Understanding Consumers’ Purchase Intention for Online Paid Knowledge: A Customer Value Perspective

Abstract

Online knowledge platforms have been undergoing a transformation from providing free knowledge to online paid knowledge (OPK). As customers play a key role in the sustainable development and success of the new business model, we focused on the factors that drive consumers’ online knowledge purchase intention. Drawing on the cognitive–affective–conative framework and customer value theory, we propose that consumers rationally evaluate the customer values of OPK in the cognitive stage, followed by generating trust and identification in the affective stage, then leading to a purchase decision. Six factors were extracted from three dimensions of customer value: Functional, emotional, and social values. The hypotheses were tested using survey data obtained from 504 respondents using structural equation modeling. The findings confirm that customer value and identification with the knowledge contributor significantly influence trust in OPK. Trust in OPK and identification with the knowledge contributor both significantly influence purchase intention, whereas trust in the platform neither influences consumers’ trust in OPK nor purchase intention. The findings of this study will help OPK platforms to increase their sales of knowledge products.

1. Introduction

Online paid knowledge (OPK) has attracted many consumers, and the OPK market is growing quickly. The OPK market size in China was about 4.91 billion Yuan (USD $720 million) in 2017, increasing by 300% compared with the previous year [1]. OPK transactions usually occur on online knowledge platforms, which are online communities where users can share, sell, seek, and buy knowledge, such as question and answer (Q&A) sites (e.g., Quora an Zhihu in China) and online course platforms (e.g., Coursera and edX). Originally, the content on these platforms was free, so most platforms’ revenue relied on advertising. However, as more content value and different knowledge platforms emerge, the quality of knowledge may not be maintained. For example, the quality of both questions and answers in Q&A sites seems to have been diluted [2]. As a result, users have to spend more time and effort screening large amounts of free knowledge to obtain the desired knowledge [3]. To attract both knowledge seekers and contributors, some knowledge platforms offer OPK as a service that allows knowledge contributors and platforms to charge knowledge seekers for specific knowledge. Such practice creates a sustainable profit model that creates tripartite benefits to platforms, knowledge contributors, and consumers. For knowledge contributors, they can obtain economic returns and create social reputation by contributing their successful methods, experiences, and unique thoughts [4,5]. For consumers, they can draw value from OPK [6] from subscriptions such as “receiving the best management tips in five minutes”. For platforms, they can receive stable revenue in the forms of direct sales, commissions, value-added services, and advertisement. To better attract and retain consumers, online knowledge platforms as well as knowledge contributors must understand why consumers pay for the knowledge, to help build a profitable and sustainable business model [4,7].

As the OPK phenomenon is relatively new in practice, the psychological mechanism that influences customer purchases of OPK is still unclear. Prior research has mainly focused on users’ knowledge-seeking or adoption behaviors in a free context, which has provided some insights. For example, people who participate in online knowledge searches are more concerned about knowledge characteristics such as specialty, preciseness, and diversity [8]. High-quality information is usually the main driver for people to use a free knowledge search engine or learning systems [9,10,11,12,13,14]. With the development of social media, social factors also became important, such as social support [10] and social responses [14]. People may also want to establish an erudite self-image by showing impressive opinions and spreading useful contents. In doing so, they may get more social support in social interactions. In contrary, if they lack knowledge, they may have negative emotions, such as anxious, in communicating with others. These people need a reliable and high-quality source of knowledge. However, with the explosion of information online, users require much more time to find and examine their wanted knowledge. Current online knowledge platforms allow users to pay for high-quality content. However, as it is a paid service, the extra value customers derive from OPK and how customers make the purchase decision are still unclear.

To answer the above research questions, we investigated the factors influencing consumers’ intention to purchase OPK on online knowledge platforms. First, drawing on customer value theory, we focused on three dimensions of customer value when consumers decide to purchase OPK: Functional, emotional, and social values. Functional value in terms of knowledge quality and price utility is a basic value for which customers pay [15,16]. In addition, we think that OPK could offer extra value that free knowledge may not. In this regard, the emotional and social values may fulfill customers’ higher-level needs, such as enjoyment and receiving comfort from obtaining knowledge, creating a positive social image as well as building social relationships. Second, we attempted to understand the process of a customer’s purchase decision by applying the cognitive-affective- conative framework [17]. According to the framework, the cognitive stage focuses on what values are important when customers consider purchasing OPK and how they rationally evaluate the OPK’s value. The affective stage is the emotional or feeling state. In our context, trust is vital, and customer value may build trust in OPK. As knowledge is subjective and conveys the knowledge contributor’s characteristics, we further examined identification with the knowledge contributor and trust in the platform during the affective stage. Finally, affective factors may influence consumers’ purchase intention (conative stage). The proposed model was tested using survey data in China.

2. Theoretical Background

2.1. Customer Value Theory

Customer value is a customer’s evaluation of a product [18] and needs gratification [16]. Previous studies considered customer value to be an important predictor of customer purchase decision [19,20,21,22]. The value framework incorporates three pertinent dimensions of customer consumption value: Functional, emotional, and social values [23].

Functional value is the perceived utility of a product based on the item’s capacity for functional or utilitarian performance. Quality factors (information quality, service quality, and system quality) are vital determinants of trust [24]. OPK’s functional value may be derived from its characteristics or attributes, including specialty, preciseness, diversity, usefulness, up-to-date information, time-saving, effort-saving, and reasonable price. Hsiao and Chen [15] proved that perceived content and perceived price are determinants of perceived value, whereas technical factors are not important in e-book subscription. Perceived usefulness and perceived fee significantly influence perceived value in paid content [22].

Emotional value is the perceived utility of a product based on the item’s capacity to generate feelings or affective states. Previous studies focused on arousing positive feelings or affective states and perceived enjoyment, and suggested that perceived enjoyment is a vital factor in purchase intention for online content services [22] and digital items in social networking communities [21]. A consumer’s perceived enjoyment is positively associated with the consumer’s intention to use [25,26,27] and trust beliefs [27,28,29]. Rouibah et al. [27] examined the role of perceived enjoyment in trust formation on online payment.

Social value is the perceived utility of a product based on the item’s ability to enhance one’s social self-concept. A consumer is influenced by the way in which they view themselves or wish to be viewed by others. Social value (self-image expression and relationship support) significantly influences purchase intention in social networking communities [21]. Knowledge community users are usually influenced by social factors [4,5,10,14]. Community social responses [14] and users’ community participation [30] help users’ social relationships and influence consumer behavior.

2.2. Cognitive–Affective–Conative Framework

Consumers’ purchase decisions involve three stages: Cognitive, affective, and conative [17]. The cognitive stage is the intellectual, mental, or rational state of a consumer’s mind. The affective stage refers to the emotional or feeling state. The conative stage indicates the striving state, related to the tendency to treat objects as positive or negative goals [17].

In the cognitive stage, consumers perceive the elements related to the product and evaluate the value rationally. Fang et al. [31] examined the role of perceived value in the cognitive stage in an e-commerce context. Cognitive aspects, such as ideas, are important in decision making [32]. Yoo et al. [33] applied the cognition-to-action framework and suggested that quality is an important factor in the cognitive phase.

In the affective stage, consumers generate feelings about OPK, knowledge contributors, and platforms. For example, consumers feel trust if they perceive a product is more valuable [31] or high quality [34]. Identification with the knowledge contributor is consumers’ sense of oneness or belonging with the knowledge contributor’s identity [35], and customer identification with the seller is positively related to purchase intention in the conative stage [36].

3. Research Model and Hypotheses

Figure 1 presents our proposed model in this study. Based on the cognitive–affective–conative framework and customer value theory, we propose that consumers rationally evaluate OPK value in the cognitive stage, including functional value (price utility and knowledge quality), emotional value (perceived enjoyment and anxiety relief), and social value (knowledge-image expression and social relationship support). In the affective stage, we propose that consumers generate feelings or affective states about OPK (trust in OPK), the platform (trust in the platform), and the knowledge contributor (identification with the knowledge contributor). We further propose that perceived customer value in the cognitive stage influences trust in OPK in the affective stage. Affective elements, such as consumers’ trust in and identification with the knowledge contributor, positively impact on purchase intention (conative stage).

3.1. Affective Elements

Trust in paid knowledge is defined as the consumer perception that the OPK provided by the knowledge contributor has characteristics that benefit the consumer. A consumer pays knowledge contributors for useful knowledge, which may save time and effort when seeking needed knowledge due to the vast amounts of available information. The literature indicates that consumers’ trust in the product provided by an e-vendor is associated with purchase intention in e-commerce [37,38,39]. Trust also influences information adoption in the wiki [13]. If a consumer believes the OPK is reliable, they may be more likely to pay for it. Therefore, the following hypothesis is proposed:

Hypothesis 1 (H1).

Trust in OPK provided by the knowledge contributor positively affects intention to purchase OPK.

Trust in the online knowledge platform is defined as the consumer’s perception that the online knowledge platform has characteristics that benefit the consumer [40]. Consumers’ perceptions of OPK may be influenced by the knowledge sources, one of which is the platform. Online knowledge platforms provide places where consumers and knowledge contributors can exchange value and provide the facilities and management to complete a successful transaction. Platforms select or invite professional knowledge contributors and may cooperate with providers to produce paid courses to guarantee consumers’ rights. Trust in the online knowledge platform may influence the consumers’ reliance on the credibility of the OPK. The literature indicates that consumers’ trust in a community or website is associated with purchase intention in e-commerce [41,42,43]. In the context of information or knowledge, trust in the platform influences the intention to obtain information [42,44], adopt information [13], and share knowledge [45]. If consumers do not trust a platform but trust the e-seller, they are less likely to engage in purchase behavior with the e-sellers through the platform and may choose other methods of transacting with them [44]. Therefore, the following hypothesis is proposed:

Hypothesis 2 (H2).

Trust in the platform positively affects intention to purchase OPK.

Perceptions about the website (e.g., usefulness) positively affect trust in the vendor [46]. According to trust transfer theory, a buyer’s trust in a platform positively affects their trust in a seller [37]. The OPK platform is a place where the professionalism of knowledge contributors and the reliability of the OPK can be demonstrated. Site trust affects information trust [47]. If consumers feel the platform is more reliable, consumers tend to trust the OPK on this platform. Therefore, the following hypothesis is proposed:

Hypothesis 3 (H3).

Trust in the platform positively affects trust in the OPK.

Identification with the knowledge contributor refers to the extent to which individuals perceive oneness or belongingness with the knowledge contributor identity [35]. Social identity theory suggests that people typically exceed their identity to develop a social identity, through which they can distinguish themselves from others in social contexts [48]. Consumers are more likely to be attracted by a company identity that has prestige or that is similar to their own identity or has distinctive traits that consumers value. Consumer identification with the company will affect consumer behaviors [49]. A consumer who highly identifies with a seller would have a higher level of purchase intention [36,50] and loyalty [51]. A company may share the same knowledge interest with consumers in the context of OPK, so the knowledge contributor’s identity may attract consumers and lead consumers to pay for OPK. Therefore, the following hypothesis is proposed:

Hypothesis 4 (H4).

Identification with the knowledge contributor positively affects intention to purchase OPK.

Consumer identification with a brand is positively related to trust [52] and resilience to negative information, which means that consumers tolerate defects in information [53]. Online knowledge platforms users tend to have a common identity and purpose, and they feel bonded with the knowledge contributor and other members interested in the same provider [54]. The knowledge contributor and brand may influence consumers. When consumers are attracted by the provider’s identity, consumers tend to believe the OPK is trustworthy and ignore the negative information about the provider or OPK. Therefore, the following hypothesis is proposed:

Hypothesis 5 (H5).

Identification with the knowledge contributor positively affects trust in OPK.

3.2. Customer Value Elements

In this section, we discuss the impacts of customer value elements on trust in paid knowledge. We extract six factors from the three dimensions of customer value. In this research context, we consider knowledge quality and price utility as functional values, perceived enjoyment and anxiety relief as emotional values, and social knowledge-image expression and social relationship support as social values.

Knowledge quality is defined as the perceived degree to which the OPK helps meet needs. Knowledge quality includes the appropriate format (e.g., texts or video) and accurate information. Consumers may perceive higher-quality OPK as containing more useful methods or up-to-date information that save consumer time and effort. Therefore, consumers may tend to believe that the knowledge contributor respects consumers’ rights and payments and offers reliable OPK. Studies suggested that high-quality information enhances users’ trust of knowledge communities and engagement because the information they gain from the communities is useful [10,14,37]. Information quality is an important determinant of online learning usage [9,55]. Perceived information usefulness significantly influences purchase online content services [22]. Studies have shown that quality is vital in trust formation and e-commerce success [24]. Therefore, the following hypothesis is proposed:

Hypothesis 6 (H6).

Knowledge quality positively affects trust in OPK.

Price utility refers to the extent to which a consumer perceives a product is worth the price. As suggested by previous research, perceived price significantly influences purchase intention of online content [15,22]. Pricing policy significantly affects retailer trust [56]. In the context of OPK, when consumers perceive the OPK’s price is rational and the use of money is more effective, consumers tend to believe that the knowledge contributor considers consumers’ rights when offering reliable OPK. Therefore, the following hypothesis is proposed:

Hypothesis 7 (H7).

Price utility positively affects trust in OPK.

Perceived enjoyment is defined as the extent to which a consumer perceives that the activity of acquiring specific OPK is enjoyable in its own right, with no consideration of instrumental benefits from the gained knowledge [25]. Perceived enjoyment is an aspect of perceived value that positively affects purchase intention in the context of online content services [22]. It also positively influences a consumer’s intention to use [25,26,27] and trust beliefs [27,28,29] in online services. Rouibah et al. [27] confirmed the role of perceived enjoyment in trust formation on online payment. In the context of OPK, consumers who perceive enjoyment when using OPK may have an emotional tendency to trust the OPK, which is helpful for establishing a long-term emotional attachment to OPK product. Achieving long-term results by learning OPK often requires consumers’ time and effort, so consumers tend to believe that enjoyable OPK is more likely to be used in the long-term. Therefore, the following hypothesis is proposed:

Hypothesis 8 (H8).

Perceived enjoyment positively affects trust in OPK.

Anxiety relief refers to the extent to which a consumer perceives the activity of purchasing or using specific OPK helps alleviate anxiety. People may experience considerable anxiety regarding success, fierce competition, time pressure, and limited energy. Negative emotions, such as fear and anxiety, significantly influence users’ behavior [57,58]. Negative emotions increase people’s engagement in problem-solving and seeking of utilitarian products [59]. Research suggests that anxiety promotes buying [60]. OPK can help consumers upgrade skills and solve problems, which provides consumers the emotional value of relieving anxiety. Anxiety negatively affects e-trust [28] and website trust beliefs for an unfamiliar site [29]. OPK is a useful tool for relieving consumers’ anxiety, so may influence consumers’ perception of its benefits and credibility. Buying behavior can alleviate negative emotional arousal [61]. Therefore, the following hypothesis is proposed:

Hypothesis 9 (H9).

Anxiety relief positively affects trust in OPK.

Social knowledge-image expression refers to a consumer’s perception of OPK’s capability to build and enhance one’s image of knowledge in the eyes of others. Consumers’ relationships with other consumers positively influence trust [62]. Self-image expression significantly influences purchase intention in social networking communities [21]. OPK may help a consumer to know more than others, be more insightful, show competence at work, etc., which help to build an image of a knowledgeable person. People tend to develop a positive self-concept of being knowledgeable [63] and professional reputations [5]. If consumers believe that OPK can help them create a knowledgeable image, they may perceive the goods are more reliable and worth the investment of time and effort. Therefore, the following hypothesis is proposed:

Hypothesis 10 (H10).

Social knowledge-image expression positively affects trust in OPK.

Social relationship support refers to a consumer’s perception of OPK’s capability to help establish, develop, and enhance interpersonal relationships. Studies suggested that social responses and emotional support affect users’ seeking behaviors and information adoption [10,14]. Consumers’ relationships with other consumers positively influence trust [62]. OPK is a kind of channel that helps consumers understand more topics and obtain up-to-date information when they communicate with others. The OPK themes can help consumers to find social circles where they share similar interests or focus on achieving the same goals. If consumers believe that OPK helps with their social relationships, they may perceive the goods are more reliable. Therefore, the following hypothesis is proposed:

Hypothesis 11 (H11).

Social relationship support positively affects trust in OPK.

We also considered several control variables that might influence OPK purchase behavior: Age, sex, education, and experience with online knowledge platforms.

4. Research Methodology

4.1. Development of Instruments

All the measurement items used in this study were adapted from prior validated scales, as shown in

Table 1. Measurement scales.

Construct	Item	Wording	Sources
Price utility (PU)	PU1	The knowledge product sold offers value for my money.	[23]
	PU2	The knowledge product sold here is appropriate for the price.
Knowledge quality (KQ)	KQ1	The knowledge product sold provides relevant knowledge that meets my needs.	[23,24]
	KQ2	The knowledge product sold presents the knowledge in an appropriate format (e.g., appropriate words, pictures, audio, live example, or summary).
	KQ3	The knowledge product sold has an acceptable standard of quality.
	KQ4	The knowledge product on the online knowledge platform is up-to-date enough for my purposes.
	KQ5	The knowledge product on the online knowledge platform provides me with the precise information I need.
Perceived enjoyment (PE)	PE1	Using the knowledge sold here stimulates my curiosity.	[21,25]
	PE2	Using the knowledge sold here arouses my imagination.
	PE3	I find using knowledge enjoyable.
	PE4	I have fun using knowledge.
Anxiety relief (AR)	AR1	Using the knowledge sold reduces the fear I feel of a lack of knowledge.	[60]
	AR2	Using the knowledge sold reduces my feeling of being upset or feeling panic about the lack of knowledge.
	AR3	Using the knowledge sold reduces my worries about the lack of knowledge.
Social knowledge-image expression (SKE)	SKE1	Using the knowledge sold enhances my image of being knowledgeable to others.	[23]
	SKE2	Using the knowledge sold improves my knowledge-expression to others.
	SKE3	Using the knowledge sold makes a good knowledge impression on others.
	SKE4	Using the knowledge sold improves how I am perceived.
Social relational support (SRS)	SRS1	Using the knowledge sold better enables me to form interpersonal bonds with others.	[21]
	SRS2	Using the knowledge sold helps me maintain my social relationships with others.
	SRS3	Using the knowledge sold helps me make new friends.
	SRS4	Using the knowledge sold enhances my social relationships with others.
Trust in online paid knowledge (TK)	TK1	The performance of this online paid knowledge meets my expectations.	[38,41]
	TK2	This paid knowledge can be relied upon as a good knowledge product.
	TK3	This paid knowledge is a reliable knowledge product.
Trust in the platform (TP)	TP1	This online knowledge platform is a reliable online knowledge platform.	[39]
	TP2	This online knowledge platform gives the impression that it keeps promises and commitments.
	TP3	I believe that this online knowledge platform has my best interests in mind.
Identification with the knowledge contributor (IKC)	IKC1	This knowledge contributor indicates the kind of person I am.	[51,64]
	IKC2	This knowledge product provider’s knowledge-image and my knowledge-image are similar.
	IKC3	I am attached to the knowledge product provider.
	IKC4	The knowledge product provider produces a strong sense of belonging.
Purchase intention(PI)	PI1	I am very likely to buy the knowledge product.	[65]
	PI2	I would consider buying the knowledge product.
	PI3	I intend to buy the knowledge product.

. These items were measured using a seven-point Likert scale ranging from one, strongly disagree, to seven, strongly agree. All the survey items were pilot-tested using samples collected from 43 participants. The results indicated that the measurement model fulfilled the criteria for reliability, convergent validity, and discriminant validity.

4.2. Sample and Procedure

A web-based survey questionnaire was distributed to collect data. As we focused on consumers’ purchase intention in the context of OPK, we limited our respondents to those who have used known online knowledge platforms. We set a screening question to allow only those who had previously used platforms to qualify to participate. Each respondent was asked to select a platform they recently used that could help them recall the experience. Then, we displayed a typical OPK introduction page on the platform they had just chosen. The respondents were asked to answer each question to the extent to which they agreed with each statement. The questionnaire contained a few nominally scaled background questions (

Table 2. Demographic statistics of the respondents.

Demographic Variable	Frequency (n)	Percentage (%)
Sex
Male	257	50.99%
Female	247	49.01%
Age
Less than 18 years	10	1.98%
18–25 years	459	91.07%
26–30 years	31	6.15%
More than 31 years	4	0.79%
Education status
Doctorate degree	7	1.39%
Master degree	166	32.94%
Bachelor degree	323	64.09%
College or less	8	1.59%
Experience with online knowledge platforms
Less than 1 year	175	34.72%
1–2 years	210	41.67%
3–4 years	108	21.43%
More than 5 years	11	2.18%
Monthly spent on paid knowledge on platforms
Less than 50 yuan	292	57.94%
51–100 yuan	141	27.98%
101–300 yuan	56	11.11%
More than 301 yuan	15	2.98%
Usage Frequency
Every day	60	11.91%
Every three days	119	23.61%
Every week	104	20.64%
Every month	93	18.45%
More than a month	128	25.40%
Monthly expenses
Less than 1000 yuan	79	15.68%
1001–2000 yuan	323	64.09%
2001–3000 yuan	71	14.09%
3001–4000 yuan	18	3.57%
More than 4000	13	2.58%

).

We collected data in a university with about 41,000 students. We prepared both online and offline versions of questionnaires and randomly distributed them to students on campus. Each student was only allowed to fill in one survey. From 29 November 2017 to 30 December 2017, 570 questionnaires were returned. We removed some questionnaires that included contradicting responses, as these respondents might not have read and answered the questionnaire carefully. Finally, we had 504 surveys for our data analysis. The descriptive characteristics of the sample are listed in Table 2.

5. Data Analysis and Results

In this research, we analyzed a path model using the partial least squares (PLS) structural equation modeling (SEM) technique using SmartPLS 3.0 (SmartPLS GmbH, Bönningstedt, Germany) [66] to test our hypotheses. PLS was used for data analysis since it places minimum restrictions on samples size, measurement scales, and residual distributions [67,68]. PLS is more appropriate for analyzing complex and larger models [69]. Specifically, the two main reasons for choosing PLS-SEM rather than other SEM technique were: (1) The PLS technique examines the measurement and structural model by ignoring other covariance that is not explicitly stated in the model [70]. This method is particularly useful for estimating the exploratory relationships and for theory building. (2) PLS is regarded as being more suitable when the purpose of the model is prediction rather than testing well-established theory [71]. In our study, we had new constructs, such as anxiety relief, which have not been previously used in the OPK context. We used a two-step approach to conduct data analysis. First, the reliability and construct validity were assessed. Then, the structure model was examined in the second step to test the research hypothesis. In the study, we used SmartPLS 3.0 and SPSS 22.0 (SPSS IBM, New York, NY, USA) in our data analysis.

5.1. Model Fit

Henseler et al. [72] recommend the evaluation of global model fit as the preliminary step in PLS model assessment. They suggested examining three model fit indexes: (1) The standardized root mean squared residual (SRMR), (2) the unweighted least squares discrepancy (d_ULS), and (3) the geodesic discrepancy (d_G). The SRMR is defined as the difference between the observed correlation and the model implied correlation matrix [73]. SRMR values below the threshold of 0.08 are considered a good fit. If d_ULS and d_G exceed bootstrap-based 95(I95) or 99% (HI99) percentiles, the model is likely inaccurate [73]. For our model, the SRMR was 0.05, indicating an acceptable model fit. The value of d_ULS was 1.16, less than the bootstrapped HI 99% of d_ULS. The value of d_G was 0.75, less than the bootstrapped HI 99% of d_G. Together, the results indicate that the model fit the data well.

5.2. Measurement Model

Data analysis was conducted using SmartPLS 3.0. The quality of a measurement model is usually evaluated based on the criteria of reliability, convergent validity, and discriminant validity [74]. Reliability is usually assessed by two indicators: Cronbach’s α and composite reliability.

Table 3. Reliability of measurement items.

Construct	Mean	SD	Cronbach’s α	Composite Reliability	Average Variance Extracted (AVE)
Anxiety relief (AR)	5.24	1.14	0.93	0.95	0.87
Knowledge quality (KQ)	5.04	1.16	0.88	0.91	0.73
Price utility (PU)	5.29	1.09	0.88	0.91	0.67
Perceived enjoyment (PE)	5.24	1.13	0.90	0.93	0.78
Trust in knowledge (TK)	4.72	1.31	0.93	0.96	0.88
Trust in platform (TP)	4.74	1.24	0.87	0.94	0.89
Social knowledge-image expression (SKE)	5.04	1.17	0.92	0.95	0.81
Social relational support (SRS)	4.71	1.26	0.93	0.95	0.83
Identification with knowledge contributor (IKC)	5.33	1.06	0.92	0.95	0.85
Purchase intention (PI)	5.07	1.26	0.84	0.90	0.75

lists the reliability indicators of the constructs. The lowest composite reliability was 0.90, and the lowest Cronbach’s α was 0.84, which is higher than the recommended minimum value of 0.7 [75], indicating the adequate reliability of the measurement for each construct.

Construct validity can be assessed using convergent validity and discriminant validity. Convergent validity is defined as the degree to which the measurement items are related to the construct to which they were theoretically predicted to be related.

Table 4. Factor loadings and cross-loadings. Factor loadings are shown in bold.

	AR	IKC	KQ	PE	PI	PU	SKE	SRS	TK	TP
AR1	0.92	0.35	0.51	0.49	0.30	0.39	0.39	0.30	0.50	0.31
AR2	0.94	0.36	0.42	0.44	0.33	0.35	0.45	0.33	0.48	0.28
AR3	0.95	0.34	0.45	0.42	0.34	0.39	0.40	0.33	0.47	0.29
IKC1	0.36	0.88	0.52	0.43	0.42	0.52	0.44	0.34	0.55	0.42
IKC2	0.29	0.86	0.43	0.33	0.40	0.46	0.43	0.40	0.44	0.34
IKC3	0.35	0.86	0.49	0.41	0.43	0.51	0.43	0.34	0.61	0.40
IKC4	0.25	0.81	0.35	0.31	0.40	0.43	0.40	0.45	0.41	0.31
KQ1	0.45	0.44	0.85	0.47	0.40	0.49	0.35	0.21	0.53	0.33
KQ2	0.41	0.46	0.80	0.46	0.39	0.54	0.35	0.26	0.47	0.36
KQ3	0.43	0.49	0.82	0.49	0.29	0.53	0.39	0.24	0.54	0.41
KQ4	0.35	0.34	0.80	0.42	0.25	0.36	0.31	0.17	0.47	0.24
KQ5	0.39	0.43	0.82	0.43	0.36	0.46	0.40	0.27	0.52	0.38
PE1	0.46	0.40	0.47	0.89	0.35	0.44	0.50	0.36	0.49	0.30
PE2	0.38	0.39	0.44	0.84	0.39	0.41	0.49	0.38	0.47	0.28
PE3	0.42	0.37	0.54	0.89	0.32	0.49	0.47	0.35	0.54	0.38
PE4	0.45	0.39	0.52	0.91	0.34	0.51	0.50	0.36	0.52	0.37
PI1	0.32	0.45	0.37	0.35	0.93	0.47	0.33	0.32	0.52	0.22
PI2	0.32	0.46	0.40	0.39	0.94	0.43	0.34	0.33	0.55	0.24
PI3	0.33	0.45	0.38	0.38	0.94	0.46	0.36	0.35	0.53	0.24
PU1	0.42	0.58	0.59	0.51	0.43	0.95	0.40	0.34	0.57	0.47
PU2	0.34	0.48	0.50	0.47	0.48	0.93	0.34	0.31	0.51	0.41
SKE1	0.40	0.44	0.39	0.48	0.32	0.33	0.88	0.50	0.45	0.31
SKE2	0.42	0.48	0.44	0.55	0.34	0.38	0.91	0.39	0.55	0.33
SKE3	0.40	0.45	0.38	0.48	0.35	0.34	0.92	0.49	0.51	0.32
SKE4	0.38	0.43	0.38	0.48	0.31	0.37	0.89	0.40	0.49	0.30
SRS1	0.31	0.44	0.28	0.38	0.33	0.33	0.49	0.90	0.35	0.29
SRS2	0.32	0.41	0.25	0.37	0.31	0.30	0.47	0.92	0.33	0.25
SRS3	0.28	0.39	0.25	0.36	0.32	0.30	0.41	0.89	0.32	0.24
SRS4	0.33	0.37	0.25	0.38	0.33	0.32	0.42	0.93	0.31	0.23
TK1	0.51	0.55	0.59	0.55	0.57	0.57	0.55	0.37	0.92	0.42
TK2	0.46	0.54	0.54	0.54	0.49	0.51	0.50	0.33	0.92	0.37
TK3	0.46	0.56	0.58	0.49	0.51	0.51	0.50	0.30	0.93	0.44
TP1	0.29	0.40	0.38	0.35	0.20	0.41	0.33	0.25	0.39	0.89
TP2	0.27	0.40	0.38	0.31	0.26	0.43	0.32	0.31	0.38	0.86
TP3	0.26	0.32	0.34	0.33	0.18	0.37	0.25	0.16	0.39	0.86

shows that all of the items exhibited a loading higher than 0.7 for their respective constructs, and Table 3 showed that all the average variances extracted (AVEs) ranged from 0.669 to 0.89 and were higher than 0.5, which is the threshold recommended by Fornell and Larcker [76], indicating good convergent validity.

Discriminant validity refers to the extent to which measures of different model constructs are unique. In this study, we evaluated the discriminant validity by comparing the correlations between constructs and the square root of the AVE of each construct.

Table 5. Correlations between latent variables.

	AR	IKC	KQ	PE	PI	PU	SKE	SRS	TK	TP
AR	0.93
IKC	0.37	0.85
KQ	0.50	0.53	0.82
PE	0.49	0.44	0.56	0.88
PI	0.35	0.49	0.41	0.40	0.94
PU	0.40	0.57	0.58	0.52	0.48	0.94
SKE	0.45	0.50	0.44	0.55	0.37	0.39	0.90
SRS	0.34	0.44	0.28	0.41	0.36	0.34	0.49	0.91
TK	0.52	0.60	0.62	0.57	0.57	0.58	0.56	0.36	0.92
TP	0.31	0.43	0.42	0.38	0.25	0.47	0.35	0.28	0.45	0.87

Notes: Diagonal elements (in bold) are the square root of the average variance extracted (AVE).

lists the correlations among constructs, with the square root of the AVE on the diagonal. The square root of each construct’s AVE is much higher than the construct’s correlations with all of the other constructs, suggesting sufficient discriminant validity [75]. Table 3 shows the cross-loading of the items on all latent constructs used in the model, also indicating reasonable discriminant validity.

We also calculated the variance inflation factor (VIF) values for all the constructs to assess the degree of multicollinearity. We used SPSS 22.0 and conducted a regression analysis by modeling purchase intention as the dependent variable and the other nine variables as the independent variables. Our results show that the VIF did not exceed 2.37, which is well below the suggested threshold of 3–5 [77]. Thus, multicollinearity is not a significant concern.

5.3. Structural Model

The theoretical model and hypothesized relationships were estimated using the SmartPLS 3.0 bootstrap approach with a sample size of 1000 to generate t-values and standard errors for determining the significance of paths in the structural model. The results for the structural model are shown in

Table 6. Hypotheses testing results.

Hypothesis	Relationship between Variables	Path Coefficients (β)	t-Value	p-Value	Testing Results
H1	TK → PI	0.45	7.99	0.00 ***	Supported
H2	TP → PI	−0.06	1.34	0.18	Rejected
H3	TP → TK	0.07	1.33	0.19	Rejected
H4	IKC → PI	0.24	4.79	0.00 ***	Supported
H5	IKC → TK	0.21	2.48	0.01 *	Supported
H6	KQ → TK	0.19	2.00	0.05 *	Supported
H7	PU → TK	0.13	2.17	0.03 *	Supported
H8	PE → TK	0.13	2.25	0.03 *	Supported
H9	AR → TK	0.14	3.29	0.00 **	Supported
H10	SKE → TK	0.18	2.35	0.02 *	Supported
H11	SRS → TK	−0.04	0.60	0.55	Rejected

Note: * p < 0.05, ** p < 0.01, *** p < 0.001.

and Figure 2. For the hypotheses, eight are supported, and three are rejected. The coefficient of determination (R²) indicates that the predictor variables (the external variables) explained 58.148% of TK (trust in OPK)’s variance. TK, IKC (identification with the knowledge contributor), and TP (trust in the platform) explained 36.45% of PI (purchase intention). For the control variables, age, education status, and experience with knowledge platforms were all found to have no significant influence on purchase intention. Sex had a significant influence on purchase intention. We discuss the results in the next section.

6. Discussion and Implications

6.1. Discussion

In this study, we focused on the antecedents of consumers’ purchase intention for OPK, including functional value, emotional value, social value, trust, and identification.

6.1.1. Affective Elements

Our results show that trust in OPK and identification with the knowledge contributor were the key antecedents of purchase intention, whereas trust in the platform was not important. Consumers are concerned about what they pay for. They are willing to buy OPK from which they could benefit and to ensure that the invested time and effort is not wasted. As OPK is subjective as a digital product, trust in OPK is an important factor affecting purchase intention.

Identification with the knowledge contributor positively affected consumers’ purchase intention. This finding is consistent with those reported in previous works [36,50]. Our findings suggest that if a consumer has formed an identification with the knowledge contributor, the consumer is more likely to be attracted to and trust the OPK from the provider. Identification will directly influence the purchase intention.

Trust in the platform had no effect on consumer purchase intention, and this result is contrary to previous e-commerce research that showed trust in the website significantly affects purchase intention [41,42,43]. As OPK delivers methods, experience, and unique insights from the knowledge contributors to consumers, the knowledge content relies on the knowledge contributors instead of platforms. This may explain that trust in the platform is being unimportant. Consumers’ trust in the platform neither influenced consumers’ trust in OPK nor purchase intention, which is in line with the situation for many OPK platforms.

6.1.2. Customer Value Elements

Customer value influenced consumers’ trust in OPK. The results showed that one aspect of functional value, knowledge quality, significantly influenced consumers’ trust in OPK, which is consistent with the findings reported by Shen et al. [13]. Another aspect of functional value, price utility, significant influenced trust in OPK as well, in alignment with previous work [56]. People see functional value as being important when they evaluate the credibility of OPK.

Two aspects of emotional value, perceived enjoyment and anxiety relief, significantly influenced trust in OPK. Learning with enjoyment makes the consumer feel the OPK is beneficial and worth the investment of time and effort, strengthening the consumer’s trust beliefs in OPK. We also added anxiety relief as a new type of emotional value. In line with Chen et al. [59], acquiring useful goods may help consumers solve problems and reduce anxiety. People obtaining OPK can improve skills and solve problems, which may reduce their anxiety about success, fierce competition, time pressure, and limited energy. Whereas previous studies about customer value focused on generating or improving positive feelings (e.g., enjoyment), we also examined how OPK helps relieve negative feelings, i.e., anxiety. To summarize, the enjoyable process of using OPK increases consumers’ trust, which further increases purchase intention. Moreover, anxiety relief influences consumers’ behaviors and increases customer assessment of value.

The findings further revealed that one aspect of social value, social knowledge-image expression, significantly influenced consumers’ trust in OPK, which is consistent with previous work [62]. People who used knowledge communities preferred to develop a positive self-concept of being knowledgeable [63] and maintain their professional reputations [5]. OPK can help consumers to solve problems, improve skills, and know more than others, which will make a consumer feel knowledgeable and be regarded as knowledgeable by others. Unexpectedly, another aspect of social value, social relational support, had no significant effect on trust in OPK, and this result is contrary to that reported by Habibi et al. [62]. One explanation could be that people see knowledge-image expression as being more important than social relationship support when they decide to purchase OPK. When consumers evaluate OPK, they pay more attention to its performance: Better quality and better self-image.

6.2. Implications for Research

The findings provide four theoretical contributions. First, this study is one of the pioneering studies focusing on consumers’ purchase of OPK, and we first integrated elements relevant to consumers, knowledge contributors, and platforms into one model. Whereas prior studies paid more attention to user behavior (e.g., knowledge seeking) in a free knowledge context, we focused on purchase behavior in a paid context.

Second, this study is the first to apply customer value theory to the OPK context to understand consumers’ OPK purchase intention. Previous studies used social exchange theory [4,14], social support theory [10], expectation confirmation theory [78], information adoption model [10,13], social capital theory [5,79], motivation theory [5,7], and self-discrepancy theory [63] from a personality traits perspective [80] to investigate why users use knowledge platforms. In terms of paid knowledge, little research examined knowledge as a commodity, and the reasons why consumers purchase OPK remain unclear. We applied customer value theory to focus on what value attracts consumers. The findings verified that functional value (represented by knowledge quality and price utility), emotional value (represented by perceived enjoyment and anxiety relief), and social value (represented by social knowledge-image expression) are important factors influencing consumer behaviors.

Third, we added anxiety relief as a new aspect of emotional value, which extends the customer value theory to the OKP context. We suggest that the unique value of paid knowledge, anxiety relief, is a particular aspect of emotional value influencing consumer behavior, thereby extending customer value theory and expanding studies on OPK. Previous studies of customer value focused on driving positive feelings (e.g., enjoyment) [16,22], but overlooked the role of reducing negative feelings (e.g., anxiety). According to previous studies, negative emotions (e.g., anxiety) would increase people’s engagement in problem-solving and the willingness to obtain utilitarian products [59], and tension release significantly influences consumer behavior [58]. Our results verified that anxiety relief is an important value influencing consumers’ trust in OPK.

Fourth, the findings suggest trust in OPK is an important antecedent of purchase intention for OPK, and two types of trust have different impacts. As OPK delivers the knowledge contributor’s personally successful methods, experience, and unique insights, consumers’ purchases are determined by knowledge contributors’ characteristics, which may explain that trust in the platform is being unimportant. Therefore, different types of trust have different underlying mechanisms influencing purchase intention. This finding should be considered in future studies, suggesting that we should pay more attention to OPK and knowledge contributors rather than platforms. We investigated trust in OPK from a customer value perspective based on knowledge characteristics.

6.3. Implications for Practice

The findings have practical implications for both knowledge contributors and platforms. For knowledge contributors, they could focus on the important values that benefit consumers and express their social identities. First, knowledge contributors may provide accurate, professional, and unique OPK at a rational price, and emphasize the specific skills or problems they may provide or help address. Second, knowledge contributors may make the use of OPK enjoyable since an enjoyable process improves consumer trust in OPK, so they are more willing to and sure about investing time and effort in learning. Knowledge contributors can focus on conveying utilitarian knowledge and emphasizing that this knowledge can help consumers relieve anxiety, such as by improving competition with others, reducing career pressure, and efficiently using limited time and limited energy. Third, they may emphasize that the OPK can help consumers establish a positive knowledge image, meaning that consumers view themselves better or are viewed better by others by using OPK. Fourth, knowledge contributors should express a distinct social identity and use each characteristic to attract consumers and gain consumers’ trust. Several methods can be used to express their social identities, such as listing jobs to attract consumers with similar jobs, expressing their values and knowledge pursuits to attract like-minded consumers, and describing their professional reputation in relevant fields.

For platforms, they could focus on the important values that benefit consumers, inviting knowledge contributors with distinct images and known reputations in relevant fields, helping knowledge contributors express their social identities, and retaining existing knowledge contributors. First, regarding knowledge content generation, platforms could provide management to ensure knowledge quality and reasonable price, and introduce stimulus measures to encourage knowledge contributors to provide high-value knowledge. Platforms may cooperate with knowledge contributors to offer high quality and high price utility OPK. Second, platforms may provide more specific product introduction forms to help knowledge contributors emphasize the OPK’s specific value and recommend knowledge to quickly meet consumers’ needs. Templates introducing functional, emotional, and social values mentioned above in the introduction should be provided and price brackets and categories could be used to quickly match the different needs of the customers, such as high-quality information, cost-effective, most enjoyable, anxiety-relieving information, and better knowledge-image. Third, platforms could pay more attention to inviting and retaining good knowledge contributors with distinct images and known reputations in the relevant fields, like society celebrities and experts, and help knowledge contributors better express their social identity characteristics to attract consumers. Fourth, platforms could present their OPK to consumers through personalized matching based on consumers’ social identities. They may display different categories and different OPK in different orders for each consumer according to their identity characteristics, such as career, followed users, and followed knowledge area. Each consumer could have a unique OPK page.

7. Limitations and Future Research

This study has three limitations. First, although studying purchase intention is appropriate in the focal context, a need remains to explore actual purchase behavior on OPK. Future research can use transaction data to confirm actual purchase behavior. Second, the sample was limited to university students. Although the university students are the majority of online knowledge platforms users, they may not precisely reflect the whole population. Conducting a study to cover larger populations is an option for future research. Third, the sample was limited to OPK platform users in China; therefore, the generalizability of the findings to other countries may be limited. There might be culture differences, for example, in customer value. Future research can examine if users in different countries have different costumer values.

Future research should search for additional antecedents and moderators that influence consumers’ purchase intention for OPK, such as personality traits (e.g., innovativeness [80]), habit [19,20], and epistemic curiosity [8]. Future research could examine other customer values and explore how personality traits affect purchase intention.

8. Conclusions

OPK is an emerging avenue for consumers to obtain knowledge and a new revenue source for knowledge platforms and providers. Figuring out why consumers purchase OPK is important for the success of knowledge platforms and providers. In this study, we integrated customer value theory and the cognitive–affective–conative framework to examine consumers’ purchase intention. The results showed that customer value, trust, and identification are important antecedents affecting consumers’ OPK purchase intention. Online consumers’ purchase intentions are influenced by their trust in OPK and identification with the knowledge contributor, and consumers’ trust in OPK is influenced by customer value and identification. Unexpectedly, the trust in the platform does not influence consumers’ trust in OPK or purchase intention. Specifically, knowledge quality and price utility are important functional values that influence consumers’ trust in OPK. Both enjoyment and anxiety relief are important emotional values that positively influence consumers’ trust in OPK. Social relationship support (an aspect of social value) also positively influences consumers’ trust in OPK. The findings have several important theoretical and practical implications for consumers’ OPK purchase behaviors.