THE EFFECT OF WEBSITE QUALITYAND ELECTRONIC WORD OF MOUTH ON TRUST TO ENCOURAGE PRODUCT PURCHASE DECISION THROUGH BERRY BENKA (CASE STUDY: BERRY BENKA CONSUMER)

ThisXstudy aims to analyze the influence of website quality, Electronic Word of Mouth, Trust,and purchasingXdecisions. The object of this study was berrybenka consumers, and the number of sempel determined was 210 respondents using a calculation method based on heir formula multiplied by 5. XPurposiveXsamplingXmethod, data collection method using questionnaire, and data analysis usingXPartialXLeast Square (PLS). The results of the study are known that Website Quality affects purchasingXdecisions, ElectronicXWord of Mouth affects purchasingXdecisions, XTrust influences purchasing decisions, TrustXinfluences purchasing decisions, XWebsite QualityXaffects Trust, Electronic WordXofXMouth affectsTrust, Trust plays a role in mediating Website Qualityto consumer decisions, Trust plays a role in mediating Electronic Word of Mouth to consumer decisions.


INTRODUCTION
ThexdevelopmentXofxanxincreasingly modern era encourages a variety of system changes, both directly and indirectly, such as trading systems, ways of transactions and marketingsystems. One type of technology implementation inxterms of improving business, sales and purchasing products is to use electronic commerce to marketxandxbuyxa wide range of products or services. xIndonesia's internet service provider projects that total internet usersxin Indonesia grew by almost 10 percent to 171.17 million people from a totalxpopulation of 262 million people. The largest percentage of consumers in internet users is workers or selfemployed with 82.2xmillion visitors or 62% and the second place is housewives with a percentage of visits from housewives of 22 million people or 16.6%. Meanwhile, compared to visitors among students only 10.3 million people orx7.8% and students who only 8.3 million people or by 6.3% and others only 796 buyers or 0.6. accompanied by the development of berrybenka itself can be seen in the Top Brand Award data in 2017, 2018 and 2019 that is from year to year berrybenka continues to decline and can not increase in 20 200berrybenka no longer enter the top 3xOf Top Brand Index This shows that the trust anddecisions of people shopping diberrybenkaxoffered is decreasing. Based on the results of research on the influence of each variable in accordance with the researchers do still give different results, therefore it is necessary the existence of further research in the hope of being able to explain the causal relationship between each variable that researchersxexaminextoday. xBasedxon the data of pre-survey results, phenomena, and background, researchers are interested in conducting this research. Website Qualitycanxbe seen as anxattribute of a websitexthatxcontributes to its usefulness to consumers (Gregg D. &xWalczak, 2010). xWebsitexQualityAccording to (Hyejeong & Niehm S, 2009) revealed thatxprevious researchersdivided thexdimensionsxofxwebsite quality intoxfive namely InformationxSecurity, xConvenience, xComfort, xand Qualityxof service e. Understanding thexElectronic Wordxof Mouth According tox (Noor, x 2011) saidxElectronic Word of Mouth as a negative or positive statement made by actual consumers, potential or previous consumers about products or companies where this information is available to people or institutions through internet media".

LITERATURE
1. Frequency of accessing information from social networking sites 2. Frequency of interaction with users of social networking sites 3. Number of Reviews written by users of social networking sites.

RESEARCH METHODS
Thexresearchxmethod usedxin this study is to use a survey method which is a data collection method by distributing questionnaires toxrespondentsxRespondentsxinxthis studyxare consumers of berry benka in West Jakata. To obtain data for the preparation of thesis the author took the place of researchxonxconsumersxof Berry Benkaxin west Jakarta. xInxthis study selectedxpopulationxresearch is consumers of berry benka in west Jakarta. In determining the number of samples in this study is five multiplied by the numberxof indicators. Thus, five multiplied by forty-two. So the number of samples in this study is two hundred. xThe data collectionxinstrument used is a questionnaire. xAnalysis method used by data is descriptivexquantitativexanalysis, assisted by IBM SPSSxStatistic 23 and Partial Least Square (PLS) software.

RESULTS AND DISCUSSIONS
Basedxonxthe resultsxofxresearchxconducted on 210 respondents, xnamely Berryxbenka consumers, it can be known the karakerisik picture of respondentsxbasedxonxgender, age, occupation. Based xonxthexresults of data processing showed that out of 210 respondents there were 68 male respoden by 32.4%, and 142 female respondents by 67.6%xand also showed that out of 210 respondents underx20 years old, 77xrespondentsxwith presentations of x36.7%,x26 -30 yearsxas many as 100xrespondents with a presentation of 47.6%, can be concluded in this study mostly aged 26-30 years.
Based on results showed that out of 210 respondents there were 36 respoden working as civil servants, private employees as many as 123 respondents, students or students as many as 11 respondents and the last answered others -another 40 respondents.
Afterxmodification of the construction contained in the modeling Seen that some value loading factor for each indicator of each variablexalreadyxhas axloading factor value that isxnot the most large comparedxto the loading value if connected with other variables. This means that each variable has a good discriminant validity where some variablesxstill have a meter that is highly correlated with other constructs. Itxcan be concluded thatxthe model inxthexconsumerxdecisionxvariable is 0.525 which means thatxthexconsumer decision can be explainedxby the variables in the model namely Trust,ewom, xwebsitexquality, and furthermore the Trust has a valuexof 0.385xwhich means the Trust can be explained by the variables in the model namely ewom, xwebsitexquality.

b. HypotheticalxTestxResults (Line Coefficient Estimation)
Lookxatxthexsignificance of thexhypothesis by looking at the parameter coefficient value and the statistical t significance value in the algorithm boostrappingxreport. To find out significant or insignificantxviewsxof the t-table at alpha 0.05 (5%)x= 1.96. Thenxttablexcomparedxtoxt-count (t-statistics). xHere'sxthextablexof Direct Effects Hypothesis Test Results and Indirect Effects Hypothesis Test Results

A. ThexInfluencexof Electronic Word of Mouth on consumer decisions
Basedxon the hypothesis test in this study, t statistics are greater than t table that is t statistic of four point five hundred and twenty-two greater than one point nine pulan six. Thus, the results of hypothetical tests in this study showed that Electronic Word of Mouth has a positive and significant effect on consumer decisions. xThe hypothetical test results in this study can be interpreted electronic word of mouth communication that occurs between consumers and consumers berrybenka run well, the information conveyed by consumers to other consumers is more directed to positive information and in accordance with consumer expectations after they make a purchase, meaning that consumer confidence about the information provided and services from berrybenka have a positive influence on consumers. The results of hypothetical tests in this study were supported by research conducted by (Jalilvand R & Samiei, 2012) that megatakan Electronic Word of Mouth has a positive and significant effect on trusts,and according to research conductedby (Bhandari & Rodgers, 2017) obtained results that Electronic Word of Mouth has a positive and significant effect on trusts.

B. Trust's Influence on Consumer Decisions
Based on the hypothesis test in this study, showed the results that t statistics are greater than t table that is t statistic of seven point zero thirty greater than one point nine pulan six. Thus, the results of hypothetical tests in this study showed that confidence has a positive and significantxeffect on consumer decisions. xIf thexberryxbenka canximprove and givexgood confidence through every indicator that exists then it will also improve the consumer's decisionxasxwell. xThe hypothetical testxresults in this study were supported by research conducted (Piarna, 2016) which obtained the resultsxthatxconfidencexhas a positive andxsignificant effectx on consumerxdecisions, and furthermore according to (Matute et al., 2016) xtrust has a positive and significant effect on consumer decisions

C. The Influencexof Website Quality on Consumer Decisions
Based on the hypothesis test in this study, t statistics are greater than t table that is t statistic greater than three point eight forty greater than in one point nine pulan six. Thus, the results of hypothetical tests in this study show website qualityhas a positive and significant effect on consumerdecisions. The results of hypothetical tests in this study can be interpreted if the relevant parties can improve the existing Website Qualityby paying attention to every indicator thatexists, it will affect consumer decisions berrybenka. The results of hypothetical tests in this study are supported by research conducted by (Soraya, 2019) who said that Website Qualityhas a positive and significant effect on consumer decisions, and according to research conducted (Giao et al., 2020) which says that Website Quality hasa positive and significant effect onconsumer decisions.

D. Effectxof Website Qualityon Trust
Basedxon the hypothesis test in this study, txstatistics are greater than t table ie t statistic greater six point eight hundredxseventyxfive greater than at one point nine pulan six. Thus the results of hypotheticalxtestsxinxthis study show website quality hasxaxpositive and significant effect on trust. Hypothetical testxresultsxin this studyxcan be interpreted if the relevant parties can improve and utilize thexsupporting factors that encourage consumers to conduct ecommerce activities through a website that is content then a company should be able to present and maintain the quality of the website they have if the company is able to present the website and maintain the quality of the website they have well then directly consumers become trust and conduct transactions online through berrybenka. The results of hypothetical tests in this study were supported by research conducted by (Octavia & Tamarlane, 2017) said that the quality of the website has a positive and significant effect on trust, as well as researchxconductedx (Kurniawan et al., 2018) xwhichxstated thatxthexqualityxof thexwebsite has a positivexand significant effect

E. Effectx of Website Qualityon Trustmediated purchase decisions
Based on the hypothesis test in this study, t statistics are greater than t table that is t statistic greater than three point eight hundred eighty-two than in one point nine pulan six. Thus the hypothetical test results in this study show website quality hasa positive and significant effect on purchasing decisions mediated by the Trust. Test results can be explained if the company already has a good website will further increase trust, and will affect the purchasing decision or can be explained initially Website Quality affectsthe trust and ultimately affect the purchasing decision. The hypothetical test results in this study were supported by research conducted by (Kurniawan et al., 2018) who said that Website Quality influenced thepurchase decision of variablemediated Trust.

F. Effectxof Electronic Word of Mouth on mediated purchase decisions Trust
Basedxon the hypothesis test in this study, statistical t is greater than t table that is t statistic greater than five point three hundredxsixty greater than one point nine pulan six. Thus the hypothetical test results in thisxstudyxshowed that Electronic Word of Mouth positively and significantly influenced the purchasexdecision mediated by the Trust. The results of hypothetical tests in this study can be explained if berrybenkaxalready have good communication and feedback then it will increase the trust they have and will ultimately affect the purchasing decision or can be explained initially Electronic Wordxof Mouth affects the trust and ultimately affect the purchasing decision. The hypothetical test results in this study were supported by research conducted by (Hendratono & Purwanto 2018), who said that ewom influenced the purchase decision of the Variable Mediated Trust.

CONCLUSIONS AND SUGGESTIONS Conclusion
Afterxthe author conducted research and discussion in the previous chapter on "The influence of Website qualityxand electronic word of mouth on trust to encourage product purchasing decisions throughxBerrybenka" then the author concluded as follows: 1. Based on thexresults of hypothetical tests in this study showed website quality has a positive and significant effect on trust. 2. Based on the results of hypothetical testsxin this study showed that electronic word of mouth has a positive and significant effect on trust. 3. Based on the results of hypothetical tests in this study showed that trust has a positive and significant effect on consumer decisions. 4. Basedxon the resultsxof hypothetical tests in this study showed the website quality positively and significantly influential trust. 5. Basedxon the resultsxofxhypothetical tests in this study electronic word of mouth has a positive and significant effect on trust. 6. Basedvon the resultsxof hypothetical tests in this study trust plays a role in mediating quality websites to consumer decisions 7. Based onxthe results ofxhypotheticalxtests in this study trust plays a role in mediating electronic word of mouth to consumer decisions.

Suggestions for Companies:
a. Basedxon the results of the respondent's answer description can be seen from the lowest average value of the variable webstie quality W11 with a value of 3.69 with the questionnaire statement "Berrybenka has an attractive visual appearance", therefore the advice that can be considered is the berrybenka do a re-evaluation of the visual appearance on itsxwebsite, xand after re-evaluatingxmore attention to existing user interfacesxsuch as visual design, colors, layouts, typography.
b. Based on thexresults of the respondent's answer descriptionxcan be seen from the lowest average value of electronic word of mouth EWOM1 with a score ofx3.69 with the questionnaire statement "I know the quality information of Berryxbenka products through social media".xTherefore, the advice that can be considered is expected berryxbenka more use of social media that they have so that theximpactxon consumers is facilitated in finding information on the quality of products they have. c. Basedxon the results of thexrespondent'sxanswerxdescriptionxcanxbexseen from the lowest average value of the T8 trust withxa score of 3.31 withxthexstatementxquestionnairex"I feel that Berrybenka will meet what its customers expect." Therefore, the advice that can be considered is berrybenka through endorsemenxtoxexisting fashionxinfluencers to do promotions through testimonials that they have in order to increase the sense of consumer confidence itself.

Suggestions for Researchers:
Based on R square results of 0.525 which means consumer decisions can be explained by variables in the model, namely trust, electronic word of mouth, website quality while the rest is explained by other variables that are not researched in this model, and suggestions for further researchers can further develop variables and indicatorsthat have not been used in this study such as servicequality, price, and easyness that can affect consumer decisions. In addition, for researchers who will conduct research in the same field and use this thesis as a reference, then it should be reviewed because it does not close the possibility of statements -statements that are not yet appropriate, because researchers feel there are still many shortcomings and limitations in completing this thesis.