Article Text

Download PDFPDF

Product retrieval time in small tobacco retail outlets before and after the Australian plain packaging policy: real-world study
  1. Melanie Wakefield,
  2. Megan Bayly,
  3. Michelle Scollo
  1. Centre for Behavioural Research in Cancer, Cancer Council Victoria, Melbourne, Australia
  1. Correspondence to Professor Melanie Wakefield, Centre for Behavioural Research in Cancer, Cancer Council Victoria, 1 Rathdowne Street, Carlton Victoria 3053, Australia; melanie.wakefield{at}


Objective We aimed to assess change in cigarette pack retrieval time in small retail outlets following the introduction of plain packaging in Australia in 1 December 2012.

Methods A sample of 303 milk bars, convenience stores, petrol stations and newsagents was selected in four capital cities, stratified by area socioeconomic status. In June and September (baseline months), the first 2 weeks of December 2012, and February 2013, stores were visited by trained fieldworkers who requested a cigarette pack of a pre-determined brand, variant and pack size, unobtrusively recording the time from the end of the request to when the pack was scanned or placed on the counter.

Results In multivariate analysis, December retrieval time (12.43 s) did not differ from June (10.91 s; p=0.410) or February (10.37 s; p=0.382), but was slower than September (9.84 s; p=0.024). In December, retrieval time declined as days after plain packaging implementation increased (β=−0.21, p=0.011), returning to the baseline range by the second week of implementation. This pattern was not observed in baseline months or in February. Sensitivity analyses showed that results were robust to the variability in purchasing circumstances in tobacco retail outlets.

Conclusions Retailers quickly gained experience with the new plain packaging legislation, evidenced by retrieval time having returned to the baseline range by the second week of implementation and remaining so several months later. The long retrieval times predicted by tobacco industry-funded retailer groups and the consequent costs they predicted would fall upon small retailers from plain packaging are unlikely to eventuate.

Statistics from

Request Permissions

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.


From 1 December 2012, all tobacco sold in Australia was required to be plainly packaged in drab dark brown packs, with 75% front-of-pack graphic health warnings and the brand name and variant limited to a standardised font size and type.1 Plain packaging aims to reduce the attractiveness and appeal of tobacco, reduce the ability of packaging to mislead consumers about smoking harms and increase the noticeability and effectiveness of health warnings.1 The introduction of Australia's plain packaging law was preceded by a vigorous campaign against the policy by the tobacco industry and the Alliance of Australian Retailers (AAR).2 ,3 Industry-funded advertising from June 2010 featured portrayals of concerned retailers saying that plain packaging would not work and would damage their business.2 ,3 The AAR (, accessed 12 January 2013) argued that plainly packaged tobacco products would be more difficult for retailers to locate and retrieve, resulting in longer transaction times and more selection errors as retailers searched to identify brands in packs that would be virtually indistinguishable without their former identifying colours and designs. It was predicted that these longer transaction times would result in long queues in convenience stores and other small tobacco retail outlets, customer frustrations and loss of trade.

The basis for the AAR's claim about these longer service times was a short report produced by Deloitte Australia in February 2011, which attempted to quantify potential impacts on small business, including extra time associated with stock management; increased time taken to complete a transaction; increased product selection errors; security costs; and channel shift from smaller to larger retailers.4 As a basis for cost estimates, Deloitte spoke to six retailers: two operators of service stations/convenience stores, two tobacconists and two newsagents. On the basis of qualitative feedback from this handful of operators, Deloitte concluded that plain packaging would extend the average transaction time to serve customers tobacco by between 15 and 45 s.4 The six retailers also thought plain packaging could lead pack retrieval errors to increase by 5%, providing the potential for increased costs from longer transaction time to obtain the correct pack and replacement of inadvertently opened incorrect stock. A second AAR-commissioned Deloitte report in June 2011, which included focus groups of small mixed business tobacco retailers, projected likely costs of the predicted shift of tobacco business from small to large retailers.5

The results of the Deloitte focus groups and interviews are at odds with the results of a peer-reviewed simulation experiment by Carter et al,2 where 52 adults retrieved a randomly ordered list of cigarette brands from an alphabetised display of either 50 branded packs or 50 plain packs, simulating a total of 5200 ‘transactions’. The study found that packs were retrieved from the plainly packaged cigarette display significantly more quickly (2.92 s) than from the branded cigarette display (3.17 s). Retrieval time reduced for branded and plain packs as the number of transactions increased. Qualitative interviews of the participants suggested that the colours and inconsistent location of brand names on packs in the branded condition often served to distract when participants scanned for brands.

Most recently, a study of four convenience stores by the Rural Shops Alliance (RSA) in the UK, which was funded by British American Tobacco, recorded transaction times and selection errors during 1 week of branded packaging followed by 1 week of ‘plain packaging’, when all packs were shrouded in plain pack sleeves specially produced for the study.6 Over the 2-week period, an unobtrusive video camera recorded a total of 3851 tobacco purchases. The total transaction time—defined as the time from the customer arriving at the retailer counter to when they left the counter—increased from 30 s for branded packs to 58 s for plain packs, but discussions with curious customers about the plain packs accounted for some of the additional time during the plain packaging week of the study. Within the overall transaction period, the total selection (or retrieval) time, defined as the time from a customer instruction for tobacco to the product being placed on the counter awaiting payment, increased from 11 s to 28 s.6 The selection error rate increased from 7% in the branded packaging week to 25% during the plain packaging week, although the overall time added was only 3 s. As with the Carter et al2 study, selection time reduced as the retailers became accustomed to retrieving the plain packs.

Aside from the plain packaging transaction time being over-estimated because it sometimes included an explanation about the study to consumers when they noticed the unfamiliar plain packs, the RSA study used only four retail outlets. Each had agreed to be fitted with hidden, high definition closed circuit television cameras for this study, so their inclusion was not random. Two of the stores already had the cigarette display covered, as is the case in a display ban, while two did not. As all states of Australia have display bans, consumers no longer expect to see cigarette packs displayed in stores, and retailers have become accustomed to retrieving packs from a closed cupboard. Most crucially, in the RSA study, retailers did not use any strategies that would be expected to be used in the real world of plain packaging and as were explicitly used in the Carter et al study2 to discreetly label or alphabetise the packs for their own easy retrieval.

The aim of the present study was to determine the time taken for retailers to retrieve tobacco products before and after implementation of the Australian plain packaging legislation. The study focused on small tobacco retailers, such as those that operate milk bars/delicatessens, convenience stores, petrol stations and newsagents (responsible for a total of about 30% of tobacco sales in Australia),7 which the tobacco industry argued would be most adversely affected by plain packaging legislation.


Store sample and selection

In May 2012, 303 stores were selected from 48 geographical areas, each comprising a primary and secondary postcode, within the metropolitan regions of the major Australian cities of Sydney, Melbourne, Adelaide and Perth. Areas were stratified by socioeconomic status (SES) and randomly selected from each SES quartile.8 For each primary postcode, a secondary postcode in close geographical proximity to the first, which matched the SES characteristics of the primary postcode, was also selected.

On maps of selected primary postcode and secondary postcode areas, all potentially useful shopping centres or strips were identified and ordered by size. Fieldworkers travelled to a pre-determined starting point in the largest shopping location within their primary postcode and, using a rigorous set of walking rules, sampled every potentially eligible store they encountered. Fieldworkers continued sampling until they had at least one of each of the required store types in the sample. If store quotas in the first location were not met, they travelled to the next location and repeated the procedure until that area's store sample quota was filled or all identified locations had been visited. If necessary, the fieldworker went to the secondary postcode and repeated the sampling procedure.

A minimum of six stores per area were required, with at least one of each of the following store types where possible:

  • Milk bars (Melbourne and Sydney) and delicatessens (Adelaide and Perth): small businesses and family-run stores, typically showing signage of ‘milk bar’, ‘deli', ‘corner store’ or ‘general store’;

  • Convenience stores: included small independent supermarkets and branded chain convenience stores (such as 7–11) with a broad range of products and extended opening hours;

  • Petrol stations: any store that sold petrol, including those branded as convenience stores or supermarket chains;

  • Newsagents/lottery outlets: included stores that sold newspapers, magazines and stationery as their primary business and/or lottery tickets and services.

To be eligible, stores were required to sell tobacco and have a visible price board listing at least three cigarette brands. Specialist tobacconists and large supermarkets were excluded. Eligible stores that were sampled before the store type quotas were filled were also included in the final sample. Three-quarters of the sample (n=229) were found in the primary postcode.

A total of 16 areas in each of the large cities of Melbourne and Sydney were sampled and 8 in each of the smaller cities of Adelaide and Perth. The areas were evenly spread across four SES categories in each state. This panel of stores was revisited for the purpose of recording cigarette pack retrieval time from 20 June to 2 July 2012 and from 6 to 24 September 2012, both prior to plain packaging implementation, and from 3 to 12 December 2012, which fell during the first 2 weeks of implementation of plain packaging, and again from 14 to 28 February 2013. Stores that were temporarily or permanently closed were not replaced in subsequent visits.

Measurement of cigarette retrieval time

Retrieval time was measured as part of a broader store audit study, in the first stage of an attempt to purchase the cheapest pack of tobacco available in the store. In each store, the fieldworker approached the counter and consistently requested one of three particular brands in a common variant and small pack size from the so-called ‘value’ market segment of factory-made cigarettes. At the moment they finished this request (ie, as they said the pack size) and discreetly began recording the time. When the retailer placed the requested pack on the counter or scanned its barcode (whichever came first), the fieldworker stopped recording. We chose brands that we believed were likely to be stocked by most stores (as they were often listed on tobacco price boards), without being the main market leaders that retailers had most practiced retrieving.

Several strategies for time recording were consistently used. Some fieldworkers used a stopwatch, or mobile phone with a stopwatch feature, concealed in their pocket or bag. Others used a voice recorder, or phone with this feature, and activated it before they went into the store. After exiting the store, they reviewed the recording and timed the duration using a stopwatch. Fieldworkers who used this approach were careful to ensure that they made an audible marker (such as saying ‘thanks’ when the pack was placed on the counter) to signal when they should finish recording. After exiting the store, fieldworkers also noted whether any interruption or other extraneous factor occurred during the procedure. Fieldworkers rotated their assigned areas in each of the months to reduce the likelihood that retailers would remember them, so as to preserve the unobtrusive nature of the study.

Fieldworker training

A large national fieldwork agency experienced in in-store research was contracted to undertake data collection. Fieldworkers (n=23) ranged in age from their 20s to their 60s with the majority aged in their 40s and 50s, and 61% were males. Prior to the May 2012 store selection, fieldworkers in each capital city undertook a 2–3-hour training session led by a senior fieldwork coordinator from the agency and our study coordinator (MB), focusing on the store selection procedure and store eligibility criteria, using mocked-up price board examples. Prior to the first month of measurement of retrieval time, fieldworkers again attended a 60–90-minute training session, which detailed retrieval time methods and measures and illustrated using modelling. Fieldworkers were able to practise these methods in unselected stores prior to their main data collection.

Statistical analysis

Data analysis was undertaken using Stata S.E. V.12.1. Two outcomes were coded that precluded the measurement of retrieval time: (1) the fieldworker did not complete the pack request correctly, requiring the retailer to ask which brand variant or pack size they wanted, or the fieldworker mishandled the recording device; or (2) the requested product was not in stock. For all other visits, retrieval time was assessed and four outcomes were coded to reflect extraneous factors that would have likely influenced retrieval time: (i) automatic dispenser used to retrieve the requested pack; (ii) shopkeeper did not start looking for the pack immediately, but chatted to the fieldworker or interacted with another customer first; (iii) the shopkeeper requested age identification; and (iv) equipment-related delays, such as a barcode scanner malfunction. Other instances that were related to the pack searching procedure were deemed to be part of usual retrieval time variability, including trainee shopkeepers, other staff members helped to locate the requested pack, if the shopkeeper erred during pack selection or where the shopkeeper asked for clarification about the requested variant or pack size (even though the fieldworker had originally specified this information as per the study protocol).

Comparisons between proportions of binary outcomes in stores across months were undertaken using logistic regression analysis, adjusting for store type, city and area SES. Wald χ2 tests were used to examine main effects of covariates with 3 categories or more. Analysis of covariances (ANCOVAs) were performed to compare mean retrieval time across months and within months across store types and to produce adjusted means. These also adjusted for store type, city and area SES. Bonferroni-adjusted post hoc pairwise comparisons were performed to examine differences in adjusted means between subgroups. Differences in the incidence of extraneous factors affecting the measurement of retrieval time across months were examined with χ2 tests.


Store characteristics

A total of 303 stores comprised the sample identified in May 2012. An even spread of SES categories was achieved with a range of store types (table 1).

Table 1

Characteristics of the original store sample, May 2012

The number of these stores that were open, still selling tobacco and were visited for the measurement of retrieval time was 300 in June, 300 in September, 297 in December and 295 in February.

Outcome of cigarette pack requests

Retrieval time was unable to be measured in 21.7% of stores in June, 9.3% of stores in September, 35.7% of stores in December and 27.8% in February, and reasons varied by month (table 2). Fieldworkers failed to correctly request the brand variant or pack size or mishandled their timing device most frequently in June, but this substantially reduced in later months as more practice was gained with the procedure. The requested product was not in stock in 35.0% of stores in December and 22.4% of stores in February, compared with 11% or less in the two baseline months. Because this occurrence was higher post-plain packaging, we further examined the characteristics of stores where the requested pack was not in stock for each month using logistic regression analyses. In December, a significant main effect of day of month was found, where the odds of encountering an out-of-stock store decreased by 18.4% for each additional day (OR=0.82 (95% CI 0.70 to 0.95), p=0.008). There were also significant main effects of city (χ2=12.81, p=0.005) and area SES (χ2=16.54, p<0.001). Bonferroni-adjusted pairwise post hoc comparisons showed that Sydney stores (25.0%) were significantly less likely to be out of stock than stores in Melbourne (47.1%; p=0.004). Stores in low disadvantage areas (20.8%) were less likely to be out of stock than those in the least disadvantaged areas (51.9%; p<0.001). In February, significant main effects were found for area SES (χ2=13.49, p=0.004) and store type (χ2=11.18, p=0.011). A trend towards an effect of city was also found (χ2=6.83, p=0.078), as was a trend towards out-of-stock stores being less common as day of month increased (OR=0.93 (95% CI 0.86 to 1.01), p=0.077). Bonferroni-adjusted post hoc comparisons showed that, in February, stores in areas of moderate disadvantage (9.0%) were less likely to be out of stock than stores in areas of least (30.8%; p=0.010) and most disadvantage (34.8%; p=0.006), while milk bar/delicatessens (11.4%) were less likely than petrol stations (21.1%; p=0.032) to be out of stock, and marginally less likely than newsagent/lottery outlets (38.4%; p=0.077). No pairwise differences between cities were found in February. No effects of day or other covariates were found for June or September.

Table 2

Outcome of store visits by month

A few extraneous factors were noted in a minority of stores in each month that would have influenced retrieval time, and these cases were excluded from the main analysis. Automatic dispensers and store equipment malfunctions were rare and did not differ significantly across months. Unrelated interruptions to the retailer in the midst of cigarette pack retrieval were uncommon, particularly in December. A few instances of fieldworkers being asked for age identification occurred only in June.

The time for the retailer to manually retrieve the requested cigarette pack was measured without extraneous factors present in 72.7% of stores visited in June, 84.3% of stores in September, 61.3% of stores in December and 67.1% of stores in February.

Cigarette pack retrieval time

An ANCOVA adjusting for store type, city and area SES showed a significant effect of month (F(3,838)=2.84, p=0.037) on retrieval time. Bonferroni-adjusted post hoc pairwise comparisons found that after adjusting for store and area covariates, following plain packaging introduction in December, the mean adjusted retrieval time (December: 12.43 s) did not differ significantly from June (10.91 s; p=0.818), but was slower than that observed in September (9.84 s; p=0.024). The adjusted average pack retrieval time observed in February was not significantly different to that in any other month. A significant main effect of state was also found (F(3,838)=10.59, p<0.001), where adjusted average pack retrieval times were significantly faster in Melbourne (8.63 s (95% CI 7.62 to 9.63)) compared with Sydney (11.58 s (95% CI 10.63 to 12.53); p<0.001), Adelaide (13.51 s (95% CI 11.89 to 15.12); p<0.001) and Perth (11.28 s (95% CI 9.71 to 12.85); p=0.031).

There were no significant main effects of month within any of the store types. However, it can be seen from table 3 that pack retrieval duration was highest in December for convenience stores, petrol stations and newsagent/lottery outlets. Within each store type, pack retrieval time in December was compared with each other month using ANCOVA. Cigarette pack retrieval time did not change significantly in the month of plain packaging introduction compared with any other month for milk bars/delicatessens or petrol stations. Adjusting for city and area SES, pack retrieval times for convenience stores were significantly slower in December than September (F(1,91)=4.70, p=0.033), but not different from June or February. Likewise, adjusted mean pack retrieval time in newsagents in December was significantly slower than in September (F(1,102)=5.55, (p=0.020), but December did not differ from June or February.

Table 3

Adjusted average cigarette pack retrieval time by month and store type

We further examined whether day of fieldwork influenced retrieval times within each month. An ANCOVA adjusting for store type, city and area SES showed that there was a main effect of day in December (F(1,171)=6.68, p=0.011). Linear regression showed that pack retrieval time decreased as the number of days since the mandatory introduction of plain packaging on 1 December increased (β=−0.21, p=0.011). In fact, by comparison with days 1–3 (3–5 December, 16.03 s), retrieval time in days 7–10 (9–12 December, within the second week of plain packaging) had halved to just 8.15 s, a figure no different from either baseline months. A trend towards an effect of day was also found for the first baseline month (F(1,207)=3.50, p=0.063), although this was in the opposite direction to December (β=0.14, p=0.063). By comparison, no differences by day were found in September or February.

Sensitivity analyses

In a sensitivity analysis, we included all cases where retrieval time had been recorded but had been compromised by extraneous interference. Overall, this increased the total number of stores to 235 in June, 272 in September, 191 in December and 213 in February. An ANCOVA adjusting for store type, city and area SES showed a similar pattern of results, although the main effect for month on retrieval time was no longer significant (F(3,898)=1.94, p=0.122). Again, December (adjusted mean time=13.08 s (95% CI 11.33 to 14.83) was slower than all other months, particularly September (adjusted mean=10.60 s (95% CI 9.59 to 11.60)). Again, June (adjusted mean=11.94 s (95% CI 10.81 to 13.06)) and February (adjusted mean=11.73 s (95% CI 10.20 to 13.26)) showed similar pack retrieval times. In this sensitivity analysis, the pattern of findings for store types also remained the same. Likewise, the same pattern of association between day of month and retrieval time persisted for December (F(1,180)=5.65, p=0.019) and became significant for June (F(1,224)=5.97, p=0.015). The direction of these effects was as before: retrieval time decreased with number of days in December (β=−0.19, p=0.018), but increased in June (β=0.17, p=0.015).

A second sensitivity analysis was performed using only those stores that were able to be visited every time, where pack retrieval times were able to recorded, and where no extraneous delays occurred in all 4 months (76 stores in total). Adjusted average pack retrieval times were again calculated for each month, controlling for state, area SES and store type. The pattern of adjusted average retrieval times was exactly the same as seen in table 3 and the earlier sensitivity analysis: December was slowest at 12.49 s (95% CI 9.65 to 15.32) and September was fastest at 9.77 s (95% CI 8.15 to 11.39). Adjusted average retrieval times were similar for June (adjusted mean=10.68 s, 95% CI 9.19 to 12.17) and February (adjusted mean=10.83 s, 95% CI 8.82 to 12.85). Finally, the effect of day on pack retrieval time was still significant in December (F(1,65)=4.80, p=0.032) and in the same direction (β=−0.29, p=0.032). This effect was no longer significant for June.


We found the introduction of the Australian plain packaging legislation to be associated with, at most, a modest 2- to 3-s increase in retrieval time, from 10.91 and 9.84 s in the two baseline months to 12.43 s in the first few weeks of implementation of the policy. The December data collection commenced just 2 days after implementation on 1 December, so as to span a period when the most retrieval difficulty might have been expected. It was notable that—unlike the baseline months or in February—cigarette pack retrieval time in December declined as the days of fieldwork progressed and experience with plain packaging increased, so that within the second week of plain packaging implementation, retrieval time had returned to the baseline range and remained so in February. Sensitivity analysis suggested that our results are robust to the variability in circumstances at the point of sale in which cigarettes are purchased.

Consistent with two studies that only simulated plain packaging,2 ,6 our study of the policy in vivo confirms that there is a substantial practice effect, with retailers becoming accustomed relatively quickly to the new appearance of the packs during the first week of its introduction. Our results provide a slightly different picture than a recent note of a study, which focused on retrieval times in 100 stores in Perth in October (baseline) and again in January (after plain packaging) and reported faster retrieval times overall and a decrease in times between these 2 months.9 The faster retrieval times overall in the Perth study compared with ours may in part be due to the inclusion of supermarkets that have a dedicated tobacco counter and could be expected to have faster retrieval times than the average convenience store. Since plain packs began appearing in stores during October, the baseline retrieval time in the Perth study was likely not assessed in a retail environment where 100% of the packs were branded. The confusion caused by a dual system operating in a number of stores could have contributed to a somewhat higher-than-expected retrieval time in those stores in the baseline period in the Perth study.

Our study provides a long-term picture of the ‘real world’ effect on retail serving time in small tobacco outlets in Australia. The stores in the study included the main types of small retailers and spanned four major capital cities and a range of neighbourhoods in these urban locations, giving confidence that the observed effect was widespread. The increase in requested packs not in stock in December suggests that distributors were slow to fully resource small retailers with plainly packaged stock, but instances of stock unavailability significantly declined over the days of fieldwork.

In almost all cases, retailers did not rely on automatic dispensers, but continued to retrieve cigarette packs themselves from behind cupboard doors or screens already present due to existing display ban legislation. Retailers were already accustomed to retrieval of packs from behind cupboard doors; an audit of tobacco retail outlets has shown that compliance with the requirements of the display ban law has been high in Australia10 as in other countries.11 ,12 The Australian plain packaging legislation requires that the brand name and variant be printed in a specific font style and the size large enough to be seen by retailers.13 The font size and type used on Australian plain packaging—Lucida Sans font size 14 for brand names and no larger than 10 point for variant names—was developed after careful readability test research.14 Retailers were free to alphabetise packs behind the doors or use other discreet labelling methods to facilitate their own easy identification of brands.

A number of design features of our study meant that we likely over-estimated retrieval time, in order to purposefully assess what might be a ‘worst possible’ scenario for retailers. We conducted fieldwork early during plain packaging implementation when most retrieval difficulty might have been expected, finding that retrieval time returned to its usual range very rapidly during the second week of implementation and remained so in February. Also, we requested brands that were not the main market leaders, which retailers had the most practice retrieving, so retrieval times might have been quicker in all months if we had requested the most common brands or a wider range of brands including the most common brands (the approach adopted in the Perth study which likely also contributed to the lower average retrieval times in that study).

Studies that employ point-of-sale audit methods and interactions with shopkeepers using a real-world random sample of stores can be challenging.15 Our study had some limitations, but also considerable strengths. First, despite attending training sessions, fieldworkers did not complete the request correctly in 11% of stores in the first baseline month of June, reducing the number of stores in which retrieval time could be measured. This percentage declined substantially by the next baseline month of September as they gained practice in fully requesting the cigarette brand, variant and pack size and recording the retrieval time. It was a virtue of the study that we had two baseline months prior to plain packaging introduction to allow for these practice effects.

Second, only one cigarette pack was requested per store each month. More requests could have been made if the fieldworker revisited each store many times, but this would have increased fieldwork costs. On the other hand, the large number and geographical spread of sampled stores were greater and more representative by far than any previous study.

Finally, we focused on the measurement of retrieval time and did not attempt to count selection errors. Retrieval time is the main underlying metric used in industry-funded studies that attempt to estimate the costs of plain packaging to retailers.4 The Carter et al2 study found an extremely low incidence of selection errors when expressed as a percentage of transactions for branded packs (1.5% of all branded pack transactions), which significantly declined with plain packaging (0.4% of all plain pack transactions). Similarly, the Perth retrieval time study noted selection errors of only 3% and 2% in both months tested.9 In the RSA study, selection errors added only 3 s to the entire transaction time on the occasions they were made, but overall, since corrective action was quick and errors were not particularly frequent, the report concluded that ‘selection errors do increase the overall store's average of transaction times, but only marginally’(ref. 6, p. 11). Therefore, we designed our study protocol to prioritise the measurement of overall retrieval time and included any instances where the wrong pack was first selected and then corrected as being within the overall retrieval time.

Our ‘real world’ findings shine a light upon the poor predictive validity of the qualitative, survey and simulation studies produced by the weak and incompletely described research methods of tobacco industry-funded consultancy groups,4 ,6 in contrast to the findings of more carefully conducted peer-reviewed independent research.2 One may recall that during introduction of smoke-free laws in restaurants and bars in several jurisdictions internationally, weakly designed tobacco industry-funded studies were used in media promotion by industry groups opposing legislation to predict massive costs to business and loss of trade, while peer-reviewed studies using strong objective outcome measures after implementation found no evidence of negative financial effects.16–18 Similarly, tobacco display bans at the point of sale have been vigorously opposed by tobacco companies,19 ,20 yet are readily complied with10–12 and have reduced both exposure to tobacco marketing and perceived prevalence of smoking, one of the normative factors for smoking uptake,21 without prompting business losses or closures.22

Our study shows that small retailers serving customers have been able to quickly adapt to serving plainly packaged tobacco. We conclude that the tobacco industry's claims that cigarette pack retrieval times would double6 or extend by a further 15–45 s4 have not proved to be accurate among small tobacco retailers in Australia, even in the very early implementation period of plain packaging. It appears that the catastrophic costs that tobacco industry-funded retailer groups predicted would fall upon small retailers are also unlikely to eventuate.

What this paper adds

  • Tobacco industry-funded retailer groups have vigorously asserted that plain tobacco packaging would make it difficult for small retailers to identify cigarette brands and variants, resulting in much longer tobacco product retrieval times, inconvenience for customers and ultimately, loss of trade as frustrated customers went elsewhere. These predictions contrasted with a peer-reviewed experimental study in a simulated store, where plain packaging was found to reduce retrieval time, compared with a fully branded cigarette display.

  • Using unobtrusive recording methods in 303 small tobacco retail outlets in four capital cities, we found the implementation of plain packaging in Australia on 1 December 2012 produced a modest 2- to 3-s increase in tobacco product retrieval time, compared with two baseline months. However, retailers quickly adapted to the new plain packs, returning to the baseline retrieval time range during the second week of implementation and remaining so several months later.

  • Tobacco industry-funded retailer groups have vastly over-estimated the impact of plain packaging on retrieval time and, therefore, the consequent costs they predicted would fall upon small tobacco retail outlets.


View Abstract


  • Contributors MW and MS co-designed the study and MB coordinated fieldwork and undertook data analysis. All authors contributed to manuscript writing and revisions and approved the final version of the paper.

  • Funding This study was supported by Quit Victoria, Cancer Council Australia, Cancer Council South Australia, Cancer Council Victoria, Cancer Research UK, Fresh, Smoke-free Southwest, Tobacco-Free Futures and Action on Smoking and Health (UK).

  • Competing interests None.

  • Ethics approval Institutional Research Review Committee, Cancer Council Victoria.

  • Provenance and peer review Not commissioned; externally peer reviewed.