problem in the development of social advertising. Technical distribution did not by itself establish appropriateness within the receiving context. Expectations about publication and control were part of the conditions under which commercial intermediation could be accepted. They belonged to the viability of the service itself.
December 2007: A correction to the original arrangement
In a December 2007 statement, Mark Zuckerberg acknowledged mistakes with Beacon. He discussed the problems of the original opt-out approach and described explicit approval and a means of turning the feature off completely (Zuckerberg, 2007).
The statement was the company’s response to difficulties with its own product. It should be read as such, rather than as an independent assessment or an explanation of every later privacy controversy. For this episode, it documents a revision to the relationship between automatic circulation and a participant’s deliberate decision.
The case also clarifies the meaning of social endorsement. A platform can display information associated with someone a recipient knows without that person actively composing a recommendation at that moment. The connection may still influence the recipient. Explaining what the visible signal actually represents therefore becomes important to interpreting the communication.
Social advertising history includes revisions and withdrawals as well as expansion. A commercially plausible design had to encounter the expectations of its participants. That process could alter the product. Such corrections help explain why the later system should not be treated as fixed at the network’s beginning or as the inevitable execution of an original plan.
Likes and the older image of the consumer as judge
Meta’s chronology dates the introduction of the Like button to 9 February 2009 (Meta, n.d.). A limited reaction became easy to perform and count. What the reaction meant still required interpretation, because the action did not answer every question an observer might wish to ask.
Stefan Schwarzkopf’s historical study, published in 2011, examined representations of the consumer as voter, judge and jury. He connected their development with market research practices of the 1930s and 1940s that appeared to make preferences directly observable (Schwarzkopf, 2011).
The comparison here is a historical interpretation, not a claim that Schwarzkopf empirically investigated this particular Facebook feature. It draws attention to a recurring tendency: a measurable choice can be elevated into a much broader statement about what an audience wants. A convenient signal can have several motivations and remain a limited indicator of commitment or willingness to purchase.
For marketing, countability consequently created both an opportunity and a conceptual task. Observing a reaction can be useful. Turning it into a defensible claim about attitudes, behaviour or economic consequences requires additional reasoning. The history of metrics concerns their interpretation as well as the technologies that make them available.
Management literature translated the platform into organisational practice
Clara Shih’s book The Facebook Era appeared in 2009. A contemporary extract discussed social networks as resources for understanding customers, preparing conversations and sustaining longer relationships. It connected the emerging information environment with sales and customer relationship management (Shih, 2009).
This records another stage of development. The platform was being interpreted as more than a place to buy advertising space. It could enter discussions of several organisational activities, including gathering information before a conversation and maintaining contact afterwards. Commercial interpretation was expanding alongside the available products.
Management literature is evidence of contemporary expectations, not proof that every business obtained the outcomes described. Its historical importance lies partly in the work of translation. New technical possibilities need to acquire names, purposes and connections to existing responsibilities before they become intelligible within organisations.
The development of a medium therefore involves more than product releases. Businesses also need accounts of what it might be for and how it relates to established work. Professional writing can organise and circulate those accounts. The history of the platform and the history of its practical interpretation developed alongside each other.
What field experiments established about social advertising cues
Eytan Bakshy and colleagues examined social cues in two Facebook field experiments. In one, a small reference to a friend’s connection with an advertised Page raised click-through rates by 3.8–5.4 per cent and new Page connection rates by 9.6–11.6 per cent, relatively. The estimates varied with the number of already connected friends (Bakshy et al., 2012).
These were relative changes under the conditions examined, not percentage-point increases or general sales gains. Their precise scope is what makes them historically valuable. A central product proposition of social advertising was being subjected to experimental investigation through defined reactions, rather than being accepted solely as a plausible account of influence.
The study also separated displayed social information from the possibility that connected people already shared interests. Someone with several friends associated with a brand might be more receptive even without seeing a friend’s name in an advertisement. An appropriate comparison is necessary to distinguish existing similarity from an effect of the cue itself.
This matters to understanding the advertising system. A relationship between friendship connections and brand responses does not, on its own, establish a cause. Experimentally changing the visible information produces a different kind of evidence from inspecting existing behavioural patterns. Product claims and research findings therefore occupy different explanatory levels and belong to different moments in the chronology.
The scale reported for 2011 was an outcome of development
Facebook’s 2012 registration statement reported approximately 845 million monthly active users at the end of 2011. Advertising revenue for 2011 was $3.154 billion, representing 85 per cent of total revenue (Facebook, 2012). These figures describe a position seven years after launch, rather than the network’s original circumstances.
They establish substantial economic scale without establishing an equally substantial benefit for every advertiser. Platform revenue and campaign profitability are different quantities. Similarly, a large overall membership does not guarantee intensive participation by the particular people important to a particular business.
Scale should therefore complement the explanation of emergence rather than replace it. The financial result becomes historically meaningful through changes in access, participation and commercial organisation. Reading a later position back into the beginning as a completed plan would remove the intermediate decisions, trials and corrections that require explanation.
The distinction allows an initially bounded service to be studied without diminishing its subsequent importance. A historical milestone can identify a consequential beginning while recognising that later outcomes were not already present. Facebook’s marketing significance developed through a sequence of changes whose timing and relationships deserve attention.
Later measurement research qualified the promise of abundant data
Brett Gordon, Florian Zettelmeyer, Neha Bhargava and Dan Chapsky compared observational approaches with results from fifteen US Facebook advertising experiments in 2019. Even with extensive demographic and behavioural information, observational methods frequently produced effect estimates different from those obtained through randomised comparisons (Gordon et al., 2019).
The underlying problem of advertising effectiveness had therefore not disappeared with data abundance. Everyday exposure is not randomly assigned. Interest, selection and previous experience can affect both advertising delivery and subsequent behaviour. A richly described sequence of events does not automatically establish the additional contribution made by an advertisement.
This was a later examination of a developed advertising system. It should not be presented as knowledge already fully available in 2004. Retrospectively, it helps define the historical contribution more precisely: Facebook expanded possibilities for reaching audiences and observing activity while leaving important questions about incremental economic effects open to investigation.
The significance of 2004 lies in the beginning of a network that progressively became an advertising intermediary. Early campus bookings, the 2007 system launch, particular brand applications and subsequent corrections explain that development better than a single invention narrative. Social media marketing emerged here through the interaction of audiences, available actions, commercial rules and continuing evaluation of what those arrangements actually achieved.
References
Bakshy, E., Eckles, D., Yan, R. and Rosenn, I. (2012) ‘Social influence in social advertising: Evidence from field experiments’, Proceedings of the 13th ACM Conference on Electronic Commerce, pp. 146–161. https://doi.org/10.1145/2229012.2229027. Full text: https://arxiv.org/abs/1206.4327.
boyd, d.m. and Ellison, N.B. (2007) ‘Social network sites: Definition, history, and scholarship’, Journal of Computer-Mediated Communication, 13(1), pp. 210–230. https://doi.org/10.1111/j.1083-6101.2007.00393.x.
Facebook (2006) Facebook expansion enables more people to connect with friends in a trusted environment. Press release, 26 September. https://about.fb.com/news/2006/09/facebook-expansion-enables-more-people-to-connect-with-friends-in-a-trusted-environment/ (Accessed: 23 October 2026).
Facebook (2007a) Facebook unveils Facebook Ads. Press release, 6 November. https://about.fb.com/news/2007/11/facebook-unveils-facebook-ads/ (Accessed: 23 October 2026).
Facebook (2007b) Facebook Ads launches with 12 landmark partners. Press release, 6 November. https://about.fb.com/news/2007/11/facebook-ads-launches-with-12-landmark-partners/ (Accessed: 23 October 2026).
Facebook (2007c) Leading websites offer Facebook Beacon for social distribution. Press release, 6 November. https://about.fb.com/news/2007/11/leading-websites-offer-facebook-beacon-for-social-distribution/ (Accessed: 23 October 2026).
Facebook (2012) Registration statement on Form S-1. 1 February. US Securities and Exchange Commission. https://www.sec.gov/Archives/edgar/data/1326801/000119312512034517/d287954ds1.htm.
Gordon, B.R., Zettelmeyer, F., Bhargava, N. and Chapsky, D. (2019) ‘A comparison of approaches to advertising measurement: Evidence from big field experiments at Facebook’, Marketing Science, 38(2), pp. 193–225. https://doi.org/10.1287/mksc.2018.1135.
Grynbaum, M.M. (2004) ‘Online Facebook solicits new ads’, The Harvard Crimson, May. https://www.thecrimson.com/article/2004/5/7/online-facebook-solicits-new-ads-thefacebookcom/ (Accessed: 23 October 2026).
Kim, K.-H. and Yun, H. (2007) ‘Cying for me, cying for us: Relational dialectics in a Korean social network site’, Journal of Computer-Mediated Communication, 13(1), pp. 298–318. https://doi.org/10.1111/j.1083-6101.2007.00397.x.
Meta (n.d.) Unternehmensinformationen: Unsere Geschichte. https://about.fb.com/de/company.info/ (Accessed: 23 October 2026).
Muniz, A.M., Jr. and O’Guinn, T.C. (2001) ‘Brand community’, Journal of Consumer Research, 27(4), pp. 412–432. https://doi.org/10.1086/319618.
Schwarzkopf, S. (2011) ‘The consumer as “voter,” “judge,” and “jury”: Historical origins and political consequences of a marketing myth’, Journal of Macromarketing, 31(1), pp. 8–18. https://doi.org/10.1177/0276146710378168.
Shih, C. (2009) The Facebook Era: Tapping Online Social Networks to Build Better Products, Reach New Audiences, and Sell More Stuff. Addison-Wesley. Extract consulted: ‘Sales in the Facebook era’, CRM, 1 June. https://www.destinationcrm.com/Articles/ReadArticle.aspx?ArticleID=54771.
Tabak, A.J. (2004) ‘Hundreds register for new Facebook website’, The Harvard Crimson, 9 February. https://www.thecrimson.com/article/2004/2/9/hundreds-register-for-new-facebook-website/ (Accessed: 23 October 2026).
Zuckerberg, M. (2007) Thoughts on Beacon. Published as ‘Announcement: Facebook users can now opt-out of Beacon feature’, 6 December. https://about.fb.com/news/2007/12/announcement-facebook-users-can-now-opt-out-of-beacon-feature/ (Accessed: 23 October 2026).
