E-wom's Role in Driving Purchase Intention during Covid-19 Pandemic

: Testimonials are one of the marketing tools that can have a big impact on consumers. Testimony is a form of electronic word of mouth (E-wom) where the information contained in E-wom will be very important as a reference for consumers when they want to make a purchase of a product. This research aims to find out the influence of E-wom quality, E-wom quantity, and E-wom credibility on the purchase intention of fashion products on the online marketplace. The population in this study are people who have visited marketplaces such as tokopedia, shopee, and bukalapak in Surabaya. The sampling method used is purposive sampling, with the number of samples being 385 respondents. Respondent data is collected by distributing questionnaires online. This study used multiple linear regression analysis in SPSS 22 as an analysis tool. The findings of this analysis show that E-wom quality affects purchase intention, E-wom quantity affects purchase intention, and E-wom credibility affects purchase intention.


Research Background
Marketplace provides a great opportunity for an online business because not a few internet users make purchases of a product. This can be shown through data compiled by We Are Social and Hootsuite (2018) which states that as many as 86% of internet users in Indonesia shop online. Of course, it has a positive impact on businesses when marketing products through the marketplace. As is known, a marketplace is a place in the form of a website or application where sellers can create an account and market products to be sold so that it can be used as a facility that supports the existence of buying and selling transactions (Aisyah and Engriani, 2019). The emergence of the marketplace certainly managed to attract the attention of the public for online shopping especially during the current Covid-19 pandemic.
Based on iPrice.com data in 2021, shows the top three marketplaces with the highest number of visitors during the Covid-19 pandemic, namely Tokopedia with 150 million visitors, Shopee with 130 million visitors, and Bukalapak with 30 million visitors. This also shows the increasing awareness of consumers' online shopping, especially during the pandemic. Various policies implemented by the government, such as PSBB, level PPKM, stay at home, work from home, study from home, as well as reducing the use of cash have changed consumer's shopping behavior.
Consumers' awareness and knowledge about shopping online has increased since the Covid-19 pandemic hit because consumers are required to avoid direct physical contact so that in making shopping transactions it is recommended to be able to use online transactions. So that during the COVID-19 pandemic, consumers are faced with alternative choices for contactless and cashless transactions to avoid and break the spread of the Covid-19 virus. This encourages an increase in online shopping using various digital devices as mobile phone and personal computer (databoks.katadata.co.id (2020)). Marketplace takes advantage of this opportunity to further encourage consumer buying interest online by promoting discounted prices and free shipping. Buying interest is a decision made by consumers related to consumer behavior, perceptions, and attitudes where buying interest is influenced by other factors such as price, quality, value, persuasive information, perceived benefits, and information credibility (Lkhaasuren & Nam, 2018;Prendergast, Ko, & Yuen, 2010;Tien, Amaya Rivas, & Liao, 2019). Buying interest arises when there is an intention of solving a problem on potential consumers when choosing and using a product (Imbayani & Gama, 2018). The indicators of purchase intention are the willingness/willingness to buy the product, the willingness to recommend others to buy the product, the willingness to continue using the product even though the price of the product increases, the willingness to buy the product at any time when you need the product, and the willingness to try the product (Prendergast et al., 2010;Lkhaasuren & Nam, 2018;Tien et al., 2019). To support this research, several previous studies were found that became references for researchers related to online buying interest which stated that research related to online buying interest was often associated with e-WOM quality, e-WOM quantity, and e-WOM credibility variables (Aini and Zuliestiana, 2019;Lkhaasuren and Nam, 2018;Mehyar, Saeed, Al-Ja'afreh, & Al-Adaileh, 2020;Tien et al., 2019;Yusuf, Che Hussin, and Busalim, 2018).
When buying products online consumers have several considerations that consumers do not need to waste time in the store, get products at a low price compared to other websites, can avoid pressure when interacting with sellers, and can avoid congestion (Harahap, 2018). Marketplace online shopping makes it easy for consumers to compare the products to be purchased, one of which is by evaluating reviews or testimonials from other buyers. Testimonials or reviews are very important for consumers because testimonials or reviews also help consumers when making decisions to buy a product or service (Syifa et al., 2019). Reviews or testimonials circulating online or in other words are referred to as electronic word of mouth or e-wom. E-wom can be interpreted as a positive or negative statement that has been made by a potential and actual customer of a product or company that has been available to many people (Hennig-Thurau et al., 2004;Yusuf et al., 2018). E-wom has three aspects namely e-wom quality, e-wom quantity, and e-wom credibility (Lkhaasuren and Nam, 2018). E-wom quality is the most important factor that can provide information that is in line with consumer needs such as information provided clear, easy to understand, complete, accurate, timely, and relevant to determine consumer perception of the quality of information as an element to determine buying interest (Tajuddin, Hassan, Razak,and Othman, 2018) (Cheung et al., 2008;Park and Kim, 2008;Teng et al., 2014;Tajuddin et al., 2018;Lkhaasuren and Nam, 2018). E-wom quality can be based on the quality of information and persuasion. The quality of information is acceptable if the information has completed, updated, and relevant details, while persuasion focuses more on the persuasive power of the information received (Tien et al., 2019) (Zhang et al., 2014Tien et al., 2019). Therefore, indicators of e-wom quality are completeness, accuracy, timeliness, relevance, supporting reasons, easy to understand and clear (Cheung et al.,2008;Park and Kim, 2008;Teng et al., 2014;(Tajuddin et al., 2018)Tajuddin et al., 2018Lkhaasuren and Nam, 2018).
E-wom quantity refers to the number of online reviews or comments that are a sign of how important the product is and with the amount of information becomes an important factor for consumers when it needs a reference to strengthen confidence and reduce risk when deciding to buy a product (Lkhaasuren and Nam, 2018). E-wom quantity is many reviews or comments to strengthen consumer confidence that the product has been purchased by many other  consumers and signify the product is popular and important to influence the decision of consumers to buy products and services (Park et al.,2007;Lin et al., 2013;Huyen and Costello, 2017;Lkhaasuren and Nam, 2018). Indicators of ewom quantity are there are many reviews or recommendations about a product online; many people talk about the product online; the number of reviews or comments online that many conclude that the product is popular; the number of reviews or comments online good concluded that the product has a good sale; the number of reviews or comments online that many recommended concluded that the product has a good reputation (Park et al.,2007;Lin et al., 2013;Huyen and Costello, 2017;Lkhaasuren and Nam, 2018). E-wom credibility is a credible source where there are trustworthy reviews or comments ( (Aini & Zuliestiana, 2019;Bataineh, 2015), but if the comment or review is considered less credible, then it does not affect consumers to buy a product (Aini and Zuliestiana, 2019). This is because E-wom credibility is how a person obtains the information that one considers to be trustworthy, true, or factual to influence a person making a purchase decision (Ohanian, 1990;Cheung et al., 2009;Fang, 2014;Lkhaasuren and Nam, 2018;Tien et al., 2019). Indicators of e-wom credibility are reliability of a product review, honest online comments about the product, credible, accurate, and factual (Ohanian, 1990;Cheung et al., 2009;Fang, 2014;Lkhaasuren and Nam, 2018;Tien et al., 2019).
Previous research explained that E-wom is part of the argument quality and source credibility contained in the information adoption model which is a development of the Elaboration Likelihood Model (ELM) theory which simplifies the identification of the determinants of the persuasion process in a more focused context (Shu & Scott, 2014). Information Adoption Model (IAM) can be seen from four components, namely argument quality, source credibility, information usefulness, and information adoption (Sussman and Siegal, 2003). The theory of information adoption model (IAM) explains that the adoption of E-WOM information is part of the persuasion process, where E-wom creates a transfer of basic information and the impact of the information received will be different for each person, but can shape one's interest in the process of adopting information for understand the extent of the influence of information on the person (Cheung et al., 2008;Fang, 2014;Lis, 2013;Tien et al., 2019).
Several previous studies that became references stated that E-wom is often associated with E-wom quality, quantity, credibility, and purchase intention variables. The results of some of these studies state that E-wom quality, Ewom quantity, and E-wom credibility has a positif effect towards online purchase intention (Aini & Zuliestiana, 2019). E-wom quality and E-wom quantity has a positif and significant effect on purchase intention, but E-wom credibility has no effect on consumer buying interest (Mehyar et al., 2020). But contrary to that, Idris (2017) also found that E-wom quality has a positive but not significant effect on buying interest. In addition, E-wom quantity also does not have a significant effect on consumer purchase intentions (Lkhaasuren and Nam, 2018), and whereas consumers really consider reviews before making online purchases, where the presence of E-WOM is stronger than WOM because it offers more business opportunities and a wider range of information for consumers. Previous research has also explained that there are differences between male and female gender in terms of processing information online, where men when they want to buy products online only pay attention to a small part of the available review/comment information, while women focus on all the available information, so that the risk the risk accepted by men is greater than women (Akram, Hui, and Sohaib, 2018). This is become an interesting thing to re-examine, so based on this research gap, this study aims to examine the effect of E-wom quality, E-wom quantity, E-wom credibility to the purchase intention of online marketplace consumers, but this time on during the Covid-19 pandemic.
E-wom quality has a significant influence on purchase intention (Lkhaasuren and Nam, 2018). Therefore, the extent to which the quality of the information provided is useful, clear, and easy to understand, it can be an important element for evaluating consumer buying interest (Mehyar, Saeed, Al-Ja'afreh, and Al-Adaileh, 2020). The quality of E-WOM refers to the persuasive power of reviews or testimonials contained in information messages where the reviews or testimonials have quality and useful information so that consumers can make purchasing decisions (Bhattacherjee and Clive, 2016;Filieri, 2015;Tajuddin et al., 2018). With reviews or comments through various high-quality platforms, of course, having a hand in influencing consumer decision making before making a virtual purchase, it can be said that the quality of E-WOM has a direct influence on consumer purchase intentions (Ismagilova et al. al., 2019;Idris, 2017;Jalilvand and Samiei, 2012). Previous research stated that women are more concerned with detailed and complete information than men who are less focused on detailed information, causing the risk that men accept in the process of buying products online is higher than women (Fan and Miao, 2012;Sohaib et al., 2018).
E-wom quantity has a significant influence on purchase intention (Mehyar et al., 2020). E-wom quantity is an important factor that can affect the evaluation of information and products recommended by other consumers (Tajuddin et al., 2018). When consumers lack knowledge of a product, consumers collect more information about the product to reduce and eliminate the risks that occur (Jeong and Koo, 2015;Tajuddin et al., 2018). The amount of information is an important factor which can be used as a reference to convince and direct consumers when making purchasing decisions to become more rational in buying products and as a sign of how popular the product (Chevalier and Mayzlin, 2006;Kim, Shin, and Lee, 2009;Lkhaasuren and Nam, 2018;Tajuddin et al., 2018). Previous research has also stated that women pay more attention on all the information available online and try to find similarities and differences between the information, whereas men tend to focus on a small amount of information that is considered useful for decision making (Putrevu, 2001;Sohaib et al., 2018). E-wom credibility has a significant influence on purchase intention (Lkhaasuren & Nam, 2018). E-wom credibility is a review or comment where people think the recommendation is trustworthy, true, or factual so that the comment will create buying interest from consumers (Aini and Zuliestiana, 2019;Lkhaasuren and Nam, 2018;Bataineh, 2015;Fogg and Tseng, 1999;). Credibility of information is very important where online recommendations (positive or negative) are the main determinants that influence consumer buying behavior to reduce uncertainty about products and increase credibility and trust (Awad and Ragowsky, 2008;Maia, Lunardi, Longaray, and Munhoz, 2018;Yusuf et al., 2018). Previous research has also stated that women trust more to online recommendations and reviews, where women rely more on existing reviews before deciding than men who focus less on available information (Sohaib et al., 2018)

Method
The population of this research is people who have visited marketplaces such as tokopedia, shopee, and bukalapak. Because the type of population in this study is the infinite population so the large sample in this study is 385 respondents obtained using the formula Davis and Cosenza (1993). The sampling technique used is purposive sampling by questionnaires where samples were selected with criteria determined by researchers based on known populations (Machali, 2018). Therefore, the sample criteria of this study are women and men aged 15-45 years who have visited and searched for products through online marketplaces such as tokopedia, shopee, and bukalapak. Where based on the Association of Indonesian Internet Service Providers (APJII) states that this age is included in the fairly large number of internet users. The first stage of data analysis in this study is the validity test that aims to see and measure whether a questionnaire is valid (Ghozali, 2018). An item is said to be valid if the significant value contained in the Pearson correlation is less than 0.05., and reliability test to measure questionnaires where they are indicators of variables. A questionnaire can be said to be reliable if the respondent's answer to a statement is always the same or fixed from time to time. Reliability measurement is approved with statistics test if Cronbach alpha value > 0.70 (Ghozali, 2018).
The second stage is to test the classical assumption which consists of multicollinearity to test whether there is a correlation between free variables by looking at the tolerance value ≥ 0.10 or the VIF value ≤ 10, heterocedassticity test to test whether residual variants in the regression model have nothing in common on one observation to another (Ghozali, 2018). A model can be said to be good if there is no heteroscedasticity with the sig value in the White test greater than 0.05. Normality Test to test whether, in the regression model, disruptive or residual variables have a normal distribution (Ghozali, 2018). The Kolmogorov-Smirnow test is used to determine whether the data is normally distributed with a sig value of more than 0.05, then the residual can be declared normal. Linearity Test aim to see if the specifications of the model used are linear or not (Ghozali, 2018). A significance value of fewer than 0.05 means there is a linear relationship. The third stage is multiple linear regression analysis to find out the results of the t test, F test, and the coefficient of determination.

Results
The descriptive statistics describe the data based on the results obtained from respondents' answers to each variable measuring indicator. In the E wom Quality (X1) variable, the highest mean value is found in the X1.6 indicator of 3.29, while the lowest mean value is found in the X1.2 indicator with a value of 2.86. In the E wom Quantity (X2) variable, the highest mean value is found in the X2.5 indicator of 3.45, while the lowest mean value is found in the X2.1 indicator with a value of 3.08. In the E wom Credibility (X3) variable, the highest mean value is found in the X3.1 indicator of 3.12, while the lowest mean value is found in the X3.2 indicator with a value of 2.92. And lastly, in the Purchase Intention (Y) variable, the highest mean value is found in the Y5 indicator of 3.31, while the lowest mean value is found in the Y3 indicator with a value of 2.68. Table 1 shows that the category of respondents by gender is dominated by women respondents (73%) which means that women visit online marketplaces more than men. In the age category, most respondents are aged 15-25 years (94%) which can be concluded that the use of online marketplaces such as tokopedia, shopee, and bukalapak more effectively targeting consumers within that age range, then the job category is dominated by students (81%).  Table 2 indicates that the validity test of all items has a significant value of less than 0.05 then it can be concluded that all research instruments are declared valid. Reliability tests showed Cronbach's alpha value was greater than 0.70 where Cronbach's alpha purchase intention is 0.820, the e-wom quality is 0.846, the variable e-wom quantity is 0.790, e-wom credibility variable is 0.874, so the data generated by the questionnaire could be said to be reliable.  Table 3. indicates that the multicollinearity test has a variable tolerance value of e-wom quality (X1) of 0.387, variable e-wom quantity (X2) of 0.447, and variable e-wom credibility (X3) of 0.482 where all calculation results show a number greater than 0.10. The VIF value of the variable e-wom quality (X1) is 2,583, the variable e-wom quantity (X2) is 2,236, and the variable e-wom credibility (X3) is 2,075 which indicates a value of less than 10 for all variables so it is said that there is no multicollinearity or no correlation between free variables in this study. For heteroscedasticity test has e-wom quality (X1) shows the number 0.293, e-wom quantity (X2) shows a value of 0.082, and e-wom credibility (X3) shows a value of 0.175 then in this study there is no heteroscedasticity or residual variant has no similarity on one observation to another where the significant value is more than 0.05. Similarly, the linearity test where the linearity value of the three free variables is 0.000 < 0.05 so that all variables are free i.e.e-wom quality, e-wom quantity, and e-wom credibility has a linear relationship with variables bound to purchase intention. In addition to multicollinearity tests, heteroscedasticity tests, and linearity tests, classic assumption tests also have normality tests. The normality test can be seen in the following Table 5 states that the sig value. (2-tailed) of the Kolmogorov-Smirnov test shows a figure of 0.199 which is worth more than 0.05 so it can be explained that normally distributed data or residual variants have a normal distribution.  Table 5 describes the overall R square value of 0.471 which means that the free variables are e-wom quality, e-wom quantity, and e-wom credibility can explain the bound variable that is purchase intention with a value of 47.1% while the rest is explained by other variables that are not contained in this study.  Table 6 shows that the overall value of the F test significance indicates that the variables e-wom quality, e-wom quantity, and e-wom credibility simultaneously have a significant effect on purchase intention and this model test can be said to be feasible because the significant value indicates a figure of 0.000 where the value is less than 0.05.The Ewom quality indicates a figure of 0.000 where the value is less than 0.05, which means e-wom quality (X1) influences purchase intention. The E-wom quantity indicates the number 0.002 where the value is less than 0.05, which means ewom quantity (X2) influences purchase intention. The E-wom credibility indicates the number 0.000 where the value is smaller than 0.05, meaning e-wom credibility (X3) influences purchase intention. From the results in Table 6 obtained the multiple linear regression analysis equations as follows Y = 2,712 + 0,284X1 + 0,190X2 + 0,236X3 + ε Table 7 shows the results of multiple linear regression studies explain that the constant value (α) is 2,712. This constant value (α) can be concluded that if the free variable is worth 0, then it can be said that the purchase intention variable is worth 2,712. The regression coefficient value in the variable e-wom quality (X1) has a positive value of 0.284 so that it can be concluded that any increase in e-wom quality variables, can increase the purchase intention by 0.284 if the other independent variables remain, which means that the better quality of e-wom's product, the more increase consumer buying interest. The value of the regression coefficient in the variable e-wom quantity (X2) is positive with the number 0.190 so that it can be concluded that each increase in the variable e-wom quantity, can increase the purchase intention by 0.190 if the other independent variables remain, which means that the more quantity of e-wom's product, the higher consumer's buying interest. The value of the regression coefficient on the variable e-wom credibility Volume 9 No 1 2022 Hlm. 44 -54 (X3) is positive with the number 0.236 so that it can be concluded that any increase in the variable e-wom credibility, can increase the purchase intention by 0.236 assuming other independent variables remain, which means that the more credible e-wom's product, the more increased consumer buying interest.  Table 8 shows the significant value of the t-test on the variable e-WOM quality (X1) is 0.012 where the value is less than 0.05, so it can be interpreted that e-WOM quality (X1) has an influence on male gender purchase intention. The significance value of the t-test on the E-WOM quantity (X2) is 0.861 where the value is more than 0.05, which means that the e-WOM quantity (X2) has no effect on male gender purchase intentions. The significance value of the t-test on e-WOM credibility (X3) shows a value of 0.002 where the value is less than 0.05, which means that e-WOM credibility (X3) has an influence on male gender purchase intention. In addition, these results also show that the -test value of e-WOM quality and e-WOM credibility is greater than the t-table value of 1.96, so it can be stated that e-WOM quality and e-WOM credibility have a positive and significant effect on the male gender purchase intention variable. Meanwhile, e-WOM quantity has no significant effect on purchase intention because the t-test value is smaller than ttable.  Table 9 shows the significant value of the t-test on the variable e-WOM quality (X1) is 0.000, e-WOM quantity (X2) is 0.000, and e-WOM Credibility (X3) is 0.001, where all these values is less than 0.05, which means that all these variables have an influence on the purchase intention of women's gender. In addition, based on the results of the of the t-test value above, it states that e-WOM quality, e-WOM quantity, and e-WOM credibility have a positive and significant effect on the dependent variable of female gender purchase intention because each t-value of the three variables is greater than the t table of 1.96.

Effect of E-WOM Quality (X1) on Purchase Intention
This study shows the influence of e-wom quality (X1) on purchase intention (Y) where the value of t-test significance is 0.000 which is less than 0.05. Therefore, the first hypothesis is acceptable in this study. The results of this study are in line with the research of Lkhaasuren and Nam (2018) namely e-wom quality has significance to purchase intention. This is also supported by the opinion of Mehyar et al., (2020) namely the extent to which the quality of the information provided is useful, clear, and easy to understand, it can be an important element to evaluate consumer buying interests. The quality of E-wom refers to the persuasive power of reviews (testimonials) contained in information messages where the reviews (testimonials) have quality and useful information so that consumers can make purchasing decisions (Bhattacherjee & Clive, 2016;Filieri, 2015;Tajuddin et al., 2018). The test on the effect of e-WOM quality (X1) on male gender purchase intention (Y) shows the number 0.012 where this value is less than 0.05 so that hypothesis 1a can be accepted. Likewise, the test of the effect of e-WOM quality (X1) on purchase intention (Y) female gender shows the number 0.000 where this value shows a number less than 0.05 so that hypothesis 1b can be accepted. Based on the significance value of male and female gender, it can be concluded that the effect of e-WOM quality on female gender purchase intention is greater than male gender. This study is in line with previous research where the female gender is more concerned with detailed and complete information than men who are less focused on detailed information, causing the risk that men accept in the process of buying products online is higher than women (J. Fan, Shao, Li, & Huang, 2018);Sohaib et al., (2018). From the results of this study, it can be said that female respondents pay more attention to detailed and detailed testimonials when they want to make a purchase of a product online which is included complete with photos and descriptions of the condition of the product when it is received or used, and there are testimonials that say the product is good so that it makes respondents believe in the testimony. In addition, before buying, women do a testimonial comparison between one store and another. Meanwhile, men are not as detailed as women, if the testimonials seen are good, not careless, honest, and complete, this is enough to decide to buy the product.

Effect of E-WOM Quantity (X2) on Purchase Intention
Based on the significance of the t-test on the variable e-wom quantity is worth 0.002 which is less than 0.05 so that the variable e-wom quantity (X2) affects the purchase intention (Y) overall. Therefore, the second hypothesis in this study is acceptable. The results of this study are following research conducted by Mehyar et al. (2020) where e-wom quantity has a significant influence on purchase intention. The number of testimonials (reviews) becomes an important factor that can be used as a reference to convince consumers when making purchasing decisions to be more rational in buying products and as a sign of how popular the product (Chevalier and Mayzlin, 2006;C. Kim, Mirusmonov, and Lee, 2010;Lkhaasuren and Nam, 2018) .
The test on male gender shows a t-test value of 0.861 which is greater than 0.05 so that the e-WOM quantity (X2) has no effect on purchase intention (Y) for male gender, which means that hypothesis 2a is rejected. For male respondents, the large number of testimonials not a major consideration when respondents want to buy a product, because respondents think that most of the testimonials given are only in the form of star ratings but do not have a complete and clear description, so that the number of testimonials is not a major consideration when respondents want to buy a product. Meanwhile, for female gender, the results show that there is an effect of e-WOM quantity (X2) on female gender purchase intention (Y) with a t-test value of 0.000 which is less than 0.05, so that e-WOM quantity (X2) on purchase intention (Y) gender female, which means that hypothesis 2b is accepted. The results of this study are in line with previous studies where women focus more on all the information available online and try to find similarities and differences between the available information, while men tend to focus on a small portion of information that is considered useful for decision making (Putrevu, 2001;Sohaib et al., 2018). In addition, for women, the large number of positive testimonials can greatly influence their consideration of buying the product because with the large number of testimonials, women feel that the product is good, popular, and many people buy the product. On the other hand, the small number of testimonials, even though they are positive reviews, is still easy to make women feels less confident and hesitant to buy the product.

Effect of E-WOM Credibility on Purchase Intention
The value of the significance t-test of e-wom credibility is 0.000 which is less than 0.05 then the declared variable ewom credibility (X3) influences the purchase intention (Y) as a whole. The results of this study are in line with the research of Lkhaasuren and Nam (2018) namely e-wom credibility has a significant effect on buying interest. Some researchers also support the importance of information credibility where online recommendations (positive or negative) become the main determinants that affect consumer purchasing behavior to reduce uncertainty about the product and Volume 9 No 1 2022 Hlm. 44 -54 increase credibility and trust (Awad and Ragowsky, 2008;Maia et al., 2018;Yusuf et al., 2018). The presence of product photos included in a review further supports e-wom credibility and adds to the belief that the reviews provided are reliable, honest, reliable, and factual. For the male gender, the significant value shows the number 0.002 which means less than 0.05 so that the e-WOM credibility variable (X3) influences purchase intention (Y). The male gender means that hypothesis 3a can be accepted. For male respondents, testimonials that are factual, honest, reliable, and trustworthy can influence the respondent's desire to buy products. Meanwhile, for female gender, there is a significance value of 0.001 which means less than 0.05, so the variable e-WOM credibility (X3) influences purchase intention (Y) for female gender, which means that hypothesis 3b can be accepted. Based on the significance value of male and female gender, it can be stated that the effect of e-WOM credibility on female gender purchase intention is greater than male gender.
This study is in accordance with previous research where women trust online recommendations or reviews more where women rely more on existing reviews before making decisions than men who focus less on available information (Sohaib et al., 2018). For female respondents, the existence of reliable, honest, reliable, and factual testimonials and accompanied by photo documentation of the product can be more convincing to buy the product. Female respondents are more confident in product photos in testimonials given by fellow consumers, rather than product photos from sellers. Therefore, testimonial reviews that are reliable, honest, reliable, factual, and complete with product photo documentation from fellow consumers can influence female respondents more when they want to buy a product.

Conclusion
The conclusion that can be taken in this study is that e-wom quality has a significant effect on purchase intention or in other words the first hypothesis (H1) is accepted, e-wom quantity has a significant effect on purchase intention or in other words the second hypothesis (H2) is accepted, e-wom credibility has a significant effect on the purchase intention or in other words hypothesis (H3) is accepted. In the e-WOM quality variable, a company or a product must have an accurate testimonial or review in accordance with the existing product, the company can remind and ask consumers to provide precise and accurate testimonials about a product. In addition, testimonials must have a complete, relevant content review, and have supporting documentation for the product being reviewed. In the e-WOM Quantity variable, the more testimonials or consumer reviews of a company's product, the more it will attract the attention of consumers, especially female consumers because they think the product is more popular and has a good reputation and sales. In the e-WOM Credibility variable, the testimonials or reviews owned by the company must be honest testimonials from consumers by asking consumers to provide product reviews that are in accordance with the product conditions and or experience when using the product, therefore testimonials can be said to be reliable, credible, accurate, and factual. Overall, to be able to create positive testimonials or reviews, companies must be able to provide good product quality and improve service quality so that consumers feel satisfied with the product and services and become more confidence to share good testimonials or reviews to increase buying interest of other potential consumers.
The results of this study indicate that testimonials or reviews given by consumers is an important thing and become one of the main considerations consumers in encouraging the emergence of purchase intention that is expected to reach the stage of purchase decisions especially during Covid-19 pandemic. So, for every online business that uses the marketplace as a forum for promotion and sales is recommended can encourage every consumer who has purchased to be able to provide a complete review with photos and or videos. This can certainly increase the purchase intention of the next prospective consumer to buy the product.