IQOS point-of-sale marketing: a comparison between Arab and Jewish neighborhoods in Israel

Background: Philip Morris International’s (PMI) IQOS, with its heatsticks (HEETS), is the heated tobacco product with the largest global market share. IQOS and/or electronic cigarettes use rate is higher among Arabs vs. Jews in Israel. This paper aims to compare IQOS point-of-sale (POS) marketing strategies, and regulatory compliance in Arab vs. Jewish neighborhoods in Israel. Methods: We integrated data from two separate studies including a cross-sectional survey with IQOS retailers (December 2020-April 2021) and audits of POS that sold IQOS/HEETS (April 2021-July 2021) in 5 large cities in Israel, after marketing restrictions including a POS display ban and plain packaging became effective in Israel (January 2020). The survey included 69 POS (21 Arab, 48 Jewish neighborhoods) and the audits included 129 POS (48 Arab, 81 Jewish neighborhoods). Comparisons of IQOS marketing strategies between POS in Arab and Jewish neighborhoods were conducted using Chi-Square test, Fisher’s exact test or Mann-Whitney test, as appropriate. Thematic analysis was used to analyze open-ended questions. Results: Most marketing strategies, such as promotions to customers, were uniform across POS in Arab and Jewish neighborhoods. The most noteworthy differences were that a higher proportion of retailers from Arab neighborhoods were invited to IQOS parties (47.6% vs. 21.7%, p<0.05) and reported personal communication with a PMI representative (80.0% vs. 51.2%, p<0.05). Additionally, PMI representatives assisted POS in both Arab and Jewish neighborhoods in implementing the display ban by providing free compliant cabinets and product placement instructions, and directly interacted with customers. POS in Arab neighborhoods were more compliant with the display ban (25.5% vs. 8.8%, p<0.05), but less compliant with plain packaging (62.5% vs. 79.3%, p<0.05). Conclusions: There were not many notable differences in IQOS marketing across POS in Arab vs. Jewish neighborhoods, but PMI utilized marketing elements of cultural significance, especially for POS in Arab neighborhoods, such as more personal communication and invitation to social events. Continuous surveillance of tobacco POS marketing and legislation compliance is needed, with a special focus on demographic/location-based differences.


Background
IQOS is a heated tobacco product (HTP), composed of an electronic device that heats tobacco sticks called HEETS, and is manufactured by Philip Morris International (PMI).IQOS was rst launched in Japan in 2014, and is currently sold in more than 70 countries, dominating the HTP global market. [1,2] [12][13] Some tobacco companies pay for dedicated counter spaces or supply the POS with free display cases. [10]16] POS marketing might be especially important for new products like IQOS.The way a new product like IQOS is promoted to retailers at POS, and their attitude towards it, might affect their direct-to-consumer approach. [5]ere is evidence to suggest that tobacco companies might use distinctive marketing strategies to differently target POS in speci c areas, for example; in neighborhoods representing greater proportions of speci c ethnic groups. [17]A 2015 systematic review revealed PMI's use of an "Integrated Retail Demographic Database Micro-Marketing Tool", which utilized data gathered from demographic census and retail pricing to customize campaigns and offerings, including price promotions, directed at speci c POS in certain areas. [18]Several studies have found more tobacco marketing and lower cigarette prices in neighborhoods with lower socioeconomic status (SES) and large proportion of ethnic minority individuals; [19][20][21][22] however, other studies have documented no clear associations. [23,24] n the United States (US), menthol cigarettes marketing was more prevalent in neighborhoods with large proportions of African American residents, while smokeless tobacco was mainly advertised in predominantly White neighborhoods. [20]rael is a predominantly Jewish country, with the Arab population being its largest ethnic minority group (21.1% of the total population). [25,26] ore than 80.0% of Arabs reside in all or majority-Arab localities, 95.0% of which are of low SES. [27]Tobacco and nicotine use is higher among Arabs than Jews (24.4% vs. 19.1% for cigarette; 2.8% vs. 1.2% for IQOS and/or electronic cigarette). [28]IQOS rst entered the Israeli market in 2016 and has been regulated the same as all other tobacco and nicotine products since 2017.[25, 26, 29] Among other measures, this includes an advertisement ban in TV, radio, digital media and POS (2019), a POS display ban where all tobacco and nicotine products should be concealed at all times (2020), and plain packaging requirements for all of these products (2020). [25,26,29] I previous Israel-based study, [3] concealed POS audits were carried out at 80 POS in 2019 and 2020 to assess marketing materials and regulatory compliance before the POS display ban and plain packaging went into effect (January 8 2020); [3] IQOS/HEETS marketing materials and price promotions were uncommon, but IQOS/HEETS special displays were found at 20% of the audited POS.[3] However, this study only included a small sample of POS in Arab neighborhoods (n = 5), and therefore could not compare whether marketing strategies differed between POS in predominantly Arab vs. Jewish neighborhoods.
This study aims to assess and compare IQOS marketing strategies to and at POS, regulatory compliance and retailers' attitudes towards IQOS between POS in Arab vs. Jewish neighborhoods in Israel.

Study design
This manuscript integrates data from two sources: 1. Cross-sectional survey among retailers at POS that ever-sold IQOS/HEETS (December 2020-April 2021).
For each data collection, POS in Jewish neighborhoods were randomly selected from the IQOS Israel website, and all POS in Arab POS were included in the sample (but data from the survey and audits were not matched). [5]oss-Sectional Survey

Settings and procedures
This study examined IQOS marketing strategies at POS in Israel via a phone survey with owners or managers of POS that ever sold IQOS/HEETS in ve large cities in Israel: Haifa, Jerusalem, Tel Aviv (all have both Jewish and Arab populations), Beer Sheva (a predominantly Jewish city), and Nazareth (a predominantly Arab city) in December 2020 to April 2021. [5]The initial sample included the total of 713 POS (n= 86 in Arab neighborhoods and n=627 in Jewish neighborhoods) across these cities, which were then called by phone to participate in the survey.This approach yielded an insu cient number of POS in Arab neighborhoods (n=5) to conduct any comparisons (n=38 POS in Jewish neighborhoods).Thus, additional recruitment efforts included attempting in-person survey data collection by research assistants (RAs) in a matched sample (by city and SES) of POS in Arab and Jewish neighborhoods that could not be reached by phone.
Of the overall n=314 POS successfully contacted/visited (n=76 in Arab neighborhoods, n=238 in Jewish neighborhoods), surveys were completed in a nal sample of n=69 (n=21 in Arab neighborhoods, n=48 in Jewish neighborhoods).Those that were contacted/visited but did not result in completed either: 1) never sold IQOS/HEETS (n=82 overall; n=16 in Arab and n=66 in Jewish neighborhoods); 2) did not consent (n=163 overall; n=40 in Arab and n=124 in Jewish neighborhoods).Retailers (either POS manager or owner) who completed the survey were compensated with a 100 New Israeli Shekel (NIS) online voucher.

Data collection tool
The survey assessed: 1) POS characteristics (i.e., type of store, belonging to a chain, neighborhood SES); 2) participant characteristics (i.e., age, sex, job position at the location, cigarette use status, IQOS use status); 3) number of HEETS avors sold [<4, 4-5, or all 6 HEETS avors], and presence of a special display for IQOS/HEETS; 4) marketing strategies directed at the POS (e.g., free HEETS samples for the retailer's personal use, incentives for sales) with an option to freely elaborate on each item; 5) marketing strategies directed at the customers including promotions (e.g., free HEETS samples for the customers, price discounts) with an option to elaborate on each item, and advertisements (i.e., if the retailer ever promoted IQOS/HEETS online, via social media, via print media, or inside the POS, recoded into yes/no for each and later recategorized to "any form of ads"); 6) a question assessing how retailers would communicate with customers about IQOS and/or HEETS ("How would you describe the IQOS/HEETS to your customers who might ask about your tobacco products or IQOS?") with 9 separate check boxes (e.g., "IQOS is less harmful compared to traditional combustible cigarettes"); 7) an additional open-ended question asking them to describe their personal attitudes of IQOS/HEETS; 8) interactions with a PMI salesperson (the manufacturer of IQOS) such as providing direction on product placement, target market, how to communicate with customers, and providing information about IQOS/HEETS in comparison to other tobacco products (recoded into yes/no for each and with an additional variable of "any interaction"); and 9) PMI's reactions to the new tobacco legislation (i.e., retailers were provided education on the new legislation; free cabinets, etc.).

Settings and procedures
From April to July 2021, trained research assistants conducted concealed, in-person POS audits among POS that currently sold IQOS/HEETS in the same large cities (Beer Sheva, Haifa, Jerusalem, and Tel Aviv, and Nazareth).The nal total sample was 129 POS (48 Arab and 81 Jewish neighborhoods).Audits were conducted using a validated surveillance tool, developed based on the Standardized Tobacco Assessment for Retail Settings (STARS), and adapted for IQOS [3]

Data collection tools
The audit tool assessed: 1) POS characteristics (e.g., type of store, belonging to a chain, neighborhood SES); 2) marketing materials inside or outside the POS such as IQOS/HEETS special display, QR code, or signage for products sold (coded as "any ad" if marketing materials were found for any product); 3) price promotions; 4) visibility of the sold products (sold and visible, sold and not visible, or not sold); 5) prices (i.e., least expensive price of a cigarette pack and HEETS) and any price promotions across products; 6) placement (i.e., within 30 cm of toys or candy, and/or within one meter of the oor); and 7) regulatory compliance (e.g., presence of minimum age signage, presence of a "no smoking" sign, all tobacco and nicotine products in plain packaging, and all tobacco and nicotine products are completely covered and not visible; if a product was visible the POS was coded as noncompliant).

Data analysis
Descriptive analysis was conducted using counts and percentages (%) for categorical variables and mean (SD) for continuous variables (e.g., age, product price).Bivariate analyses were conducted using Chi-Square test, Fisher's exact test or Mann-Whitney test, with Bonferroni correction as appropriate, to assess the differences between POS in Arab and Jewish neighborhoods.For all analyses, SPSS v27 was used, and a p<0.05 was considered statistically signi cant.Open coding and thematic analysis were used by the lead researcher (AK) to code and analyze free-text and open-ended questions, which were validated by a senior researcher (YBZ).

Results
Cross-Sectional Survey POS and participant (manager/owner) characteristics Table 1 summarizes the POS and participant (manager/owner) characteristics from the retailers' survey, overall and comparing retailers at Arab vs. Jewish neighborhoods.The surveyed POS were mainly grocery stores (37.7%, n=26) or convenience stores not within a gas station (34.8%, n=24).A higher proportion of POS in Arab neighborhoods were grocery stores (71.4% vs. 22.9%, p=0.001), and all of them (100.0%)were located in low-and medium-SES neighborhoods, compared to 64.6% of POS in Jewish neighborhoods (p=0.009).2).Only retailers in Arab neighborhoods mentioned receiving lighter stands (n=2), ights abroad (n=1), swimming competitions (n=1), and receiving points when referring customers (n=1).More than half of all POS carried IQOS promotional materials (Table 2); these were mainly electronic and/or non-electronic signs that said "here you can buy heated tobacco products" or "heated tobacco units", small ags that say the same or advertise a price promotion, and/or special display cases for IQOS/HEETS.A few retailers mentioned that PMI sent saleswomen to set up a small stand and promote IQOS directly to customers (6 in Arab neighborhoods and 3 in Jewish neighborhoods).

Open-ended questions
Retailers from POS in Arab neighborhoods emphasized their part in promoting the product to their customers and connecting the customers with sales representatives.For example, one participant used her personal experience as a promotional strategy: "When people come to my shop and see me use it they get curious and start asking me about it, I tell them about my personal experience and how I used to smoke Marlboro but when I switched to IQOS I stopped coughing in the morning and it doesn't stink your clothes or furniture".
Another participant stated that the POS was acting as a "middle man" by connecting the customer with a PMI sales representative: "the shop was the intermediary; the company's representative asked us to connect him with the customers if anyone asks about IQOS or was interested in trying it", and mentioned collecting ID numbers and phone numbers to register customers for a user database, and received points for each person.Others referred to the representative's direct interaction with customers; "the representative and I try to tell customers about IQOS, that it can meet their requirements and is less harmful and smoke-free".

Retailers' attitudes towards IQOS and interactions with PMI representatives
More retailers from POS in Arab neighborhoods stated that IQOS is an e-cigarette (61.9% vs. 27.1%,p=0.006) and found its avors to be appealing (42.9% vs. 12.5%, p=0.009) ( gure 1: retailers' attitudes towards IQOS).Overall, 42.0% of retailers stated that IQOS is less harmful compared to cigarettes (43.8% and 38.1% among retailers from POS in Jewish and Arab neighborhoods, respectively).3).This included, for example, being given instructions on how to arrange the products behind the cover to make it easier to access and sell them, being directly informed about new campaigns and promotions, and repeatedly information about the products.

Marketing materials, prices and price promotions
The vast majority of POS (79.8%, n=103) had internal and/or external ads for any tobacco or nicotine product (72.9%Arab, n=35 and 82.9% Jewish, n=68); of which more than half was IQOS-indirect internal signage (57.1% Arab, n=20/35 and 57.4% Jewish, n=39/68), such as signs that said "heated tobacco units", or "here you can buy heated tobacco".The majority of POS that had any IQOS/HEETS signage included HEETS brand colors (70.0%Arab, n=14/20 and 79.5% Jewish, n=31/39), and IQOS/HEETS special displays (55.0%Arab, n=11/20 and 71.8% Jewish, n=28/39) (Table 4).Some of the special display cases provided by PMI were "discreet"; they had a light switch that makes the product visible only when turned on ( gures 2A and 2B: IQOS special displays in Arab and Jewish neighborhoods, respectively).Placement, visibility and regulatory compliance IQOS/HEETS were placed within 1 m of the oor only in POS in Jewish neighborhoods (12.2%, n=10).
IQOS was highly visible in the POS that sold it (68.8% in Jewish and 66.7% in Arab neighborhoods), but the visibility of HEETS was higher among POS in Jewish neighborhoods (62.2% vs. 39.6% in Arab neighborhoods, p=0.010).
A signi cantly higher proportion of POS in Arab neighborhoods had products in their original packaging (i.e., not in plain packaging as required by law) compared to Jewish neighborhoods' POS (37.5% vs. 20.7%, p=0.027), but a higher proportion were compliant with the display ban (25.5% vs. 8.8%, p=0.011).

Discussion
Findings from these studies show that, in general, PMI employed similar marketing strategies at POS in Arab and Jewish neighborhoods, without clear speci c targeting.Nonetheless, a more personalized marketing approach (personal communication, social events) was more prevalent in POS in Arab neighborhoods.
Our ndings of no over-targeting of POS in Arab neighborhoods compared to Jewish ones is in alignment with our previous study that explored PMI's marketing in print media across population groups and media outlets in Israel. [25,26,29] H[21][22] This might be attributed to IQOS being a relatively new product in the market, which would warrant a focus on the majority population in order to increase market share.However, it is surprising that despite PMI's claims that this product is intended for adult smokers, [26, 30] there seems to be no apparent targeting of Israel's largest smoking population -the Arab population. [26,28] e use of different marketing strategies for POS in Arab or Jewish neighborhoods, such as more personal communication with retailers and invitation to social events could be attributed to a form of close network marketing that might be due to cultural differences.It has been suggested that in the Arab population, business relationships might be heavily in uenced by personal ties. [31]In addition, this might re ect the presence of different marketing teams (Arab and Jewish) to carry out in-person communications and outreach activities.
The majority of surveyed retailers had a positive attitude towards IQOS/HEETS.These perceptions are of great importance because they might in uence how retailers communicate with customers and could be the result of PMI's marketing efforts directed at the POS.While some retailers talked about themselves becoming salespeople or intermediaries who connected potential customers with a PMI representative, retailers in Arab neighborhoods were more inclined to use their personal experiences using IQOS to increase their credibility when promoting the product to customers.The majority of POS in Arab neighborhoods were small businesses (i.e., grocery stores; 71.4%), owned and operated by community members who can promote IQOS sales by in uencing their customers through personal relationships, cultural cohesion, and embedded trust. [31]r ndings suggest that the POS were used by PMI as a tool to directly market IQOS to consumers, by sending saleswomen to set up small IQOS stands inside the POS, and allowing the representative the freedom to talk to customers in order to promote the product, both of which were more prevalent at POS in Arab neighborhoods.The use of saleswomen is a newer form of what was known in the past as "cigarette girls", which was previously used by PMI's branch in Australia in 2000. [32]I's active involvement in helping retailers implement the display ban might have also contributed to the retailers' positive perception of IQOS, especially with signage that serve as constant reminders of the product.Research from Scotland also highlighted the tobacco industry's role in helping POS implement display bans and how to work around them. [33]A survey by The Israel Democracy Institute in 2021 indicated that Arab residents expressed lower trust in the local authorities (32.0% vs. 62.0% of Jewish residents), [34] which might create an opportunity for external entities to in uence the retailers by offering guidance, such as advising on how to work around the display ban.
Internal IQOS signage at audited POS was found in more than half the POS in Arab and Jewish neighborhoods.However, these were mostly indirect and did not explicitly state the brand name, but used more general statements (such as "heated tobacco").Nonetheless, a high proportion of the signs (76.3%) used speci c colors that correspond to HEETS avors which could be interpreted as branded advertisement, and therefore forbidden by law.Additionally, we found that some retailers were provided with display cases that only show the product when a switch is turned on, thereby subtly violating the display ban.These ndings strengthen previous results showing the various ways in which PMI circumvents legislation. [3,35] itations These two studies used cross-sectional data that might not be re ective of all POS in Arab and Jewish neighborhoods in Israel.We only assessed POS in major cities, and our data collection efforts resulted in a small sample size for both studies.PMI might have employed other marketing tactics in other cities, or in the periphery.However, the IQOS website list of POS includes very few IQOS POS located in other dominant Arab cities and periphery localities.Additionally, we collected data both via phone, online, and face-to-face, with differences in data collection between Arab and Jewish POS (Jewish POS 79.2% by phone/online [n = 38/48], Arab 76.2% face-to-face [n = 16/21]), which might have impacted our results.
The majority of POS in Jewish neighborhoods were of middle-and high-SES, while the POS in Arab neighborhoods were of either medium-or low-SES, suggesting that differences might be based on economic, rather than ethnic factors, which cannot be differentiated within the scope of these studies.In addition, there might be some differences based on other factors, such as the city, including predominant population (mixed, predominant Arab, predominant Jewish), and store type.The small sample size precluded us from running more analyses to adjust for these factors.The data for both studies was collected at different points in time, which might have impacted the results.However, there were no differences in legislation or implementation during these times.
Nazareth was the only majority Arab city included in these studies, and it has both a Christian and Muslim population which could have affected the results.Currently, there is no available data on differences in smoking rates between Muslims and Christians, but a study conducted in 2012 showed that Muslim Arabs had a higher secondhand smoking exposure at home (55.4%), compared to 49.0% of Christian Arabs. [36]This could be attributed to differences in smoking behaviors in these two subpopulations.

Conclusions
Even

Table 1 .
POS and retailer characteristics, overall and across Arab vs. Jewish neighborhoods † Chi square test, unless stated otherwise.#Fishers exact test.^Mann-Whitneytest.*

Table 2
NISoff the price of HEETS, or buy IQOS and get a free HEETS package.Bold indicates between-group statistically signi cant differences (Bonferroni correction).More retailers in Arab vs. Jewish neighborhoods received invitations to IQOS events/parties (47.6% vs. 21.7%,p=0.032) and paraphernalia (30.0% vs. 2.2%, p=0.002).The most prevalent form of promotions targeting customers were price discounts (18.8% and 44.4% of POS in Arab and Jewish neighborhoods, respectively, p=0.069) (Table summarizes IQOS/HEETS marketing strategies overall, and compares retailers in Arab vs. Jewish neighborhoods.Compared to Jewish neighborhoods, a higher proportion in Arab neighborhoods carried less than 4 HEETS avors (66.7% vs. 17.4%, p<0.001) and less IQOS special displays (25.0% vs. 53.2%,p=0.034)(Table2).Table2.POS IQOS/HEETS marketing strategies, overall and across Arab vs. Jewish neighborhoods POS -Point-of-sale.† Chi square test, unless stated otherwise.# Fishers exact test.* Missing: Number of HEETS avors n=2; Special display n=2; Promotions to POS: free HEETS samples n=8; price discounts for your own purchases n=8; paraphernalia n=3; other gifts n=7; price discounts, rebates, or incentives based on promoting their products n=4; incentives for sales n=4; invitations to parties n=2; Promotions to costumers: free HEETS samples n=9; paraphernalia is zero for all POS; other gifts n=9; price promotions n=10; price discounts n=8; coupons n=12; special prices for members n=11; Special discounts for military/students was zero for all POS; Advertisements: any form of ad n=4 (if all items were missing); online n=4; social media n=6; print media n=5; inside the POS n=6.§ Other gifts given to retailers included lighters, lighter stands swimming competitions, and ights abroad.‡ Other gifts given to customers included lighters.$ Price promotions offered to customers such as buy one get one free, 2

Table 3
lists the retailers' interactions with a PMI representative and PMI's reaction to the POS display ban.More retailers from POS in Arab than Jewish neighborhoods reported having any form of interaction with a PMI representative (80.0% vs. 51.2%,p=0.029) with no statistically signi cant differences in regards to the detailed nature of those interactions.

Table 3 .
Interactions with a PMI representative and PMI's reaction to the POS display ban, overall and across POS in Arab vs. Jewish neighborhoods PMI representatives assisted the majority of POS implement the display ban (89.5% in Arab and 77.3% in Jewish neighborhoods), the only borderline signi cant difference was more POS in Arab neighborhoods being advised on how to navigate and overcome regulatory restrictions, (26.3% vs. 6.8%,p=0.05) (Table POS -Point-of-Sale.PMI -Philip Morris International.$ Includes ticking at least one statement from a-g. † Chi square test, unless stated otherwise # Fisher's exact test.* Missing: Speci c IQOS/HEETS salesperson n=3; Interaction with a PMI salesperson: Any interaction n=6; provide direction on placement n=5; target market n=7; communicate with consumers n=9; information on product n=9; PMI's reaction the POS display ban: All items n=6.

Table 4
summarizes the POS characteristics, marketing material, placement, promotion and regulatory compliance data from all audited and across Arab vs. Jewish neighborhoods.The audited POS were mainly convenience stores not within a gas station (45.0%, n=58) or convenience stores within a gas station (31.0%, n=40).Signi cantly more POS in Arab neighborhoods were located in areas of low SES (75.0%, n=36), compared to only 2.4% (n=2) of POS in Jewish neighborhoods (p<0.001)(Table4).
though IQOS marketing was very similar across POS in Arab and Jewish neighborhoods, PMI utilized marketing elements of cultural signi cance, especially for POS in Arab neighborhoods, such as more personal communication and invitation to social events.Despite Israel having both a POS advertisement ban and a display ban in place, our results indicate a high level of regulatory noncompliance and legislation circumvention in both Arab and Jewish populations, stressing the need for better implementation and enforcement.Additional research with continuous surveillance is needed to fully understand demographic and ethnic-based differences in tobacco advertising and to reduce smoking-related health disparities.AK and YBZ were involved in the paper's conceptualization, design, and developing the data collection tools.AK oversaw the RA training, data collection, and conducted the data analysis and interpretation under the guidance of YBZ.AK drafted and revised this paper, under the guidance of YBZ.CJB, LCA and HL were involved in the paper's conceptualization, developing the data collection tools, data interpretation, editing and revising this paper.ZD, YW, YC and CRL were involved in conceptualizing, editing and revising this paper.DE provided the list of POS (with names, locations, and neighborhood SES status), and was involved in editing this paper.