POLITICS, MARKETING AND SOCIAL MEDIA IN THE 2018 LOCAL ELECTIONS IN ICELAND

: The importance of marketing techniques in political campaigning has increased as communicating politics has become more complex in a highly fragmented media environment. With different media logics interacting in a hybrid media system political marketing methods through social media have drawn considerable attention and even been seen to pose a threat to democratic processes. This paper looks at the extent and nature of the use of marketing techniques in the 2018 municipal elections in Iceland, by using a mixed methods approach. The findings of a candidate survey and interviews with campaign managers suggest that the methods used are by and large a technical extension of previous methods and not qualitatively different from traditional electioneering. Both social media and traditional media are important marketing vehicles, but the importance of social media clearly on the rise. However, in lager communities in the capital region there is a higher degree of professionalism than in other parts of the country and the size of municipality is important, while the type of party or age of candidates is not.


Introduction
The use of social media in elections has in recent years become a heated political topic for a variety of reasons.One aspect of this discussion relates to the collection, purchase, and misuse of personal data gathered by different social media platforms, in particular by Facebook, in conjunction with the dissemination of targeted disinformation which has been seen to have influenced electoral outcomes.Other more general aspects of the discussion of the use of social media in political election campaigns relates to political marketing and the overall quality of democracy.The well-known criticisms and ethical considerations that have been articulated to describe the increased proficiency of both political entities and media outlets that accompanied the introduction of spin and professional "communication machines" (Mancini 2001), are now reinforced with the advent of social media.Indeed, the dichotomy between the intellectual autonomy of citizens to make political choices on the one hand and the freedom to compete for political power through influencing citizens on the other hand has become a more pressing issue as the sophistication of political marketing becomes greater (Beckman 2018).Indeed, the targeting of groups is at the essence of marketing and political electioneering, as became highly visible in the aftermath of the municipal elections in 2018 in Iceland.The Icelandic Data Protection Authority investigated a research project run in Reykjavík by the City of Reykjavík and the University of Iceland, gathering information aimed to increase voting participation of young people and women of foreign decent.The result of the investigation was that the use of information gathered from the National registry for targeting these groups were contrary to the then existing privacy legislation, a legislation that has since become even more protective of personal information (Data Protection Authority 2019).Following the ruling of the Data Protection Agency a series of complaints of malpractice were filed on the basis of the ruling by the city council member from the Centre Party, demanding that the elections should be nullified.That demand was however declined in the end with a ministerial ruling (Ragnarsdóttir 2019).
The impact of targeting and of social media on elections and democracy and political marketing in general thus raises both theoretical and ethical normative questions on democracy as well as purely empirical questions on whether and in what way political marketing through social media is actually conducted.This paper seeks to answer empirical questions about the use of social media in political marketing in Iceland by focusing on the municipal elections of 2018.The aim is twofold, firstly, to map the extent and nature of social media and traditional media use in local elections in Iceland and secondly, to establish the degree of sophistication of marketing expertise used by candidates and the political parties.
Some research has been done on political communication and media use by politicians in Iceland, but by and large, the field is under researched.This is the first systematic attempt to map political marketing through social media in Iceland.

Political communication and marketing in Iceland
Political communication is a broad term that refers to the relationship between politics and the media.It includes the political and cultural values that characterize the communication of politicians and media actors such as political journalists and editors (Pfetch 2014) and it includes as well the actual way in which politicians use different media, media logics and also the general mediatization of politics (Strömback 2008;Stömback & Esser 2014).The concept of political marketing on the other hand refers to the way in which the techniques of marketing are applied to politics and thus to the category of political communication but is narrower in scope.Although the basics of these marketing techniques are the same, they may vary somewhat according to the task they are applied to.The most common type of task is an election campaign and political marketing before elections will be the focus in what follows.
Some of the first definitions of marketing stated that marketing was the process of selling and buying goods.Over the course of time, these definitions have changed towards marketing being the process of delivering value to the customer (Ringold & Weitz 2007).According to Kotler and Keller (2006), marketing is the art and science of choosing target markets, getting, and holding on to customers by delivering superior value.In a way, that can be said about political marketing as well, as the aim of political marketing is to provide important benefits to the customers or in this case the voters (Beckman 2018).Political marketing involves using the same marketing tools as general marketing does.As declared by Menon (2008) political marketing is the result of intertwining marketing and politics.The political campaigner uses marketing techniques such as celebrity endorsements, online campaigning, segmentation, and microtargeting precisely as the general marketer does.Political marketing, as explained by Clemente (1992, 308), is described as: The marketing of ideas and opinions which relate to public or political issues or to specific candidates.In general, political marketing is designed to influence people´s votes in elections.It is different from conventional marketing in that concepts are being sold as opposed to products or services.Political marketing, however, employs many of the same techniques used in product marketing, such as paid advertising, direct mail and publicity.
Although different opinions do exist as to the extent to which analogy can be drawn between politics and corporate marketing (Kolovos & Harris 2005) the professionalization of political campaigning and political communication in general is firmly rooted in the marketing principles.In their seminal definition of the third age of political communications characterized by media abundance, Blumler and Kavanagh (1999) pointed out vital social developments in the latter half of the 20 th century that created the stage for the third age of political communication.These include modernization, individualizations, secularization, economization, and aestheticization.All of these elements can

STJÓRNMÁL & STJÓRNSÝSLA
Politics, marketing and social media in the 2018 local elections in Iceland be seen -partly at least -to back up a marketing approach to political communication in an era of intensified competition in politics and media, professionalizing imperatives, fragmented communication platforms and new ways in receiving political information (Blumler & Kavanagh 1999;Blumler 2013). Lilleker (2016) notes that the advent of new media has accelerated the integration of marketization "in terms of targeting, following media logic and reaching key voters through every means with the most manipulative message possible" ( Lilleker 2016, 683).
With the use of social media, anyone can be an advocator and possibly influence their audience (Baines & O´Shaughnessy 2014).The use of online platforms, such as social media, enables the micro-targeting of political messages to target voters that are likely to vote for a specific party or candidate.Using online micro targeting campaigners can send messages adapted to the receiver, for example, information regarding student loans to students or childcare information to young families.In a way, it can be said that online micro targeting can lead to more relevant information being sent to the voter groups and the campaigner might be able to reach audience they would otherwise not reach.But it does come with a large hidden cost: online micro targeting threatens privacy as it involves gathering and collecting personal data about the voters.Therefore, it has been suggested that transparency regarding online micro targeting is important, one way could be to disclose how much is spent on online political micro targeting that is how much the parties spend on advertisement on Facebook, Instagram, and other social media platforms (Borgesius et al. 2018).
Social media has dramatically changed political communication as they introduce a new media logic into communication, a network media logic which is different from the traditional media logic.Network media logic is based on the interaction of many to many, where connective information is shared between like-minded equals, while traditional media logic is based on communication of one too many and subject to journalistic scrutiny (Klinger & Svenson 2014, 2018).In present day media systems, political communication is conducted through both types of media logics as is demonstrated in Chadwick's thesis on the hybrid media system where power is determined by a successful use, for one's own goals, of information flows "across and between a range of older and newer media settings" (Chadwick 2013, 207).In a hybrid media system, political actors therefore make use of both old and new media platforms and media logics in their individual communication strategies when competing for power.
Word-of-mouth (WOM) and the internet/social version eWOM are the most persuasive marketing tools and much more credible than any other marketing tool such as advertising, direct marketing or support advertising (Iyer, et al. 2017).Political messages delivered via eWOM use one of five communication channels: Face-2-face, phone, text, social media or e-mail.The political campaigner should not ignore the perceived credibility of the WOM/eWOM messages to the receiver.WOM/eWOM messages from members of reference groups; that are family, friends, social media influencers etc. to the receiver are perceived more positively than messages received through other marketing channels.Since voters assign higher credibility to their reference groups, it can be

STJÓRNMÁL
& STJÓRNSÝSLA important for the political campaigner to identify individuals that have high numbers of followers (Iyer et al. 2017).
Decreasing voter turnout and engagement among younger demographics and the challenge of the political campaigners to reach that group, makes the use of online platforms particularly attractive.Due to their high engagement with social media platforms, digital marketing seems to be one solution to engage young voters as it has been proven difficult to reach young voters via traditional media (Leppäniemi, et al. 2010).This is a highly relevant issue in Iceland as the voter turnout in the 2014 municipal elections was quite low for voters under the age of 30 or under 50%, a situation that only marginally improved in the 2018 elections (Statistics Iceland n.d.;Eyþórsson & Önnudóttir 2017).
However, a political marketing campaign cannot solely be run via the digital media, as it will then not reach voters that mainly use traditional media.A dual political marketing strategy seems to be the best way to go, that is a traditional political campaign for the loyal older supporters mixed with internet campaign to attract newer and younger voters (Leppäniemi et al. 2010).That goes hand in hand with Chadwick's (2013) thesis and the hybrid media system and is further affirmed by Jouet, Vedel and Comby (2011) stating that even though digital media practises are developing, traditional media practises still prevail in order to reach different target groups.
Prior research on Icelandic political communications suggests that the political media system is characterized by hybridity (Guðmundsson 2016(Guðmundsson , 2019)).Using data on the media -use and the perceived importance of media platforms of Icelandic politicians in the parliamentary elections of 2016 and 2017, Guðmundsson concludes that a hybrid media system is in place with a precarious balance between old and new logics, where traditional media logic is still dominating.However, younger candidates utilize new media logics to a significantly greater extent than older ones suggesting that the balance might tilt as new and younger politicians move more centre stage (Guðmundsson 2019).Guðmundsson further draws attention to the resources that Icelandic parties have at their disposal with respect to the increased complexity of forming and carrying out communication strategies.These strategies are likely to have to be quite complex in a hybrid media system due to the abundance of media and the interaction between different platforms and logics that are at play.It is however precarious to assume that the same characteristics apply in national elections to Alþingi as they do to municipal elections.The diversity of the 72 Icelandic municipalities in terms of size and local concerns is vast and the number of municipalities make generalization difficult.As Hlynsdóttir (2018) has pointed out in her historical analysis of local self-government, it has "persisted on a path of maintaining high autonomy and a large number of small and densely populated municipalities since the establishment of the current system in the late 19th century" (Hlynsdóttir 2018, 96).Contrasting this to six large constituencies in national parliamentary elections highlights differences in communication strategies and the varying need for targeting groups in political marketing.Size and population of municipalities is thus likely to play a role in political marketing in an elections campaign and for political communication in general.Overall however, communication profes-

STJÓRNMÁL & STJÓRNSÝSLA
Politics, marketing and social media in the 2018 local elections in Iceland sionalism, including marketing expertise, be it WOM/eWMO, micro targeting or other expert input, can be expected to be in high demand in Iceland in order to gather, analyse and organize information used in political campaigning.
On the basis of the overview above some hypotheses can be proposed: Firstly, it needs to be determined if similar patterns of media use apply in municipal elections as they do in parliamentary elections, i.e. if traditional media and new media (social media) are used in a hybrid manner.Looking to the hybrid patterns established in parliamentary elections and the widespread use of social media in Iceland the following hypothesis is generated: H1. Social media and traditional media were used in a hybrid manner by political parties and candidates in the 2018 municipal elections.
With reference to the targeting possibilities of social media and the general trend of professionalisation of political campaigning it is reasonable to expect political parties and their candidates to use targeting of groups as an instrument in getting their political message across.Hence the second hypothesis, that comes in two parts: H2a Targeting was widely used by politicians and parties in the 2018 municipal elections.

And considering the nature and possibilities of social media:
H2b Social media rather than traditional media were used to get messages across to target groups.
With reference to the tradition of high autonomy of Icelandic local government and the large number of municipalities the need for targeting groups varies.Small municipalities are less segmented in terms of communication, pathways are shorter and the need for a targeted approach for politicians to reach voters not as pressing as in larger municipalities.This leads to the third hypothesis that like the second one comes in two parts: H3a Systemic targeting is used to a greater extent in larger urban municipalities than in smaller ones.
And following the logic of Hypothesis 3a that targeting is more common in larger municipalities the use of targeting and other marketing strategies can be expected to be more sophisticated as well.
H3b.More professional marketing strategies are used in large municipalities than smaller ones.

Methodology
In establishing whether the three hypotheses are substantiated by empirical data a single study mixed methods approach (Creswell & Plano 2006) was used, applying both qualitative interviews with key organizers of the election campaign of different parties as well as a candidate survey among candidates in 11 large and medium sized municipalities.By using a mixed method research design, the authors hoped to capture both the breadth and the depth of the importance of social media and the professionalism of marketing in the campaigns.Advantages that have been pointed out are e.g. that the mixed method can offset weaknesses of both quantitative and qualitative research, that it provides more comprehensive evidence, and that it helps answer questions that cannot be answered by qualitative or quantitative approaches alone (Creswell & Plano 2006;Davíðsdóttir & Ólafsdóttir 2013).Political communication in general and political marketing are clearly complex phenomena and involves both strategy-building and planning and then the actual execution of that strategy.While the actual combination of planning and the way in which the campaign is carried out may vary between municipalities and parties, generally, the emphasis of planning lies with the campaign office and managers of the parties or lists but the actual campaigning rests to a large degree on the shoulders of the candidates themselves.Thus, in order to collect holistic information both candidates and the persons responsible for the planning and organisation were approached.Due to the large number of candidates standing, a candidate survey was the most natural way to collect general information from them.In most cases, the campaign managers are the persons that have the best knowledge and inside information of the different strategies and peculiarities of the campaigns of their parties.By interviewing them, the difference between strategies could be assessed.Interviews were considered more appropriate in this case rather than other forms of qualitative methods such as e.g.focus groups of campaign managers or party office workers.This is partly due to practical reasons but also because the number of staff working for individual parties or lists varies making focus groups of party workers difficult and since the aim was to examine the differences between strategies, which might have been difficult in, inter party focus groups.In the present case then, the survey data gives an overview of the extent and actual use of certain marketing techniques, in particular targeting, but the interviews provide a valuable insight into the communication strategies the techniques are supposed to fit into.

The candidates survey
A descriptive comparative study was conducted in the form of a web-survey among candidates in eleven large and medium-sized municipalities.These were: Akureyrarbaer, Árborg, Borgarbyggð, Fljótsdalshérað, Hafnarfjörður, Ísafjarðarbaer, Kópavogur, Norðurþing, Reykjanesbaer, Reykjavík and Stykkishólmsbaer.
The top four candidates of all lists standing in each of the eleven municipalities received the survey, thereby capturing more or less the whole population of potential deputies that could be elected in the eleven municipalities.These eleven municipalities are evenly spread out through the country and should represent fairly the municipalities

STJÓRNMÁL & STJÓRNSÝSLA
Politics, marketing and social media in the 2018 local elections in Iceland where a list voting took place, but in the smallest communities votes are cast for individuals without them formally standing in the elections.The survey questions were part of a more extensive survey focused on political communication and it was conducted during the election campaign, opened two weeks before and closed two weeks after Election Day, which was the 26 th of May 2018.In total 237 survey links were sent out and 143 or just over 60% answered, although to some of the questions the response rate was a little less.The division of respondents between parties, sex and age was relatively even and no obvious reason to expect that the respondents would answer in a different manner that the non-respondents.The number of parties in the different municipalities varied and thus the number of candidates in different municipalities that received the questionnaire also varied.A party is here defined as the organized group behind a certain candidate-list offered in a municipality.This does not influence negatively the type of information sought here, and in addition the number of lists in communities reflects roughly the size of the municipality, e.g. the largest number of party-lists being in Reykjavík the largest municipality.In aggregate, the numbers from Reykjavík might therefore have somewhat greater weight, but that also reflects the important proportional position of the capital.
Four questions will be dealt with here as they deal directly with the issues raised by the hypotheses.The first one is about which media candidates used the most to convey their political messages and the second one was what medium they thought the most important (Hypothesis 1).These two questions are intended to establish the social media (and traditional media) use and perception of the candidates.The third and fourth questions look at the level of marketing practises that exists among the candidates themselves and their parties, by asking about the sophistication of targeting of voters (H2a, H2b and H3a, H3b).The questions were the following: Question 1) Please indicate how much or little you have/intend to use the following communication channels in your campaign for the upcoming parliamentary elections.Question 2) In principle, how important or unimportant do you think the following media gateways or media are for getting your political messages across in the upcoming elections?Question 3) Do you make use of information from experts or the party office about target groups?Question 4) Please indicate how strongly or little you agree with the statement: "Lack of communication expert help is a liability for my party´s campaign." All questions had predetermined options to be answered on a five-point Likert-scale: very much; rather much; neither much nor little; rather little; very little/not at all.The options dealt with in this paper in the first two questions were: Facebook; Twitter, Ins-

STJÓRNMÁL
& STJÓRNSÝSLA tagram, Snapchat; personal blog; radio; television; local media; other online media.The Research Centre at the University of Akureyri administered the survey.

The interviews
In order to establish more conclusively the level and sophistication of marketing techniques used on social media, campaign coordinators were interviewed.Iceland usually scores high on various measurements regarding internet use, in the year 2014, 97% of Icelanders aged between 16-74 were regular users of the Internet and that is the highest percentage of Internet users in Europe.There is some evidence that the use of Internet differs between the rural areas and the capital region, for example 61,7% of individuals in the capital region connect to the Internet via mobile devices versus 54,2% of individual of other regions (Hagtíðindi 2015).Research on Swedish and Norwegian voters suggest that the politicians in larger regions or cities might have more tech-savvy voters to gratify than their colleagues in more rural areas.As such the size of the municipality might influence their social media activity (Larson 2013;Larson & Kalsnes 2014).Since the social media activities were one of the main themes of the research the municipalities selected were Akureyri and Reykjavík, the largest municipality in the north and the capital in the southwest, respectively, all together 11 campaigners were interviewed.The interviews were half-standardized mostly face to face, but two interviews were conducted through the distance communication program Zoom.The length of the interviews varied from 20 to 70 minutes.All interviews were taped and transcribed, analysed and themes extracted.All interviewees were promised full confidentiality and that their responses would be completely anonymous and untraceable.

Results
Following is a report of results from both parts of the study.First an account will be given of the outcomes of the candidate survey followed by an account of the interviews.In both cases elements that relate to the three hypotheses will be the focus of attention.

Survey results
The results from the candidate survey suggest that local politicians did indeed use social media a lot in their campaign but did not do so by relying on information of specific target groups.Clearly, there are interesting variances although Facebook is by far the most used medium and it is also considered the most important.This is demonstrated in Figure 1 which also draws out that traditional media is still quite strong, both widely used and considered important.Hypothesis 1 thus receives support in the data.Social media, other than Facebook, was only moderately used in the 2018 election campaign, although Instagram seems considerably strong both in perceived importance and usage.However, it is Facebook that dominates the social media scene of candidates, which is indeed in line with other studies on media use of Icelandic politicians in elections (Bergsson 2014;Guðmundsson 2016Guðmundsson , 2019)).In general, candidates do not seem to consider a lack of assistance from communication experts to be a problem, as only one-fifth agree with a statement to that effect (Figure 2).
The possibility of targeting specific groups is one of the attractions of social media and Facebook offers several possibilities to do that.Overall, the candidates surveyed do not seem to be using information from experts or their party organization to target groups when they are using Facebook.Figure 2 demonstrates this clearly, as only 23% of candidates say they use targeting information much.Social media, other than Facebook, was only moderately used in the 2018 election campaign, although Instagram seems considerably strong both in perceived importance and usage.However, it is Facebook that dominates the social media scene of candidates, which is indeed in line with other studies on media use of Icelandic politicians in elections (Bergsson 2014;Guðmundsson 2016Guðmundsson , 2019)).In general, candidates do not seem to consider a lack of assistance from communication experts to be a problem, as only one-fifth agree with a statement to that effect (Figure 2).The possibility of targeting specific groups is one of the attractions of social media and Facebook offers several possibilities to do that.Overall, the candidates surveyed do not seem to be using information from experts or their party organization to target groups when they are using Facebook.Figure 2 demonstrates this clearly, as only 23% of candidates say they use targeting information much.Thus, if candidates are in fact systematically targeting groups through social media, they are doing it on their own and not using expert or central party knowhow.Hypothesis 2a is thus not supported by the data.However, these are aggregate descriptive numbers and considerable variance might be between different groups of candidates.

STJÓRNMÁL & STJÓRNSÝSLA
In order to establish the connection between social media use and the use of targeting information a linear regression was performed to see how well the use of targeting information (H2b) predicted the use of the three most used social media, i.e.Facebook, Instagram, and Twitter.A significant regression equation was found (F (1; 126) = 13,448, p= < 0.000) with a R 2 of .096.The use of targeting information thus significantly predicts the use of these three social media platforms.On the other hand, an examination of how well the use of targeting information predicted the use of traditional media gives an insignificant regression equation (F (1; 126) = 3,074, p = 0.082) and R 2 is 0.024.The pattern therefore supports Hypothesis 2b and the use of targeting information does not significantly predict the use of traditional media while it does predict the use of social media, suggesting that targeting -to the extent it is used by candidates -is primarily through social media.Thus, if candidates are in fact systematically targeting groups through social media, they are doing it on their own and not using expert or central party knowhow.Hypothesis 2a is thus not supported by the data.However, these are aggregate descriptive numbers and considerable variance might be between different groups of candidates.
In order to establish the connection between social media use and the use of targeting information a linear regression was performed to see how well the use of targeting information (H2b) predicted the use of the three most used social media, i.e.Facebook, Instagram, and Twitter.A significant regression equation was found (F (1; 126) = 13,448, p= < 0.000) with a R 2 of .096.The use of targeting information thus significantly predicts the use of these three social media platforms.On the other hand, an examination of how well the use of targeting information predicted the use of traditional media gives an insignificant regression equation (F (1; 126) = 3,074, p = 0.082) and R 2 is 0.024.The pattern therefore supports Hypothesis 2b and the use of targeting information does not significantly predict the use of traditional media while it does predict the use of social media, suggesting that targeting -to the extent it is used by candidates -is primarily through social media.
There does not seem to be much difference between parties, municipalities or even age when it comes to using social media, which is in fact also indicated by the practically universal use of Facebook.Similarly, in light the point made by Hlynsdóttir and Önnudóttir (2018) that organizational capabilities of established parties is grater than of local community parties or lists it might be expected that the nationally organized parties would use social media more than the local lists as these would be important

STJÓRNMÁL & STJÓRNSÝSLA
Politics, marketing and social media in the 2018 local elections in Iceland instruments of professionalized campaigning and organized targeting.However, a linear regression shows that being a candidate from a nationally organized party does not predict the use of social media (p = .938).Party affiliation, municipality, and age do not either predict the use of social media as shown in the coefficients Table 2.When these three independent variables are used to predict the use of targeting information, the outcome is similar except for municipality.Municipality predicts significantly the use of targeting information as shown in Table 1.The significant value is bolded.
Table 1.Multiple linear regression examining the relationship between the independent variables and three dependent variables: social media, the use of target information and the need for expert communication help would use social media more than the local lists as these would be important instruments of professionalized campaigning and organized targeting.However, a linear regression shows that being a candidate from a nationally organized party does not predict the use of social media (p = .938).Party affiliation, municipality, and age do not either predict the use of social media as shown in the coefficients Table 2.When these three independent variables are used to predict the use of targeting information, the outcome is similar except for municipality.Municipality predicts significantly the use of targeting information as shown in Table 1.The significant value is bolded.

STJÓRNMÁL & STJÓRNSÝSLA
definition of "functional urban areas" (FUA) be considered hinterlands from the main urban centre (OECD 2013;Eurostat n.d.).For a sharper picture of this relationship the municipalities were dummy coded and put through a multiple linear regression using Reykjavík as a reference variable.

Table 2. Multiple linear regression of dummy coded variables using Reykjavík as a reference variable
area, but Reykjanesbaer on the other hand and indeed Árborg, can, with reference to the definition of "functional urban areas" (FUA) be considered hinterlands from the main urban centre (OECD 2013;Eurostat n.d.).For a sharper picture of this relationship the municipalities were dummy coded and put through a multiple linear regression using Reykjavík as a reference variable.

Reykjavík as a reference
Note: Significant values are in bold.
As the results in Table 2 show, the pattern is by and large confirmed in that there is not a significant difference is found between Reykjavík and the above-mentioned municipalities in the South-West area.There is, on the other hand, a significant difference between Reykjavík and most municipalities in other parts of the country, the municipality of Árborg being a borderline case.The municipality of Ísafjörður in the Westfjords is, however, a notable exception.The data does therefore not support conclusively Hypothesis 3a although it can still be said that targeting information tends to be primarily used by candidates in Reykjavík and the greater capital (FUA) area.Special reasons might apply to the municipality of Ísafjörður, but more research is required before a sober suggestion can be made as to why Ísafjörður does not differ more from the targeting situation in Reykjavík.

Interview results
The interviews were all based upon a common interview-frame drawn from the hypothesis put forth.This frame consisted of questions on three main fields.Firstly, on the media use in the campaign, i.e. what media and platforms were mostly used and why?Secondly, the use of targeting and targeting information and the way in which this was carried out.Thirdly, the field of external or expert assistance used in the campaign.The

STJÓRNMÁL & STJÓRNSÝSLA
Politics, marketing and social media in the 2018 local elections in Iceland interviews were half standardized and these issues were not necessarily addressed in this order and other issues came up as well in many cases.
In presenting the analyses of the interviews it was decided to construct an analytical model based on themes extracted and the hypothesis.The analytical model is shown in Figure 3, and sums up the main findings of the qualitative data, thus provides a clearer picture of the different marketing strategies used by the campaigners.The interviews were, as stated before, taken in two different municipalities, Reykjavík, the capital, and Akureyri, the largest city in the northern part of Iceland.Five themes or categories were identified from the interviews for further traditional media, social media, micro targeting, Face2Face/WOM and lastly professional assistance.In the following sections an interpretation of the themes and their subcategories will be presented.

Traditional media
Traditional media refers to broadcast, print, and support media like billboards and other out-of-home media.Organised events by the parties also fall under the category of traditional media.
The interviewees were asked if they used print media, for example newspapers or leaflets, and broadcast media, for example radio or television advertisements to reach their voters.All of the interviewees said that they used the local media heavily.In Akureyri, there is a local ad-paper called "Dagskráin" that is printed and delivered to all homes once a week.That platform has high readings, 93% of women in the area and 85% of men read it (Vikudagur n.d.).One respondent explained: We used the local papers a lot.Just one example "Dagskráin" here in Akureyri, it is quite unique here, this paper lies on your table for a whole week until the next one arrives, the people in Akureyri don´t throw it away, then an advertisement on the front or back page is the best.

Traditional media
Traditional media refers to broadcast, print, and support media like billboards and other out-of-home media.Organised events by the parties also fall under the category of traditional media.
The interviewees were asked if they used print media, for example newspapers or leaflets, and broadcast media, for example radio or television advertisements to reach their voters.All of the interviewees said that they used the local media heavily.In Akureyri, there is a local ad-paper called "Dagskráin" that is printed and delivered to all homes once a week.That platform has high readings, 93% of women in the area and 85% of men read it (Vikudagur n.d.).One respondent explained: We used the local papers a lot.Just one example "Dagskráin" here in Akureyri, it is quite unique here, this paper lies on your table for a whole week until the next one arrives, the people in Akureyri don´t throw it away, then an advertisement on the front or back page is the best.

STJÓRNMÁL & STJÓRNSÝSLA
Politics, marketing and social media in the 2018 local elections in Iceland

Social media
Social media covers a variety of platforms, for example Facebook, Instagram, Snapchat and Twitter that are the most common ones in Iceland.All the interviewees were asked which social media platforms they used in their campaign.All the parties used social media in one form or another to reach their target groups and all agreed that social media is an important tool when it comes to political marketing.One respondent said: I am convinced that for us social media plays a huge role, much bigger than the other media and the videos and all that stuff … social media reaches our voters much better than print advertisement.
The four social media platforms stated earlier were the ones that were used in these elections.It was evident that Facebook was the main platform by all campaigners.All the interviewees said that they analysed what messages would fit which medium that is for example that some messages fitted better on Facebook while other on Instagram.All interviewees agreed that Instagram worked better to reach the younger voters while Facebook worked better for the older groups.
Instagram works for the younger people, we used it almost solely for that group, there are big groups missing, young groups on Facebook.Just like it is not common today, the use has at least lessened, these kids, these young people are much more on Instagram.
According to Leppäniemi et al. (2010) in order to reach out to younger voters the political campaigners will need to understand how to communicate with them via digital media and social media.The interview results support this, as one of our interviewees in Reykjavík said: Next time around I would skip all advertisements in newspapers and all advertisements in radio too, and TV also.Yes because people under the age of 25 they do not use these mediums, just not at all, TV is losing its power as well, it all depends on the TV program coming up to elections whether it is worth the money to buy advertisements on TV.I think I would carefully evaluate this in four years' time.
Snapchat and Twitter were not commonly used.The candidates for example used Snapchat for a day or two in order to give the viewers a glimpse into their daily life.The users of these platforms were considered to be the loyal voters of the party and not the general public.

Micro targeting
While analysing whether the campaigners used micro targeting, they were asked if they selected variables like; gender, age, interests and so on while advertising e.g. on Face-

STJÓRNMÁL
& STJÓRNSÝSLA book.The results show a difference in the use of micro targeting between Akureyri and Reykjavík.Respondents in Akureyri said that there was no use in selecting various variables when boosting advertisements on Facebook, the most commonly used social media.All interviewees in Reykjavík, except one, said they used micro targeting when using advertisements on social media.The one that didn´t use it said: We have been working with The Data Protection Authority.It is my understanding, our understanding, that it is not legal to micro target groups on social media.So, we didn´t do it.If I am wrong, then we can correct that next time around.But we did, you know, select the metropolitan area and of course the age 18-65 years old, because that is the voting age.And you can select from 18 to 65 plus, so we haven't been targeting you know only women or only men or whatever.
The others said that they did use micro targeting for example by the content used.
Yes, we did have target groups.And we did have specific issues that were targeted to certain groups, for example we did have videos regarding playschools and we micro targeted that to certain age groups to reach young parents.And then we had other videos that were targeted for everybody.
However, in Akureyri none of the campaigners said that they used micro targeting.When asked why they didn´t use this method to better target their voters, size was mentioned as a barrier.As explained by one campaigner, he didn´t think it was necessary to go to such lengths.The campaigner said: You have 22 people on the list of candidates and by that you cover almost 80-90% of the voters.All of the candidates have around thousand friends on Facebook and simply by that you have reached most of the voters that are around 13 thousand.

Face2Face and WOM
Personal communication, e.g.Face2Face and phone calls, can be important to influence voters and start WOM.This is important to start a dialog with the voters and have a two-way communication.All interviewees said that they used a Face2Face approach to reach their voters.The candidates went on company visits and out in public areas, for example shopping centres, to meet and talk to their voters.One mentioned that this was highly important to start WOM: We did this neighbourhood by neighbourhood, started with one, then the next one and so on.It is best if you get the ball rolling in one neighbourhood, because it grows and continues.If people are ready to talk about you, then we don´t have to do it ourselves.

STJÓRNMÁL & STJÓRNSÝSLA
Politics, marketing and social media in the 2018 local elections in Iceland Another respondent said: Face2Face in a municipality that is minus 20 thousand is the key.In my opinion.And social media is obviously very important, but then you need to have a team … And in Akureyri Face2Face is the way to go.But in Reykjavík it is more difficult.
Reaching voters by calling was considered important to the campaigners.One party used a call centre and called potential voters while others called their list of members.

Professional assistance
All campaigners were asked if they hired professional consultancy for example advertising agencies, or if they sought professional help in communication such as training for interviews, writing and television appearance.All campaigners had a certain strategy regarding the publishing of articles and what themes to cover.Two parties in Reykjavík mentioned that they hired marketing agencies to help them with their political marketing.None of the campaigners in Akureyri did.One party in Reykjavik mentioned that they hired an actress and a journalist to support and train their candidates.This party also invited their candidates and campaigners in other municipalities to take part in some of this work.In Akureyri the campaign management seemed to be almost entirely in the hand of the campaign manager with some support of others, and one party in Akureyri even outsourced to an external actor their social media communication.In Reykjavík however, it was noticeable that there was a bigger team active around the campaign managers, in order to support and help with the campaign.

Summary of findings
Taken together most of the hypotheses are supported in the data but surely, some more than others.
The first hypothesis suggested that hybrid media use, with new and traditional media used interchangeably, characterized the elections campaign in 2018.That was firmly supported in the survey data and in the interviews (Fig. 1).
The second hypothesis was put forth in two parts.Hypothesis 2a maintained that targeting and targeting information was widely used by parties and candidates in the 2018 campaign.This was not the case and only some 23% of candidates used targeting information rather or very much, while 59% used it rather little, little or not at all.However, the campaign managers in Reykjavík did say that they did use targeting on social media.Thus, Hypothesis 2a was only partially supported by the data and if there was extensive systematic targeting going on, it was done by only one in four of the candidates possibly with support from the central party organization or communication experts (Fig. 2).Hypothesis 2b on the other hand stated that the targeting that did take place was done through social media rather than traditional media.That part of the hypothesis was indeed supported by the survey data as well as the interviews.
The third hypothesis, also put forth in two parts, related to the size of municipalities

STJÓRNMÁL & STJÓRNSÝSLA
and the sophistication of marketing techniques.Hypothesis 3a stated that systematic targeting was used more in larger urban areas than in smaller ones.This received support in the data as the urban centre of the capital area used targeting more than municipalities elsewhere in the country.Municipalities that can be considered as hinterland (FUA) to the capital area, Reykjanesbaer and Árborg, are also candidates for more targeting practices, and Reykjanesbaer more so than Árborg.Ísafjörður in the west fjords is an outlaw in this respect, proofing not to be significantly different from Reykjavík.Hypothesis 3b is also supported in the data as not only is targeting more prevalent in the metropolitan area as show by the survey data but this is also conclusively suggested in the interview themes.

Discussion
The main findings of this study suggest that political parties and candidates in local elections in Iceland only use targeting or highly sophisticated marketing techniques through social media in a rather limited manner.Any use of personality profiling or big data analysis seems completely absent.In those instances where targeting is used in social media it can be seen as a technologically advanced extension of traditional marketing methods, such as finding and approaching target groups through micro targeting and enhancing the power of the WOM/eWOM approach.Indeed, the campaign seems largely to have unfolded along the line Leppäniemi et al. (2010) described for Finnish general elections a decade ago, namely a "dual political marketing strategy" appealing on the one hand to established supporters largely through traditional media and WOM methods and to new and perhaps somewhat younger voters through eWOM in social media and to a certain extent micro targeting.
Evidence from the interview data suggests that the role of social media might increase even further in the future, as information wells needed for targeting are increasingly drying up with stricter rules and greater awareness on personal data protection.One of the campaigners interviewed had checked with the Icelandic Data Protection Authority if targeting certain groups was in order and was told that it was not.As mentioned at the beginning of the paper this actually became an issue after the elections with a ruling of the Data Protection Authority (2019).The targeting through social media, however, is based on prior consent of the users of these platforms and thus restrictions on the grounds of official data protection can prove somewhat complex.That in turn might be an additional impetus for political actors to bypass these restrictions and rely more on social media platforms, at least until a formal ruling of the Data Protection Authority has been made on the targeting practices on social media.
The candidates themselves seem to feel relatively confident in the new media system, the hybrid system with different media logics simultaneously at play, as they do not believe that a lack of expert help it is a liability for their party´s campaign.
Furthermore, candidates use targeting information relatively little although candidates in the larger municipalities in the metropolitan area do more so than they do in the regions.That trend is indeed confirmed in the qualitative data as the interviews

STJÓRNMÁL & STJÓRNSÝSLA
Politics, marketing and social media in the 2018 local elections in Iceland suggest a more developed and extensive micro targeting in Reykjavík than in Akureyri.On the other hand, neither party nor age seem to predict the use of targeting, a finding that was strongly supported in the interviews as well.This finding seems to contradict Guðmundsson´s point discussed above, that resourceful parties might be able to conduct communication in a hybrid media system in a more effective or expedient manner that less resourceful parties.To be sure, resourceful parties might of course still have an edge, as the current data does not rule that out convincingly.However, that additional effectiveness does not seem to be achieved by using marketing techniques or access to better information or communication expertise at the local level of politics.
It has become apparent in this study that the use of marketing techniques such as micro targeting and the use of expert communication or marketing advisers in the local elections are related to the size of the municipality in question.Despite significant statistical relations between size of municipality and targeting, other cultural and placespecific factors also play a role and might explain the difference between municipalities in the metropolitan hinterland such as Reykjanesbaer and Árborg.The special position of Ísafjarðarbaer might also be explained with reference to such factors.
Overall, it can be suggested that a certain threshold in terms of numbers is needed for professional marketing methods to be more beneficial than simple ones e.g.some variant of WOM/eWOM methods.Therefore, a difference is likely to occur in the marketing techniques of candidates and parties between local elections and national elections.In national elections the constituencies resemble more the municipality of Reykjavík in terms of the number and diversity of voters, the capital city itself being divided into two constituencies in national parliamentary elections.Thus, it seems probable that somewhat more sophisticated marketing techniques would be used in those national cases than in the local settings.The current data however does not give information about that, but nevertheless poses this question for further research.

Figure 1 .
Figure 1.The use and importance of media platforms

Figure 1 .
Figure 1.Use and importance of media platforms Note: Figure entries show the proportion of candidates that said they used a medium much or very much and of candidates that thought a media platform is rather important or very important.

Figure 2 .
Figure 2. The use of target information and lack of expert help Notes: Figure entries show the proportion of candidates using targeting information from experts or party organizations in their campaign and candidates agreeing that lack of expert help was a liability to their party´s campaign.

Figure 2 .
Figure 2. Use of target information and lack of expert help Note: Figure entries show the proportion of candidates using targeting information from experts or party organizations in their campaign and candidates agreeing that lack of expert help was a liability to their party´s campaign. 17

Figure 3 .
Figure 3.An analytical model showing the main results based on themes extracted and the research questions

Figure 3 .
Figure 3. Analytical model showing the main results based on themes extracted and the research questions

Table 1 .
A multiple linear regression examining the relationship between the independent variables and three dependent variables: social media, the use of target information and the need for expert communication help Note: Significant values are in bold.not at all.In other words, the use of targeting information is greater in Reykjavík and some urban centres in the South-West than in other places in the country.The cities of Reykjavík, Kópavogur and Hafnarfjörður are for all intents and purposes a unified urban area, but Reykjanesbaer on the other hand and indeed Árborg, can, with reference to the

Table 2 .
A multiple linear regression of dummy coded variables using Reykjavík as a reference variable