Background As limitations on traditional marketing tactics and scrutiny by tobacco control have increased, the tobacco industry has benefited from direct mail marketing which transmits marketing messages directly to carefully targeted consumers utilising extensive custom consumer databases. However, research in these areas has been limited. This is the first study to examine the development, purposes and extent of direct mail and customer databases.
Methods We examined direct mail and database marketing by RJ Reynolds and Philip Morris utilising internal tobacco industry documents from the Legacy Tobacco Document Library employing standard document research techniques.
Results Direct mail marketing utilising industry databases began in the 1970s and grew from the need for a promotional strategy to deal with declining smoking rates, growing numbers of products and a cluttered media landscape. Both RJ Reynolds and Philip Morris started with existing commercial consumer mailing lists, but subsequently decided to build their own databases of smokers’ names, addresses, brand preferences, purchase patterns, interests and activities. By the mid-1990s both RJ Reynolds and Philip Morris databases contained at least 30 million smokers’ names each. These companies valued direct mail/database marketing's flexibility, efficiency and unique ability to deliver specific messages to particular groups as well as direct mail's limited visibility to tobacco control, public health and regulators.
Conclusions Database marketing is an important and increasingly sophisticated tobacco marketing strategy. Additional research is needed on the prevalence of receipt and exposure to direct mail items and their influence on receivers’ perceptions and smoking behaviours.
- Tobacco industry
- Tobacco industry documents
- Advertising and Promotion
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
As limitations on traditional marketing tactics and scrutiny by tobacco control have increased, the tobacco industry has benefited from the use of direct marketing, which allows tobacco companies to transmit messages directly to carefully targeted consumers with the aim of obtaining an immediate response and cultivating lasting customer relations.1 A major element of this strategy is direct mail to individuals on tobacco companies’ extensive mailing lists. Direct mail is both a marketing strategy and a major distribution channel for other marketing materials (eg, coupons, sweepstakes offers, gifts).2 It is currently subject to minimal restrictions in the USA and operates largely out of sight of the general public and under the radar screen of tobacco control, public health and regulatory agencies. Direct mail has been identified as a priority area for research,3 but research on direct mail marketing has been limited and little is known about its extent and its role within the larger tobacco industry marketing mix.
Data on industry expenditures for direct mail are also limited. US Federal Trade Commission (FTC) reports of expenditures for cigarette advertising through direct mail have varied since they were first reported in 1986. The most recent report put direct mail expenditures for 2012 at $45.6 million (all references to dollars are US dollars throughout).4 However, FTC reports underestimate spending for direct mail by reporting expenditures for items and promotions routinely distributed by direct mail in other categories. This includes expenditures on coupons and specialty item distribution, which together accounted for $267.4 million in 2012.4 Although it is difficult to estimate what proportion of that amount involved direct mail, if expenditures for promotional activities distributed through it were reported in the direct mail category, spending estimates would undoubtedly balloon. In fact, in 1988, when the FTC added separate categories for coupons and specialty items, direct mail expenditures fell from $187 to $42.5 million.4 An analysis of 1177 pieces of direct mail in one collection showed that 69% contained coupons,5 thus the exclusion of coupons alone ($240 million in 2012) represents a serious underestimation of the cost (and by extension the extent) of direct mail marketing.
Direct marketing/mail programmes rely on a customer database, an organised collection of data about individual customers, including contact, demographic, psychographic (eg, activities, values, interests) and behavioural information (eg, buying preferences and practices, response to previous contacts). Although databases are essential to tobacco industry targeted marketing as they allow identification of prospects for marketing messages and mining of information for fine-tuning marketing to meet preferences and behaviours of customers,1 ,6 they are a largely unexamined tobacco industry asset.
Tobacco marketing activities both utilise information contained in the databases and collect additional information for the database. Smoker names and information for databases have been obtained in different ways, including at industry-sponsored events (eg, bar/club promotions), through brand-specific websites and from sweepstakes forms or signed coupons. Direct mailings also solicit information that is recorded and utilised for future marketing efforts, such as surveys asking cigarette or lifestyle preferences or coded coupons identifying redeemers.7 In addition, as noted in Davidson's book on marketing of socially unacceptable products, use of database/direct mail marketing provides tobacco companies with “insurance against the time when the mass media might be shut off to them by some stepped-up level of governmental, or even social, regulation.”8
Information on database size and contents is limited and no research has examined reasons for their development. We examined direct mail and database marketing by RJ Reynolds (RJR) and Philip Morris (PM), the two largest US tobacco companies, utilising internal tobacco industry documents. Our main research question was: What were the intentions and purposes for tobacco industry direct mail marketing? As we collected more documents about direct mail, its strong link to the development of extensive databases of smokers also led us to address: Why and how did RJR and PM develop smoker databases and what information do they contain? We also asked: What are the perceived advantages and disadvantages of direct mail for RJR and PM tobacco companies?
We systematically searched previously secret internal tobacco industry documents from RJR and PM in the Legacy Tobacco Document Library (http://legacy.library.ucsf.edu/). Initial searches were conducted between 2006 and 2008 and rerun to confirm earlier findings and identify newly added documents between Fall 2013 and Spring 2014. We used standard techniques,9 including initial keyword searches combining terms related to direct mail (eg, mail*, marketing, budget, advantages), database marketing (eg, database, name generation, size) and marketing strategies utilising direct mail (eg, sweepstakes, coupons). Initial searches yielded thousands of documents. After initial review, the earliest and most extensive efforts in this area were found to be from Philip Morris and RJR tobacco companies, so subsequent analysis focused on these two companies. Documents addressing markets outside the US or the activities of other tobacco companies were not included in this analysis. Documents were reviewed to discern the themes and context, such as the timeline of activities, and to compare and contrast strategies of the two tobacco companies. Initial searches were followed by snowball searches utilising the names of specific projects (eg, Project MD), and other information (eg, key individuals, third parties involved, meeting dates, consecutive reference (Bates) numbers).
We wrote and reviewed summary memoranda based on documents retrieved, and resolved questions by gathering additional data. To validate and contextualise activities discussed we triangulated findings both from within the documents library and from other sources, such as general online search engines (eg, Google), the Trinkets and Trash tobacco advertising archive (eg, http://www.trinketsandtrash.org) and marketing literature. We repeated the iterative search process until we reached saturation, and additional keywords yielded no new relevant documents. This analysis is based on a final collection of approximately 275 documents, most of which originated from 1970 to 1992.
Documents are inconsistent in how they refer to direct mail and database marketing with the terms direct mail, database marketing and direct marketing used interchangeably. All references were carefully reviewed to establish that the document was specifically referring to marketing through direct mail and its databases but quotes retained the terms used in the document.
The early years of RJR direct mail marketing
Tobacco industry direct mail marketing can be traced to RJR's distribution of coupons through mailings utilising the databases of commercial cooperative mailing firms, beginning by at least 1970.10 These mailings distributed an RJR coupon along with those of other products to names on the mailing firm's database. Three mailings in 1970 and 1971 through the firms Metromail and RH Donnelly to a combined 9.14 million names aimed at increasing awareness and trial among smokers of non-RJR brands living in regions where RJR brands were less popular.11–13 The programmes were evaluated through various means, including contacting and interviewing recipients regarding product trial, use and awareness and analysing redemptions of the mailed coupons and sales data for test areas. Evaluations showed increased awareness and a higher recall for direct mail than for magazine ads and found mailings were more efficient at encouraging brand switching than sweepstakes, then considered to be RJR's most successful promotional strategy.14 Comparison of costs for sweepstakes and a 25 cent coupon mailing showed estimated cost per participant as almost equal but payoff as greater for the mailing, with redemptions among smokers of non-RJR brands (the real targets of the promotion) twice as high for a mailing as for sweepstakes (40% compared to 20%) and the average cost of the promotion per pack sold four times as high for sweepstakes as for a mailing (20 cents vs 5 cents).14
Although RJR's couponing through cooperative mailings continued throughout the 1970s, the company recognised that there was a need for more precise targeting that would eventually allow them to target smokers of specific brands.
RJR builds its own database
In 1980 RJR began considering a system for targeted mailings to specific potential customers utilising its own database, and in April 1980 approved Project MD to ‘develop a totally new marketing system’ utilising direct mail and tracking participant response.15 This was partly driven by the need to address the decreasing pool of available smokers and a growing and increasingly fragmented market. By 1980, smoking rates were falling and projected to shrink further (from 1970's 51.6 million to a projected 47.3 million in 1990) but the number of cigarette product options had doubled (from 93 to 184) with many more choices of flavour, tar level, length and package style.15 A 1983 article in Advertising Age quotes RJR's vice president for brand marketing Martin Orlowsky calling this increasingly segmented market ‘the most significant change in the tobacco industry in the past 30 years’.16
Project MD planners noted that being able to target specific groups would be vital in the future when growth would come from concentrating marketing against specific target smokers, especially those most likely to switch.15 RJR's ability to target high potential users was restricted by limitations of available marketing channels, especially for products such as ultra-low tar brands where the market was small compared to the circulation of traditional channels (eg, magazine advertising). Direct mail allowed the company to target marketing messages to specific segments of the smoker population. It was also flexible and could be used both to recruit new users to a brand and to retain existing ones,15 ,17 and was relatively cheap—with a per unit cost of between 11 cents for a self-mailer and 21 cents for a mailing containing a letter, brochure, envelope and reply card.17
However, RJR needed a large, continuous supply of smokers—a databank of smokers—and recognised a significant financial investment would be required. Project MD's 1980 budget included over $2 million in fees to an outside firm to contact and screen households to identify smokers and supply their names and relevant info (eg, demographics, brand and style smoked, likelihood of switching). Planners felt costs would be offset by financial gain and would give RJR a competitive advantage that could not be easily duplicated.15 The plan noted that each converted smoker generated $90 in direct variable profit.15 These ‘name generation’ (collecting names and information) expenditures continued and in 1985–1986 RJR contracted with the firm M/A/R/C (one of four selected suppliers of competitive smoker names) to supply over 2.4 million names for over $6.6 million.18–20 By then, RJR's programmes included offensive or conversion mailings encouraging a switch to an RJR product, and defensive or continuity mailings to RJR smokers offering incentives (eg, coupons) for continued use.21
Competition moves in—Philip Morris
In 1983, RJR considered its new system “the state-of-the art approach to cigarette marketing.”21 There were, however, external threats. RJR noted movement by its competitors Brown and Williamson and Lorillard into targeted marketing and recommended doubling 1984's budget to $60 million and its mailings to $4 million to maintain RJR's leadership.22
As it turned out, the largest threat to RJR leadership in direct mail and database marketing was its primary competitor, Philip Morris, which had been exploring the use of direct mail since 1981,23 partly in response to competitive pressures from RJR.24 A 1981 memo from Herb Foster, Assistant to PM's Executive VP of Marketing, discouraged cooperative coupon mailings such as RJR's early efforts saying trial from these mailings came mostly from ‘coupon clippers’, and group mailings gave the brand a ‘marked down, discount image,’ delivering PM's coupons along with ones for garbage bags and dog food.23 He recommended the system RJR had later adopted: mailings to a targeted list of competitive smokers compiled by an outside service and development of a database for use in future efforts. Estimated cost for name generation was $200 per thousand and total cost for a mailing to 30 million names was $10—$12 million.23
By 1985, a confidential competitive analysis of RJR's database marketing performed for PM from Select+Save Computerised Marketing Technologies estimated that RJR's database had increased from 4 million names in 1983 to 18 million in 1985, gained share from competitors other than PM, and saved RJR money. It reported that use of the list by RJR's market research department, which required 1.5 million names quarterly, had so far saved approximately $10 million in research costs, presumably by reducing costs for identifying subjects for research, and predicted the database programme would “reduce the level of conventional promotional media expenditures in 1986.”25
The report recommended that PM develop a database of 30 million names at a cost of $250 million; $50 million for acquiring names and $200 million for programme execution. It specified the inclusion of 20 million smokers of competitive brands and 10 million smokers of PM brands. In fact, PM was apparently already generating names; a 1985 overview of direct marketing reported over 7 million names had been collected from four different brokers in 1984, with another 4.5 to 5 million names coming from redemption of coupons containing codes identifying redeemers.26 PM's 1993 Direct Marketing Handbook puts the company's entry into direct marketing at 1986.27
In 1989, PM's ‘current status’ report noted that PM's direct marketing had become increasingly more efficient than mass marketing and was the most efficient and effective way of engendering brand loyalty. It estimated that in 1989 PM would mail 9.1 million marketing pieces at a cost of $26 million and that 4.7 million smokers would receive at least one of PM's mailings, with many of them being contacted two to four times.28
Initial databases contained variables recording contact information, brand preference and information to categorise individuals according to likelihood of switching brands. Additional variables were added over time. PM's 1989 ‘current status’ report listed the contents of RJR and PM databases. Both companies collected name, gender, age, address and telephone number, regular brand and flavour and preference for pack or carton. PM had more information on brand used (eg, length, tar level, soft pack or box, length of time smoked and coupon usage) and noted ‘out of stock behaviour’ (what you would do if a store was out of your brand) and share of requirement (percentage of purchases accounted for by your favourite brand)28 both of which served as indicators of propensity to switch brands.29 ,30
More information was to be added to PM's database by mid-1989, including receipt of and response to direct mail offers and lifestyle and activities data including credit card, automobile and home ownership and preferences for music, sports and other activities (presumably for use in targeting but also in ‘developing programme concepts’). Database information allowed development of targeted programmes based on lifestyle and interests of specific audiences, such as a mailing introducing a new brand (Cartier) to competitive female smokers from upscale demographic areas and one to Benson and Hedges movie buffs offering free movie tickets with proofs-of-purchase.28
Variables continued to be added and PM's 1992 Direct Marketing Review Executive Summary listed the database as containing lifestyle information including marital status, ethnic background, education, occupation, hobbies and interests as well as history of response to PM direct marketing.31 By 1994, RJR's database included geography; source of name; recency of update; response history (programmes received and those participated in); usual brand; second usual brand; style, length, flavour; pack/carton; promotion sensitivity, savings brand propensity; lifestyle, interests; and income/occupation.32 ,33
Populating the databases
From the beginning, a major issue for RJR and PM was increasing the number of database names. This was driven both by the need to add names and to increase specific types of smokers. They continued to contract with outside vendors, but also generated names through other strategies. A 1992 report by PM's marketing firm Leo Burnett said PM's competitors typically generated names through branded media, including coupons or mail-in-offers, magazine inserts/ads, telemarketing, direct mail and point of sale ‘take-one’ offers.34 A 1994 document shows PM also used these types of promotions as well as carton and pack inserts, refer-a-friend options and events.35 By 1995, according to written testimony of PM Vice President David R Beran, PM generated names through bar programmes or events, website or 800 number requests and purchasing names of people who indicated an interest in receiving tobacco coupons and promotional materials on a third-party survey.36 Both RJR and PM made special efforts to increase the number of names from under-represented groups, particularly young adults and minorities. This was necessary partly because earlier name generation methods had produced databases that skewed toward older people,37 females and carton buyers.28 To address this, PM's 1989 name generation efforts focused on minimising the skew toward deal-seekers and increasing the number of young adults on the database.38
New strategies included adding ‘front-end name generation devices’ such as a direct response ad offering product samples during Virginia Slims Superslims’ introduction to generate names for further marketing. A ‘smoker survey’ self-mailer in People magazine generated younger female names, and planners considered generating young male names by placing the ‘Smoker Survey’ in non-traditional media such as catalogue inserts and truck stop take-ones. Additional names came from a list of young male names supplied by Omni/Penthouse in exchange for merchandising credits accrued for advertising.28
Database size could vary from year to year based on name generation, rescreening confirming existing and collecting new information, names becoming unusable for various reasons and whether the number reflected only usable names or included those missing essential information (eg, mailing address and brand preference).36 A 1989 presentation by Leo Burnett comparing the size of ‘usable names’ on the PM and RJR database put RJR's at more than 30 million and PM's at 8.2 million.39 By 1992, a Leo Burnett market research report said that all major tobacco companies were conducting direct marketing activities with RJR and PM being the most active in database driven targeted mailings.34 It estimated RJR's database at 25–40 million smokers (noting that information on these might be limited) and PM's at more than 21 million smokers with adequate information for mailing. Those of other cigarette companies were judged to be much smaller. In terms of number of pieces of direct mail, the report stated that at that time RJR produced almost twice as many direct mail pieces as PM.34 PM's database grew to 30 million names in 1994 (figure 1)35 RJR's 1995 Year End State of the Database reported having 22 million names on the database, 13.3 million of them available for use. RJR aimed to increase the number of usable names to 20 million at year's end.40
Perceived advantages of direct mail/database marketing
Both PM and RJR perceived a number of advantages to direct mail/database marketing. A 1984 RJR presentation noted the benefits of targeting, personalised delivery of messages and flexibility in the timing of messages.33 Tracking of consumer response facilitated evaluation of effectiveness, and in general direct mail was known to be ‘well read by recipients.’33 In May 1986 PM noted that direct marketing could be used to build and/or reinforce brand image, generate product trial and conversion and introduce new products as well as defend an established brand's share, and encourage repurchase and larger purchases among current consumers.41 Figure 2 shows the strategic framework for direct marketing programmes from PM's 1992 Direct Marketing Handbook.42
An additional factor driving direct mail development and growth was limitations on other advertising and promotion. One of the earliest (1980) documents promoting direct mail at RJR noted that it was a private means of communicating which, unlike print media, had low visibility, allowing for ‘aggressive promotion opportunities without fear of PR/legislative/retailer whiplash’.13 According to a 1986 presentation to PM, “If traditional advertising is further restricted, Direct Marketing will become an invaluable tool to this corporation.”43 Indeed, a 1989 presentation by Leo Burnett to PM's President and CEO Ehud Houminer said one reason PM started direct marketing was the need for a method of ‘alternative delivery’ of brand images and messages if other means were cut off.44 Direct marketing was unlikely to be restricted for at least two reasons: (1) the major objection to tobacco advertising was that it reached children and non-smokers while people on PM's database certified that they were over 21 and that they were smokers who wanted to receive cigarette samples/offers in the mail, and (2) so far direct marketing had not been mentioned in any proposals for ad restrictions.44
In 1989, Leo Burnett estimated that if direct marketing became one of the few means of communicating brand images “a database of at least 50% of all smokers would be needed to obtain the reach necessary to keep those images alive.”39 A topline database report shows they had achieved this by 1999.45
This is the first study to focus on the development, intent, extent and advantages of tobacco industry direct mail and database marketing, essential components of tobacco industry direct-to-consumer marketing activities for which research has been limited. We found that the top two US tobacco companies (RJR and PM) spent considerable time, energy and money developing direct mail/database marketing capabilities and valued in particular the resulting flexibility, efficiency and unique ability to deliver specific messages to particular groups as well as direct mail's limited visibility to tobacco control, public health and regulators.
Tobacco direct mail marketing using industry databases grew from the need for a targetable promotional strategy to deal with declining smoking rates, growing numbers of product choices and a cluttered media landscape. These factors have become even more important over time. Smoking rates in the USA have fallen further, additional limits on cigarette advertising and promotions have been enacted and the number of cigarette brands and sub-brands and competing products such as snus and e-cigarettes continues to expand. These developments suggest the importance of direct mail and database marketing will continue to grow. In addition, industry direct-to-consumer marketing has now expanded to include digital delivery of brand-appropriate images and incentives through brand-specific websites, direct e-mails and apps on mobile devices. These serve a function similar to that of conventional direct mail and like it rely on information from and provide information for the databases making them even more central to tobacco marketing.
The study also documents the development and contents of industry databases. The industry's ability to segment populations and develop communications tailored to and capitalising on their preferences, attitudes and lifestyle has been noted as a strength of tobacco marketing46 and current segmentation strategies rely heavily on information from the databases. Direct mail promotions and consumer databases can be utilised both to recruit new users or encourage brand switching (offensive) and to encourage continued use among current customers (defensive), to reinforce brand image and cross promote non-tobacco products (such as with branded lifestyle magazines sent to smokers on mailing lists)47 or to conduct consumer research. Our findings document name generation efforts and strategies utilised to collect names and targeting information for the databases, and add to the literature concerning targeting of individuals in priority groups. Previous research has shown direct mail campaigns targeting women,48 young adults,49 hipsters,50 older smokers51 ,52 and low SES females.52 It has also been utilised to facilitate viral marketing of Eclipse cigarettes53 and promote menthol cigarettes to young urban Blacks.54 ,55 Limiting or countering direct marketing efforts that target vulnerable groups may be an important strategy to decrease tobacco use disparities.
There is evidence that direct mail marketing reaches smokers and impacts smoking behaviour. A 2001 study shows one in three adult smokers in New Jersey received direct mail.7 More recent survey research showed self-reported direct mail receipt by young adults56 and youth and its role in smoking initiation and continued or even increased consumption in these groups.57 ,58 In these studies it is not clear if direct mail was addressed to the youth; we did not find evidence of efforts to put youth on industry databases. Tobacco companies have stated they do not to market to youth since the 1970s,59 but youth may circumvent age checks and exposure to direct mail marketing may result from other adults in the household receiving promotional items. Additional research is needed to further examine direct marketing both through traditional and new digital strategies and to establish the prevalence of receipt and exposure to these items and their influence on receivers’ perceptions and smoking behaviours.
Knowledge of direct mail/database marketing has been hampered by its limited visibility, the difficulty of directly observing its influence on receivers’ smoking behaviours, and the lack of existing data establishing its magnitude and importance. More complete reporting of direct mail expenditures (including that carrying coupons, specialty items and other materials) and efforts to acquire direct mailings for review and analysis are needed. Current knowledge about the contents of direct mail comes from surveillance efforts utilising monitors who receive direct mail/email. While valuable, these efforts undoubtedly yield only a limited sample of direct mail/email. The FDA should request disaggregated expenditure data from tobacco companies and require the industry to submit all direct mail/emails including information about distribution, such as date of distribution, number distributed and description of receivers. Access to this information could be used to inform the development of future regulations on these types of communications.
Although it is generally recognised that a comprehensive ban on tobacco marketing is the most effective way of addressing tobacco use, policy options for restricting or banning direct mail and other direct to consumer marketing to adults are limited in the USA, where tobacco companies argue that advertising aimed at adults is protected by the First Amendment to the Constitution.60 Internationally, Article 13 of the WHO Framework Convention on Tobacco Control61 calls for a comprehensive ban on tobacco advertising, promotion and sponsorship, including direct mail. Although many signatories to the FCTC have banned direct mail marketing, others have chosen not to address it or adopted limited bans—such as allowing direct mail addressed to adults.62 Where constitutional barriers do not limit options, growing evidence of the purpose and potential impact of direct mail and database marketing on smoking behaviour should be used to promote increasing limitations on direct mail and database marketing.
Study limitations include the possibility that some relevant documents may not have been found or were not available. The study focused on the development and intentions of direct mail marketing in two companies and primarily utilised documents from the 1970s and 1980s. However, the information on the purpose and process of direct mail/database marketing was consistent through the 1990s and continues to influence tobacco marketing today. Indeed, review of the contents of more current direct mail confirms the continued use of this strategy for essential marketing needs such as reinforcing brand image encouraging switching, building brand loyalty, promoting new products and distributing price-reducing coupons and gifts.2 ,5 ,56 ,63 ,64
More awareness is needed regarding the extent and purposes of direct mail and database marketing and its potential influence on smoking behaviour and therefore health among tobacco control professionals, policymakers, smokers and members of the general public. Given the non-public nature of this marketing strategy educational efforts may be needed to increase awareness of this form of marketing and to alert smokers to its potential to derail quit attempts. Smokers who receive direct mail but are trying to quit smoking should be advised by clinicians and tobacco control programmes to take their names off mailing lists.
What this paper adds
Research on direct mail marketing has been limited and little is known about its extent and role in the tobacco industry marketing mix and still less about the industry's extensive consumer databases that support it. This is the first study to examine the development, purposes and extent of these industry strategies at RJ Reynolds and Philip Morris
Review of industry documents shows the industry valued direct mail/database marketing's flexibility, efficiency, and unique ability to deliver specific messages to particular groups as well as direct mail's limited visibility to tobacco control, public health and regulators.
The study adds to very limited information available to tobacco control regarding database size, name generation strategies and categories of information in industry databases supporting direct mail.
These findings point to the need for research into the prevalence of receipt and exposure to direct mail items and their influence on receivers’ perceptions and smoking behaviours.
Thanks to Mia Hanos Zimmermann for her help in preparing the final version of the paper, Mary Hrywna for her helpful review and comments and Sara Elnakib, Jeanne Garman and Chris Pai for assistance in identifying and reviewing tobacco documents. This research was funded by National Cancer Institute Grants R03 CA-1264330, R21CA-155956 and R01 CA-141661.
Contributors MJL conceptualised the study, identified and reviewed tobacco industry documents and led the analysis, interpretation and writing. PML contributed to conceptualisation, interpretation and writing. MJL wrote the first complete draft of this article. Both authors contributed substantially to the manuscript revision, and have approved the final manuscript.
Funding National Cancer Institute.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.