Figure 3 Smoking habits by age

On the Figure 3 it is easily seen that the number of respondents were mostly people from less than 18 to 30. The number of underaged smokers is 4 that is equal to 16,7% of this age group and 2 of them are already quitted smoking (8,3%). The number of smokers at the age of 18-25 is 119 respondents that is equal to 43,27% of this age group and between 26-30 the number of smokers is 7 that is equal to 50%. Overall, the results get from the survey illustrate the popularity of smoking for people from 18 to 30.

Thirdly, going to the IQOS promotion that is in the main focus of this work two distribution analyzes are conducted. According to two questions in the survey: “Have you seen IQOS advertising?” and “Have you ever been offered to try IQOS?”.

By the Figure4 below, the statistic shows that majority of asked respondents have ever seen the IQOS advertising because 73,6% of people answered as “yes”. Comparing with results from the first distribution analysis only 42% of the respondents are smoking, so the actual target group is much higher. Below is a graphical distribution of the answers received to this question.

Figure 4 IQOS advertisement

According to the question: “Have you ever been offered to try IQOS?”, - the Figure 5 is built. It shows the high level of IQOS successful promotion because 52,8% were offered to try IQOS and it is also necessary to compare it with 42% of smokers.

Figure 5 IQOS promotion level

To summarize, due to the distributional analysis of the respondents women are more addictive to cigarettes than men are. The most smoking age group is from 18 to 30 and there are underaged smokers and some of them have already quitted smoking habit. IQOS promotion works in a successful way for the company because the majority of the respondents have seen the advertising of IQOS and were offered to try it. Part of these respondents are not smoking but were targeted by IQOS marketing campaign.



5.2 Statistical data analysis

For successful data analysis, a binomial logistic regression is applied. To process the analysis the statistical program RStudio is used. For binomial logistic regression dependent binomial variable is whether the person will switch to IQOS or not and as for independent variables are used according to Likert Scale answers on each characteristic. Due to the analysis each characteristic should be coded as x1, x2, x3…x23 in the way it was descripted in the methodological section. To prove the hypotheses the p-value should not exceed 0,05 and where the p-value is lower than 0,05 these variables are understood as significant influence on dependent variable.

Firstly, the first model was built that is showing the prediction of respondent likelihood to use or not to use IQOS based on independent variables (see ). The variables for this model are x1, x2, x3, x4, x5, x6, x7, x8, x12, x14, x15, x16 and x19.



For example, taking x15, the value of which IQOS does not cause public censure, like regular cigarettes. According to the results of the model calculation, we see that x15 is significant, and as this variable increase by one conventional unit, the probability that a person will use IQOS, compared to what he will not use, increases by 0.4. That leads to the fact that as more a person is sure that IQOS does not cause public censure, like regular cigarettes, the probability that it will use it increases by 0.4. The same level of significance is shown by x5 that IQOS attracts consumers by its design.

However, less significant but also important level are shown by x6 and x8, about attractiveness of the loyalty program and that IQOS is less harmful than regular cigarettes. Overall, all of these four factors (x5, x15, x6 and x8) characterizing the image of IQOS in the society proves the Hypothesis 3 that IQOS public image makes it more desirable to try this device.

It is important to note that when doing the model, a person smokes or not (smnotsm), a person who does not smoke is more likely to use IQOS than a person who does not smoke 1.15 times higher. That reject the Hypothesis 2, that smokers are more likely to switch from the cigarettes to IQOS than nonsmokers.

The next stage is using the odds ratio formula to check the propensity to consume (see). When x15 and x5 are increased by one conventional unit, the chance that a person will use IQOS will increase 0.65 times.

Further stage uses pseudo R2s formula is checking if our model has a predictive ability: looking at the McFadden graph only, in order for the model to have a predictive ability, the indicator should be equal to 1, in fact it is close enough to 1 to say that the model has a predictive ability.

Secondly, another model was applied to evaluate how advertising affects whether a person will use IQOS or not (see ). The variables for this model are x10, x11, x13, x17, x18, x20, x21, x22 and x23.



The valuable significance is depicted only by characteristics x13 and x20. That are the points about hate being targeted by IQOS advertising and that children will definitely see advertisement of IQOS. Characteristic x13 shows that the more unpleasant the exposure of an ad is to a person, it means that he sees it constantly and everywhere, the more likely he will use IQOS, because advertising will still work and it is statistically proven. However, the fact that x20 influences positively the likelihood of buying attitude toward advertising could not play role in this way logically. However, the sample of respondents mainly consists of young people who have no children, hence they will see it or not, but they will use IQOS, and there are also people younger than 18, they don't even think about it, and they are advertised. In this way, Hypothesis 1 was partly rejected and partly proved by the results of the data analysis.

5.3 Interviews from PMI

The following research paper is needed to have additional comments from PMI company and IQOS unit that is why five personal interviews were held by the telephone. The problem that was faced on this stage is that all the employees have signed a confidential document that does not give them a chance to give comments on some of the questions. However, five interviews were collected to give the overall picture on the market situation and the atmosphere of the employee's attitude toward such product as IQOS. The full description of interviews is presented in the Appendixes.

IQOS employees evaluate product marketing campaign as successful and that the product is successful on the market:

“… the success of IQOS is cannot be underestimated. Progressive young generation considered smoking cigarettes unfashionable due to the promotion of a healthy lifestyle, and many of them quit smoking. Quit smoking in the direction to use IQOS, was a "step" in the direction of a healthy lifestyle, respectively, this was not to be criticized, but rather supported by other people”.

There are some insights that investments in the marketing campaign does not correspond to predictable level of sales:

“ from the point of view the launching of a new product, there is a positive trend, since the product has become widespread both in the domestic market and abroad but nevertheless, relative to the total costs allocated for the promotion, supply, maintenance of the company's image, the total income does not correspond to the planning level”.

The whole perception from IQOS representatives is that IQOS is a successful product that would lead “progressive young generation” to the smokeless future.

However, only one interviewer comments that IQOS could be listed as a product with nicotine others perceive IQOS as harmless:

“When smoking IQOS there are no risks associated with second-hand smoking and this has been proved, therefore, it is possible not to regulate the smoking zone by law (like now); if you equate cigarettes and IQOS - it will be impossible to smoke in a cafe. However, process of smoking IQOS is based on consumption of the nicotine that leads to harmful effect on health, so the legislation on advertising and product promotion should apply to IQOS as it does to cigarettes”.



6. Conclusion

IQOS is a tobacco consumption product with smokeless technology that allows to minimize the health consequences comparing to cigarettes. In this research paper 322 people were collected to give their perception on being targeted by IQOS marketing campaign to understand how unethical their performance is. The unpredictable result was received that IQOS is more likely to be bought by nonsmokers than by smokers. Despite the fact that it rejects the Hypothesis 2 it gives unpredictable result of unethical marketing campaign of IQOS. Their main goal of marketing campaign is to attract people who already smoke to switch to this harmless device but results shows the not only smokers are targeted and that influence their attitude.

However, IQOS being on the Russian market only from 2015 has already built a powerful reputation and the public image makes it more desirable to try this device that was stated as the Hypothesis 3 and proved by statistical analysis. The main factors that makes IQOS so desirable and attractive to buy and try is only the credit only of the marketing campaign. People whenever they smoke or not perceive IQOS as harmless and received comments from PMI confirm it. Nevertheless, the in-depth analysis of the literature provided in the second sections depicts other point of view due to the chemical reports of PMI to FDA and conducted experiments. Latest researches highlight that IQOS could be even more harmful than regular cigarettes. IQOS Heets sticks and the effect of heating could lead also to serious diseases.

To conclude, the research aim to identify and to prevent unethical ways of IQOS marketing campaign was achieved. For the further research, I would recommend to conduct the research in the similar manner but with bigger sample and more responds achieved. However, as far as this topic or related ones will be explored there is more possibility that it would change current situation because for now the research gap is too unexplored that led to several difficulties in the paper. For the PMI company:

· In the ethical scope of this research: IQOS marketing campaign should not be banned but the channels of spreading the advertisement should be limited in order to target only current cigarettes smokers. There is an ethical need to put pictures of possible diseases on the Heets packs because according to the researches IQOS do harm and people should be warned about it.

· In commercial scope: the results of the analysis showed that both smokers and nonsmokers do not like being too much targeted with advertisement, that is why there is a need to weaken the marketing strategy. Moreover, women show higher rate of addictiveness that is why there might be a need to switch the focus of marketing on women.

Overall, this paper is the first research exploring the heating tobacco system from the point of consumer perception of marketing campaign in ethical context. I hope that there will be more researches in this area and they will succeed to cope with all ethical dilemmas that it arises.



Reference list

1. Anastasia, S. (2019). Project Propolsal.

2. Auer, R. (2017). Heat-Not-Burn Tobacco Cigarettes: Smoke by Any Other Name. JAMA International Medicine, 1050-1052.

3. Benowitz, N. (2008). Clinical pharmacology of nicotine: Implications for understanding, preventing, and treating tobacco addiction. Clinical Pharmacol- ogy and Therapeutics, 83(4), 531-541.

4. Benowitz, N., & Henningfield, J. (2013). Reducing the nicotine content to make cigarettes less addictiv. Tobacco Control, 22, 14-17.

5. Choi, K., Hennrikus, D., Forster, J., & Moilanen, M. (2013). Receipts and redemption of cigarette coupons, perceptions of cigarette companies and smoking cessation. Tobacco Control, 22(6), 418-422.

6. Corey, C., Wang, B., Johnson, S., Apelberg, B., & Husten, C. (2013). Notes from the field: electronic cigarette use among middle and high school students - United States. MMWR Morb Mortal Weekly Report.

7. Cumming, M. (2015). Is it not time to reveal the secret sauce of nicotine addiction? Tob Control, 24(5).

8. Davis, J., & Gould, T. (2008). Associative learning, the hippocampus, and nicotine addiction. Current Drug Abuse Reviews, 1(1), 9-19.

9. DiFranza, J., Ursprung, S., & Gisele, A. (2011). Assessment of Tobacco Addiction in Adolescent. International Journal of Mental Health, 65-84.

10. Dingel, M., Karkazis, K., & Koenig, B. (2011). Framing Nicotine Addiction as a "Disease of the Brain": Social and Ethical Consequences. Social Science Quarterly, 92(5), 1363-1388.

11. Dutra, L., & Glantz, S. (2012). Electronic cigarettes and conventional cigarette use among US adolescents: a cross-sectional study. JAMA Pediatr, 1-125.

12. Etter, J. (2012). The Electronic Cigarette: An Alternative to Tobacco? 1-125.

13. Felberbaum, M. (2013). Marlboro maker Altria jumping into e-cigarettes.

14. Friedman, L., Cheyne, A., & Givelber, D. (2015). Industry use of personal responsibility rhetoric in public relations and litigation: disguising freedom to blame as freedom of choice. Am J Public Health, 250-260.

15. J., K. (б.д.). Can't buy my love: How adverising changes the way we think and feel. New York: Simon and Schuster, 55-78.

16. Kmietowicz, Z. (2014). E-cigarettes latest: users on the up but rules tighten. British Mediacl Journal, 349.

17. Kmietowicz, Z. (2014). Use of e-cigarettes in UK has tripled in two years, finds survey. BMJ.

18. Kotler, R., & Rackham, N. (2006). Ending war between Sales and Marketing.

19. Kress, M. (2017). Is heat-not-burn the next big thing in tobacco? Convenience Store News.

20. McCarthy, G., & Wakefield, M. (2010). Smoker sensitivity to retail tobacco displays. 159-163.

21. Miyashita, L., & Grigg, J. (2018). Effect of the iQOS electronic cigarette device on susceptibility to S. pneumoniae infection. Clin Immunol.

22. Moxham, J. (2000). Nicotine addiction: Should be recognised as the central problem of smoking. British Medical Journal, 391-392.

23. Pepper, J., & Brewer, N. (2013). Electronic nicotine delivery system (electronic cigarette) awareness, use, reactions and beliefs: A systematic review. Tobacco Control, 1-10.

24. Rose, J. (2006). Nicotine and nonnicotine factors in cigarette addiction. Psy- chopharmacology,, 274-285.

25. Rosenthal, D., Weitzman, M., & Benowitz, N. (2011). Nicotine addiction: mechanisms and consequences. International Journal of Mental Health, 40(1), 22-38.

26. Shaw, D. (2015). Hospitals are wrong to ban e-cigarette. British Medical Journal.

27. Thun, M., Carter, B., & Feskanich, D. (2013). 50-year trends in smoking-related mortality in the United States. N EnglJ Med, 351-368.

28. U.S. Department of Health and Human Services. (2010). How tobacco smoke causes disease: the biology and behavioral basis for smoking-attributable disease.

29. Vardavas K., G. C. (2013). Changes in tobacco industry advertising around high schools in Greece following an outdoor advartising ban: a follow-up study. Tobacco Control, 299-301.

30. Wang C., K. C. (б.д.). Online Shopper Behaviour: Influences of Online Shopping Decision. Asian Journal of Business Research, 66-74.

31. Yoshida, J. (2018). The Rise of Heat-Not Burn Tobacco in Japan: A “Hot” Issue for Tobacco Control. JP Journal of Thoracic Oncology.



Appendix 1



Appendix 2

Load the data

y <- readxl::read_xlsx("iqos.xlsx")

Preparing data for analysis

y$iqos <- ifelse(y$iqos == "1", "0", "1")

y$gender <- ifelse(y$gender == "1", "0", "1")

y$pr <- ifelse(y$pr == "1", "0", "1")

y$smnotsm <- ifelse(y$smnotsm == "1", "0", "1")

y$iqos <- as.numeric(y$iqos)

y$smnotsm <- as.numeric(y$smnotsm)

y <- na.omit(y)

Logistic regression: model 1

model1 <- glm(iqos ~ x1 + x2 + x3 + x4 + x5 + x6 + x7 + x8 + x12 + x14 +

x15 + x16 + x19 + gender + smnotsm, data = y, family = binomial(link=logit))

summary(model1)

##

## Call:

## glm(formula = iqos ~ x1 + x2 + x3 + x4 + x5 + x6 + x7 + x8 +

## x12 + x14 + x15 + x16 + x19 + gender + smnotsm, family = binomial(link = logit),

## data = y)

##

## Deviance Residuals:

## Min 1Q Median 3Q Max

## -2.6335 -0.6075 0.4860 0.7071 1.5354

##

## Coefficients:

## Estimate Std. Error z value Pr(>|z|)

## (Intercept) 2.292932 0.844802 2.714 0.006644 **

## x1 0.165533 0.164833 1.004 0.315259

## x2 -0.044084 0.157973 -0.279 0.780198

## x3 0.040537 0.127311 0.318 0.750172

## x4 -0.153154 0.125571 -1.220 0.222596

## x5 -0.534180 0.165820 -3.221 0.001275 **

## x6 0.429483 0.177539 2.419 0.015559 *

## x7 -0.025101 0.171530 -0.146 0.883656

## x8 -0.326170 0.164283 -1.985 0.047098 *

## x12 0.004345 0.177214 0.025 0.980438

## x14 -0.030033 0.153957 -0.195 0.845337

## x15 -0.429205 0.158672 -2.705 0.006831 **

## x16 0.149038 0.142950 1.043 0.297139

## x19 0.303181 0.178675 1.697 0.089729.

## gender1 -0.025568 0.323531 -0.079 0.937011

## smnotsm 1.158467 0.317688 3.647 0.000266 ***

## ---

## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

##

## (Dispersion parameter for binomial family taken to be 1)

##

## Null deviance: 369.63 on 321 degrees of freedom

## Residual deviance: 300.79 on 306 degrees of freedom

## AIC: 332.79

##

## Number of Fisher Scoring iterations: 5

Compute the odds ratios for all coefficients (model1)

exp(coef(model1))

## (Intercept) x1 x2 x3 x4 x5

## 9.9039333 1.1800216 0.9568735 1.0413702 0.8579979 0.5861500

## x6 x7 x8 x12 x14 x15

## 1.5364629 0.9752114 0.7216824 1.0043547 0.9704139 0.6510267

## x16 x19 gender1 smnotsm

## 1.1607176 1.3541602 0.9747564 3.1850457

Compute pseudo R2s for the model (McFadden)

BaylorEdPsych::PseudoR2(model1)

## McFadden Adj.McFadden Cox.Snell Nagelkerke

## 0.18623774 0.09425465 0.19248020 0.28193685

## McKelvey.Zavoina Effron Count Adj.Count

## 0.29565359 0.23778969 0.80745342 0.26190476

## AIC Corrected.AIC

## 332.79349203 334.57709859

Logistic regression: model 2

model2 <- glm(iqos ~ pr + x10 + x11 + x13 + x17 + x18 + x20 + x21 + x22 + x23 + smnotsm,

data = y, family = binomial(link=logit))

summary(model2)

##

## Call:

## glm(formula = iqos ~ pr + x10 + x11 + x13 + x17 + x18 + x20 +

## x21 + x22 + x23 + smnotsm, family = binomial(link = logit),

## data = y)

##

## Deviance Residuals:

## Min 1Q Median 3Q Max

## -2.5586 -0.4869 0.4879 0.6802 1.8438

##

## Coefficients:

## Estimate Std. Error z value Pr(>|z|)

## (Intercept) 0.8368 0.9105 0.919 0.35808

## pr1 -0.3310 0.3801 -0.871 0.38386

## x10 0.2395 0.1422 1.685 0.09198.

## x11 0.1548 0.1167 1.327 0.18445

## x13 0.2840 0.1283 2.214 0.02686 *

## x17 -0.5117 0.1223 -4.182 2.89e-05 ***

## x18 -0.4414 0.1709 -2.583 0.00980 **

## x20 0.4045 0.1682 2.405 0.01615 *

## x21 -0.1884 0.1625 -1.160 0.24613

## x22 0.3329 0.2106 1.581 0.11384

## x23 -0.1680 0.1738 -0.967 0.33354

## smnotsm 0.9126 0.3168 2.881 0.00397 **

## ---

## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

##

## (Dispersion parameter for binomial family taken to be 1)

##

## Null deviance: 369.63 on 321 degrees of freedom

## Residual deviance: 292.89 on 310 degrees of freedom

## AIC: 316.89

##

## Number of Fisher Scoring iterations: 5

# Compute the odds ratios for all coefficients (model2)

exp(coef(model2))

## (Intercept) pr1 x10 x11 x13 x17

## 2.3088676 0.7182218 1.2706670 1.1674434 1.3283936 0.5995003

## x18 x20 x21 x22 x23 smnotsm

## 0.6431365 1.4986208 0.8282614 1.3950413 0.8453245 2.4908630

Compute pseudo R2s for the model (McFadden)

BaylorEdPsych::PseudoR2(model2)

## McFadden Adj.McFadden Cox.Snell Nagelkerke

## 0.2076326 0.1372926 0.2120712 0.3106329

## McKelvey.Zavoina Effron Count Adj.Count

## 0.3332106 0.2607294 0.7981366 0.2261905

## AIC Corrected.AIC

## 316.8852377 317.8949465

Размещено на Allbest.ru

...