Two computer demonstrations became the occasion for a dispute about who could use a network to reach its users. On 3 May 1978, Gary Thuerk of Digital Equipment Corporation sent an unsolicited promotional message over the ARPANET. The invitation advertised members of the DECSYSTEM-20 computer family. The Computer History Museum identifies the episode as an early documented case of commercial bulk electronic mail. What a supplier could regard as useful product information became a challenge to the expectations surrounding a shared communications system (Computer History Museum, n.d.a).
This was more than a curious first in the history of the internet. It exposed a tension at the heart of digital marketing: the ability to reach an address does not establish that its user expects or welcomes an advertisement. The episode brings together business-to-business selling, event promotion, the use of contact information and the distribution of communication costs. Its historical importance comes from the interaction of these elements, rather than from the press of a single send command.
Defining the milestone accurately
Thuerk did not invent electronic mail, group communication or the idea of delivering advertising electronically. His episode is better understood as a documented commercial appropriation of an existing network: a supplier used access to a research community to announce products to a substantial group without first establishing individual interest. Narrowing the claim in this way makes the event more historically useful. It distinguishes a particular promotional intervention from the many other developments that made email possible.
Elizabeth Feinler and John Vittal’s historical timeline separates a non-network CTSS mass message in 1971 from commercial ARPANET spam in 1978. It also records Paul Baran’s anticipation of electronic direct advertising in 1971. Thus neither unsolicited computer messages in general nor the idea of electronic advertising began with DEC’s invitation (Feinler and Vittal, 2022, pp. 23-24, 60).
The distinction matters because several different things can be called a beginning. The earliest surviving example is not necessarily the earliest event. The first commercial use is not the same as the introduction of a communication technology. A controversial experiment is not automatically the start of an established industry. Each claim needs its own definition and its own evidence. In this case, the precise reference to commercial bulk mail on the ARPANET is more defensible than an unrestricted claim about all electronic advertising everywhere.
The familiar expression “first spam email” is therefore useful as a point of entry, provided that the account explains its limits. It should lead readers into the history rather than replace that history with a slogan. The significance of 1978 survives this qualification: the documented encounter between a promotional objective and a network community remains unusually revealing.
A technical supplier addressing a specialist environment
Digital Equipment Corporation was promoting a substantial computing investment, rather than an ordinary consumer purchase. The surviving announcement presented members of the DECSYSTEM-20 family and emphasised their connection with the ARPANET. Its argument depended on technical characteristics and the possibility of seeing the machines in operation. This was a specialist invitation addressed within a technically informed environment (Thuerk, 1978).
Viewed as a marketing artefact, the message has the structure of a business-to-business campaign. A product that requires explanation is presented through a next step: attend a demonstration, ask questions and potentially continue a sales conversation. A demonstration can reduce uncertainty in ways that a short description cannot. This interpretation describes the communication’s logic; it does not suggest that DEC already operated the automated lead-management systems associated with later digital marketing.
Technical relevance, however, is not identical to purchasing relevance. A researcher might be interested in the design of a machine without having any responsibility for buying one. An administrator might recognise the importance of a feature while having no budget for a replacement system. An institution might be committed to an existing installation. Membership of the same technical community therefore establishes only a possible connection between the product and its audience.
That connection explains why the campaign was not simply a random advertisement. It also explains why a defence based entirely on relevance was incomplete. Information can concern a recipient’s professional interests and still arrive through an unwelcome channel. The subject of a message, its timing, its distribution and the recipient’s relationship with the sender all contribute to its acceptability. A specialist product does not remove those distinctions.
The immediate objective was attendance
The invitation announced presentations on 9 May in Los Angeles and 11 May in San Mateo. It provided locations, offered a demonstration and directed people unable to attend towards DEC offices for further information. The first requested action was therefore participation or an enquiry, rather than a purchase completed within the message itself (Thuerk, 1978).
This arrangement linked an electronic invitation to a physical sales setting. The network did not replace the demonstration. It was intended to help bring people into it. A history that treats the episode simply as the arrival of online selling misses this division of labour. The communication channel carried an invitation; further explanation and evaluation were expected to take place elsewhere.
The two locations also make geography important. The nominal size of an address list is not the same as the number of people who can reasonably act on an invitation. Distance, availability and organisational circumstances affect attendance. Someone unable to visit either location could still enquire, but that would be a different response, requiring different subsequent work. The existence of an alternative contact route does not make every recipient equally likely to become a useful prospect.
These distinctions allow a more careful reading of the campaign. Delivery, awareness, interest, attendance and purchasing are separate stages. Their separation here is analytical; the surviving invitation does not supply reliable counts for each. It records what the sender wanted people to do. It cannot by itself show how far any particular recipient proceeded. The message is evidence of a commercial intention and a proposed next step, not a complete record of its consequences.
Why the audience cannot be reduced to one exact number
Published accounts do not agree entirely on the size of the distribution. Guinness World Records gives 397 email accounts. Other retellings round the number or use different totals. Describing the attempt as reaching several hundred addresses is therefore safer than treating one precise number as an independently established count of people who read the message (Guinness World Records, n.d.).
The issue is more substantial than a small discrepancy in a historical anecdote. An account does not necessarily correspond to a single person. An address in a list does not establish successful delivery. Delivery does not establish reading, and reading does not establish interest. A responsible description of reach must state which of these things is being counted. Otherwise numbers that appear comparable can refer to quite different events.
Consider an entirely hypothetical invitation sent towards 400 addresses, accepted by 360 mailboxes and noticed by 120 people. These figures are not DEC’s results. They illustrate how a technical delivery rate differs from an attention measure. Attendance divided by accepted messages would be a different ratio from attendance divided by people who actually noticed the invitation. Neither would directly reveal additional sales.
Historical evidence rarely supplies all the observations required for such calculations. Filling the gaps with plausible estimates would create an appearance of precision without improving knowledge. In Thuerk’s case, the promotional content and the controversy are more accessible than a complete campaign dataset. That uneven survival of evidence should shape the narrative. It is possible to explain why an episode matters without manufacturing the performance report that it does not preserve.
The communication system was part of the campaign
In November 1977, RFC 733 set out a format for ARPA network text messages, including their headers and addressing conventions. Its purpose was to improve consistency between message systems (Crocker et al., 1977). Such arrangements are easy to overlook when attention centres on the advertisement itself. Yet they help determine whether a document is processed as a usable message across different machines.
For a marketer, this infrastructure has two sides. Common formats facilitate distribution, but actual programs still impose practical constraints. Text that appears to a reader to be an address does not necessarily function as a recipient instruction wherever it is placed. A long list can be intelligible to a person and unsuitable for a particular implementation. Campaign preparation therefore includes the way a system interprets the message, rather than only the wording people eventually see.
This is not just a matter of avoiding clerical mistakes. New media change the knowledge required to communicate effectively. Persuasive product arguments have little value if they arrive in an unusable form. Equally, flawless transmission cannot make a poorly judged approach welcome. Reliability and appropriateness are different qualities, and a campaign needs both. The historical episode makes their relationship particularly visible because the medium itself was still developing.
Later procedures should also remain in their proper chronology. Postel’s SMTP specification, RFC 821, was published in August 1982 (Postel, 1982). The 1978 message should not be described as if it used that later specification. Familiar modern terminology can conceal substantial changes in the systems behind it. Preserving the dates prevents a retrospective account from turning an early network message into an ordinary contemporary email campaign with an old timestamp.
A dispute over the purpose of a shared network
The Computer History Museum records a swift and strongly negative reaction to the DEC mailing (Computer History Museum, n.d.a). The significance of that reaction extends beyond individual annoyance. Different participants could understand the same network in different ways: as a route towards potential buyers, as a working environment or as infrastructure whose purpose could not be redefined by an individual supplier.
Three questions need to be separated. Was the product potentially interesting? Was a bulk promotional message an appropriate way to announce it? Was the channel available for that use? Someone could answer the first question positively and the others negatively. The advertisement’s technical relevance did not resolve the dispute. Instead, the case demonstrated that useful information could still be considered an improper claim on shared attention and resources.
This also distinguishes DEC’s invitation from later examples involving invented products, deception or malicious software. The advertisement concerned real computer systems. Its controversial character did not depend on a fraudulent offer. That makes it an especially instructive marketing case: a legitimate supplier can damage acceptance through the manner of contact, even when the product itself is genuine and professionally relevant.
A history concerned only with increasing reach would miss this dimension. The development of digital promotion also involves questions about rule-making, complaints and responsibility for consequences. A campaign does not take place exclusively between an advertiser and a series of isolated individuals. Operators and other users can be affected by the same action. Understanding the network as a shared arrangement makes those additional participants visible.
Unwanted traffic had already become a technical question
Jon Postel’s November 1975 RFC 706 was entitled “On the Junk Mail Problem”. It considered unwanted messages as a burden on hosts and network resources and discussed ways of refusing traffic from particular sources. Its subject was a technical issue at the host and interface level, rather than an account of an already established commercial advertising industry (Postel, 1975).
The document changes the sequence in which the story can be told. Concern about unwanted traffic did not have to wait for commercial bulk advertising to become large. The architecture itself made the possibility of burdensome incoming communication intelligible. A receiving system could incur costs even when its eventual decision was to discard a message. Refusal was therefore not equivalent to the absence of work.
For marketing history, this supports a broader economic interpretation. A cheap communication for its initiator can impose effort elsewhere. Sorting, administration, storage and complaint handling may occur outside the advertiser’s budget. Looking only at the price of an additional send can consequently make a campaign appear efficient while leaving part of its consequences uncounted.
This is an interpretation of the cost structure, not a reconstruction of DEC’s expenses or the recipients’ losses in 1978. No comprehensive figures are offered here. The point is that the party causing a message does not necessarily bear all the effort required to receive and evaluate it. As repeated distribution becomes easier, that separation becomes more consequential. It helps explain why a communication can be commercially attractive to a sender and collectively unwelcome at the same time.
Earlier and international comparisons
An earlier comparison comes from Britain. The Computer History Museum preserves an item dated 1 June 1864 under the heading “Telegraph spam”. It describes it as an early unsolicited electronic message while noting the limiting effect of the cost of telegrams (Computer History Museum, n.d.b). This provides a predecessor outside computer mail without requiring telegrams and email to be treated as the same technology.
The differences are precisely what make the comparison useful. Telegraphic messages had different payment, transmission and delivery arrangements. Nevertheless, faster reach could already create a conflict over unwanted promotion. The British example does not disprove a narrowly defined ARPANET milestone. It establishes that the underlying social tension had an older history, while the conditions governing its scale changed between media.
The CTSS example of 1971 offers another distinction. The museum timeline describes a noncommercial message distributed within a single computing system (Feinler and Vittal, 2022, p. 24). An unsolicited communication to many users need not be an advertisement, and a mass computer message need not travel between machines. Purpose, scale and network transmission are separate characteristics. Keeping them separate makes historical comparisons more informative than a competition for one universal first.
A later case involved Canter and Siegel’s commercial postings to Usenet groups in 1994, discussed in the Internet Society’s history of spam (Internet Society, n.d.). Here the contested setting included topic-based discussion groups. That differed from sending an invitation towards individual mail accounts. Collapsing both incidents into a single category of “first email advertising” would erase important differences in where the communication appeared and why users objected.
An international perspective does not require an unsupported catalogue of national firsts. British telegraphy, an American research network and later cross-border online communication already demonstrate why the comparison needs a defined unit. Countries and institutions did not adopt every medium in the same sequence. A claim about the first electronic advertisement anywhere would require evidence far beyond the best-known American computer-network anecdote.
Customer understanding predates digital personalisation
The relationship between a sales approach and a recipient’s situation has a much longer history. In the Journal of Historical Research in Marketing, Tadajewski and Jones analyse the 1913 training series The Knack of Selling. Their findings include a context-sensitive view of selling and attention to customers’ changing needs (Tadajewski and Jones, 2020). Adaptation to a prospective buyer therefore did not begin with digital databases.
This helps locate what was distinctive about a network mailing. A new delivery mechanism does not automatically introduce a more developed understanding of relationships. A message can travel over advanced infrastructure while revealing little knowledge of any individual recipient. Conversely, a conversation conducted without electronic assistance can respond closely to a particular organisation’s purchasing problem.
A CHARM paper by Proença, Fernandes and Kannan similarly examines relationship marketing through several historical lines of development (Proença, Fernandes and Kannan, 2007). It does not establish a direct influence on DEC’s mailing. Its relevance is comparative: customer relationships belong to continuing exchanges and organisational contexts. They cannot be inferred simply from the existence of contact details.
The Routledge Companion edited by Jones and Tadajewski also places marketing practice across activities, countries and historical periods (Jones and Tadajewski, 2016). Within that broader framework, 1978 appears as a change in the means of communication within older practices of selling, inviting and seeking customers. This does not diminish the event. It clarifies which part was technologically distinctive and which parts inherited familiar commercial purposes.
Why a famous revenue claim is not a campaign evaluation
Popular versions of the story often associate the mailing with a large sales figure. A surviving invitation, however, does not demonstrate completed purchases or show what caused them. Without independent records linking recipients, enquiries, offers and orders, a claim of commercial success remains a claim about effects. It should not be converted into a verified performance result simply because it has been repeated frequently.
Even a documented order after a demonstration would leave questions unanswered. Did the buyer receive the message? Was the organisation already speaking with the supplier? Did the invitation initiate demand or merely accompany an existing negotiation? These questions distinguish sequence, participation and causation. They determine whether a campaign can be credited with additional demand rather than merely associated with a sale.
Imagine a fictional campaign followed by ten customer orders, seven of which would have occurred anyway. Attributed sales and additional sales would then differ substantially. These numbers do not describe DEC. They illustrate why a headline order value cannot establish return on investment on its own. A small mailing followed by a large contract is an interesting sequence, but it is not automatically evidence of the contract’s cause.
The relevant cost base would also need definition. Preparation, technical assistance, the event and subsequent selling all belong to the commercial process, even if the act of transmission was inexpensive. A calculation confined to delivery could overlook the work necessary to convert interest into an order. The defensible historical conclusion is therefore narrower: the message had a commercial objective and a specific action for recipients. Its additional profit cannot be independently established from that fact.
From a disputed mailing to an economics of spam
Later research examines spam through the distribution of incentives and costs. Rao and Reiley’s 2012 study analyses the economic structure of an environment very different from DEC’s experiment. Its estimates belong to that later period and should neither be presented as current statistics nor applied to the 1978 campaign (Rao and Reiley, 2012).
The useful conceptual distinction is between an initiator’s return and the burdens imposed elsewhere. A very low response rate can be sufficient for a sender when distributing another message costs little. Recipients still incur work evaluating messages that produce no sale. Advertising success and unwanted resource consumption are consequently compatible outcomes. One does not disprove the other.
Hambridge and Lunde’s 1999 RFC 2635 also addresses why unsolicited mass messages create problems for users and operators (Hambridge and Lunde, 1999). Its existence documents the development of explicit guidance around a communication problem. What could initially appear as an argument about one message became something that organisations needed to manage as a recurring category of activity.
This later development should not be narrated as the inevitable consequence of Thuerk’s action. Many technical, commercial and social changes intervened. One invitation did not create the later spam economy by itself. Its importance is that it exposed a recognisable tension early: distribution could offer a low-cost opportunity to an advertiser while imposing an unwelcome claim on the receiving side. Later scale altered the consequences without making the original distinction irrelevant.
An address list is not yet a market segment
The episode can also be read as a problem in segmentation. An audience may be assembled around an available characteristic that only imperfectly predicts its members’ information needs. In DEC’s case, network involvement had an obvious relationship to a product advertised through network-related features. Yet people connected to the same technical environment need not occupy the same position in a purchasing decision.
Consider three fictional employees at a research institution: a systems administrator, a scientist and a procurement specialist. All might have reasons to be interested in the same computer. Their questions could nevertheless concern integration, research applications and purchasing conditions respectively. Shared institutional membership would not make their interests interchangeable, nor would it reveal which of them could authorise a purchase.
For the history of direct marketing, this suggests a more demanding question than identifying the first large list. How did a sender move from an available contact characteristic to an expected action? The inference can seem sensible and still be too broad. If the action also requires attendance at a particular place and time, practical constraints narrow the audience further. The number of reachable accounts can substantially exceed the number of realistically relevant prospects.
Later software terminology should not conceal that uncertainty. There is no evidence presented here of a sophisticated scoring procedure, automated behavioural analysis or continuing individual personalisation. The historical interest lies partly in the opposite imbalance: the capacity to distribute widely is easier to establish than detailed knowledge of each recipient. Expanded access did not by itself create a correspondingly differentiated approach.
Reading the evidence without inventing the missing report
The episode offers different kinds of evidence, each suited to different questions. The announcement records the intended message and proposed action. A museum description supplies a later historical classification. A technical specification sets out intended procedures. A scholarly analysis offers a wider framework. These sources can inform one another, but they are not interchangeable.
An invitation can establish a date without proving attendance. A retrospective account can illuminate motives while remaining affected by the distance between event and recollection. Repetition of a story does not create an additional independent witness. This is especially important in marketing history, where a memorable account of campaign success can also become an argument for the power of a medium.
Missing records need equally careful treatment. The absence of a complete attendance list does not mean nobody attended. The absence of independently verified revenue documentation does not prove commercial failure. The appropriate conclusion is that some outcomes cannot be determined from the evidence used here. That is a limit to the claim, rather than a reason to replace missing observations with invented metrics.
The same caution applies internationally. A well-preserved American case can be significant even when other computing environments are less accessible to researchers. “Early documented” is therefore a meaningful historical description. It identifies the evidence available without claiming that every older possibility has been ruled out. For early digital communications, where preservation is uneven, that can be more informative than an unrestricted declaration of a worldwide first.
The place of 1978 in email marketing history
Hambridge’s 1995 Netiquette Guidelines distinguished expectations surrounding personal communication and communication to groups (Hambridge, 1995). This later document illustrates an important part of the medium’s history: expanding use produced explicit guidance about conduct as well as new technical capabilities. Advertising operated within those expectations and could not define acceptable communication solely through its own performance measures.
The DEC episode connects three developments. Contact information became a possible route towards demand. Electronic distribution supported an event taking place outside the network. The same action provoked questions about authority, burden and acceptance. None of these developments is adequately captured by saying that someone was simply the first to send a sales message.
Its continuing value lies particularly in the distinction between accessibility and relationship. A large list may be commercially weak if recipients do not expect the approach. An invitation can be professionally relevant and still be considered inappropriate. Successful technical delivery can coexist with reputational conflict. These are interpretations of the documented episode, rather than measurements collected retrospectively for it.
Gary Thuerk’s place in marketing history rests on a visible transition: a supplier treated network mail as an advertising channel, and the community contested the conditions of that use. The invitation of 1978 belongs near the beginning of a significant documented line of commercial bulk email. Its enduring importance is that the possibility of reaching prospective customers digitally appeared alongside an immediate argument about the limits of doing so.
References
Computer History Museum (n.d.a) May 3: Earliest Known Case of Spam. https://www.computerhistory.org/tdih/may/3/ (Accessed: 22 October 2026).
Computer History Museum (n.d.b) Telegraph “spam”. Object 500004786, dated 1 June 1864. https://www.computerhistory.org/revolution/the-web/20/392/2172 (Accessed: 22 October 2026).
Crocker, D. H., Vittal, J. J., Pogran, K. T. and Henderson, D. A. (1977) Standard for the Format of ARPA Network Text Messages. RFC 733. https://www.rfc-editor.org/rfc/rfc733 (Accessed: 22 October 2026).
Feinler, E. and Vittal, J. (2022) Email Innovation Timeline. Computer History Museum, 1 July. https://archive.computerhistory.org/resources/access/text/2022/08/102806104-05-01-acc.pdf (Accessed: 22 October 2026).
Guinness World Records (n.d.) Oldest electronic spam. https://www.guinnessworldrecords.com/world-records/oldest-electronic-spam (Accessed: 22 October 2026).
Hambridge, S. (1995) Netiquette Guidelines. RFC 1855. https://www.rfc-editor.org/rfc/rfc1855.txt (Accessed: 22 October 2026).
Hambridge, S. and Lunde, A. (1999) DON’T SPEW: A Set of Guidelines for Mass Unsolicited Mailings and Postings (spam). RFC 2635. https://www.rfc-editor.org/rfc/rfc2635.txt (Accessed: 22 October 2026).
Internet Society (n.d.) What is spam? A History of Spam. https://www.internetsociety.org/wp-content/uploads/2017/08/What20Is20Spam_0.pdf (Accessed: 22 October 2026).
Jones, D. G. B. and Tadajewski, M. (eds.) (2016) The Routledge Companion to Marketing History. Abingdon: Routledge. https://www.routledge.com/The-Routledge-Companion-to-Marketing-History-1st-Edition/Jones-Tadajewski/p/book/9781315882857. Publisher description and table of contents consulted.
Postel, J. (1975) On the Junk Mail Problem. RFC 706. https://www.rfc-editor.org/rfc/rfc706 (Accessed: 22 October 2026).
Postel, J. B. (1982) Simple Mail Transfer Protocol. RFC 821. https://www.rfc-editor.org/rfc/rfc821 (Accessed: 22 October 2026).
Proença, J. F., Fernandes, T. M. and Kannan, P. K. (2007) ‘The Relationships in Marketing: Contribution of a Historical Perspective’, Proceedings of the Conference on Historical Analysis and Research in Marketing, pp. 172-179. https://ojs.library.carleton.ca/index.php/pcharm/article/download/1580/1424 (Accessed: 22 October 2026).
Rao, J. M. and Reiley, D. H. (2012) ‘The Economics of Spam’, Journal of Economic Perspectives, 26(3), pp. 87-110. https://doi.org/10.1257/jep.26.3.87.
Tadajewski, M. and Jones, D. G. B. (2020) ‘The Knack of Selling: scientific salesmanship, relational themes and military metaphors in early marketing thought’, Journal of Historical Research in Marketing, 12(2), pp. 239-262. https://doi.org/10.1108/JHRM-10-2019-0035. Publicly accessible abstract consulted.
Thuerk, G. (1978) DECSYSTEM-20 product announcement. Computer History Museum, ‘First spam on the ARPANET’, object 500003129. https://www.computerhistory.org/revolution/artifact/392/2366 (Accessed: 22 October 2026).
