"Trust is essential to good market operation because ultimately it decreases operating costs and risk exposure. A well-trusted marketplace will need to spend less in attracting customers and in managing their interactions, which means that it will be much easier to scale."

Marco Iansiti and Roy Levien

Trust is very important to companies that participate in electronic markets and the keystone organizations that operate these markets. No company wishes to deal with a keystone that is not trustworthy or purchase a solution from a supplier that it does not trust. To grow a community, the keystone and the suppliers that are its marketplace members must be trusted.

Providing users of a marketplace with tools to measure trust in suppliers’ solutions may reduce transaction costs and increase the number of deals closed. The objective of this paper is to examine how to measure trust in suppliers’ solutions offered in a marketplace. The discussion on how to measure trust in a keystone is deferred to a later paper.

This paper is organized into six parts. The first part defines trust and the second examines the concept of trust transitivity, which is the use of indirect trust in a trust network. The third part describes how trust can be measured. The fourth provides an overview of the author's research, which examined how the numbers of observations about a solution-supplier’s ability, integrity, and benevolence affect a customer’s uncertainty and belief in the solution offered in the Eclipse Marketplace. The fifth part discusses the implications of this research for keystone operators. The last section provides conclusions and summarizes the relevance of this research.

What is Trust?

Trust is an element of the relationship between two parties, a truster and a trustee. A working definition of trust is drawn from the work by Mayer, Davis, and Schoorman: "Trust is the willingness of a party to be vulnerable to the actions of another party based on the expectation that the other will perform a particular action important to the truster, irrespective of the ability to monitor or control that other party." Trust is, therefore, a willingness to take risk. Moreover, individuals take risks only when they have something to gain.

The three dimensions of trust are: ability, integrity, and benevolence. Ability refers to the truster’s belief that a trustee is able to deliver a given product or service in a given manner. Integrity refers to the truster's belief that the trustee will fulfill agreements as promised. Benevolence is the truster’s belief that the trustee wants to do good for the truster. These dimensions embody the main elements that one can observe about a trustee. One will decide whether or not to trust based on the trustee’s ability to deliver, integrity to deliver, and benevolence to deliver. When the parties have never before had a business interaction, the truster seeks indirect evidence to support a decision.

Trust Transitivity

Trust can be established from the truster’s direct experience with the trustee or can be indirectly derived from recommendations made by others.

Trust transitivity is the use of indirect trust in a trust network. A trust network is established when one party derives indirect trust in a third party through their relationship with someone they trust directly. For example, Alice wants to contract a Java developer, but she does not know anyone who she can really trust who has the required skills. Alice trusts Bob, and Bob knows and trusts Charlie as a good Java developer. The trust network is established: Alice -> Bob -> Charlie. The level of belief or disbelief that Alice has in Bob will contribute to Alice’s final decision whether or not to hire Charlie. Consequently, even though Bob’s trust opinion of Charlie will not be the only factor in Alice’s hiring decision, that opinion will hold a degree of influence in her final trust decision.

In this example, Bob’s direct experience working with Charlie was essential for establishing Alice's indirect trust in Charlie. This concept extends to organizational trust. As Bruce Yandle argues, mechanisms that ensure trust include “the reputational capital of the service firm or owner, and recommendations from other customers [...] that can be assessed directly only through experience.” Therefore, leveraging experience from other customers helps a truster to establish indirect trust in a service or product supplier. But, what if there is no direct experience to draw upon? In the next section, a method of measuring indirect trust by observing a supplier’s interactions is described.

Measuring Trust in a Marketplace

Trust can be measured using an approach called Trust Network Analysis with Subjective Logic (TNA-SL). With TNA-SL, it is possible to calculate values of trust opinions and to combine multiple trust opinion values. In the example above, TNA-SL enables us to measure Alice’s level of belief in Bob and Bob’s level of belief in Charlie, and then to combine them to determine an estimated indirect level of belief of Alice in Charlie. These trust opinion values are measured based on the number of positive and negative observations that Alice has about Bob, and the number of positive and negative observations that Bob has about Charlie.

TNA-SL can be used to measure trust opinion values of marketplace members about products and services offered by suppliers in that marketplace. Trust opinion values can be calculated based on the community members’ feedback about the solutions and services published in the marketplace. It is also possible to measure the trust opinion values about the members themselves based on their activity and participation within the community.

Research

In the author's master’s thesis, TNA-SL was used to examine whether or not providing additional observations on a solution-supplier’s ability, integrity, and benevolence help decrease a customer’s uncertainty and increase a customer’s belief about the supplier’s solution. This research used 621 observations on members and suppliers’ solutions inferred from information collected from the Eclipse Marketplace website. Observations about 232 members and 16 solutions were drawn from two solution categories in the Eclipse Marketplace: Documentation and Team Development.

The research delivered:

1. A model to measure trust in suppliers’ solutions offered in a marketplace.

2. An approach to automatically calculate trust values of solutions offered in a marketplace.

3. Insights on ways to increase trust and decrease a customer’s uncertainty about a solution offered in a marketplace.

The research results suggest that customer uncertainty decreases with additional observations on solution-suppliers. All scenarios we examined show that uncertainty decreases when additional information about a solution supplier’s ability, integrity, and benevolence are made available. On average, uncertainty values decrease 3.23% when observations were added.

Providing additional observations on solution-suppliers does not necessarily increases the belief in a solution. Belief increases or decreases based on the judgment made about such observations to identify them as positive or negative.

Providing ways to allow solution-supplier’s to share knowledge with marketplace members, for example by releasing whitepapers, makes it possible for potential customers to identify which suppliers share and which suppliers do not share information with the community. While the group that shares their knowledge would have their belief value increased, the second group identified as not willing to share would have their belief decreased.

Implications for Keystones

A marketplace is operated by a keystone that supports the platform and the community anchored around the platform. A customer goes to a marketplace when they trust the keystone that is operating it is a legitimate platform operator and community hub. The trust network for a vendor neutral marketplace can be represented as: customer -> keystone -> members -> suppliers.

Keystone operators can take advantage of the research findings described above by providing as much information as possible about their members and about the solutions announced in their marketplaces. This will allow members to make observations along the three dimensions of trust:

1. Ability: the keystone should enable observations that show a solution suppliers’ ability help to increase customer’s belief. Solution-suppliers should provide as much evidence as possible about their ability to deliver a solution.

2. Integrity: keystone operators can enable solution-suppliers to show their integrity to potential customers by providing them with ways that attest their correctness to deliver what they said they would deliver.

3. Benevolence: if the keystone wishes to act in the best interest of their customers, it should provide suppliers with as many mechanisms as possible to allow them to show their benevolence. It is on the suppliers’ hands to use such mechanisms to gain customers’ belief in their solutions. Information that shows a solution-suppliers’ benevolence should always be made public on a marketplace website. Examples include suppliers participating in organizations that advocate on behalf of customers and statements from suppliers that their success depends on their customers’ success.

Conclusion

Trust can be defined as the “willingness to take risk”. This willingness depends upon positive observations about a trustee’s ability, integrity, and benevolence. When no direct observations about a supplier or the supplier’s solution are available, then trust transitivity (relying on the recommendations of others) becomes important. In a marketplace operated by a keystone, the members should be encouraged to provide customers with recommendations of products and services.

The potential benefits of the results of the research described in this paper are:

• reduced transaction costs for members who access a marketplace

• increased number of deals closed by marketplace members

This research is relevant because it provides an approach to calculate trust opinion values of the solutions offered in the marketplace operated by the keystone. These trust opinion values are visible and may be able to reduce members’ search and information costs.

This research is also relevant to the academic community for two reasons. First, the research provides concrete examples of how to calculate trust opinion values in a real marketplace. Second, this research brings together two separate literature streams: trust network analysis with subjective logic model proposed by Jøsang et al. and trust dimensions analysis by Mayer et al.

Finally, this research is relevant to operators of keystones for two reasons. First, the research can guide a keystone to help its members increase trust in their solutions. Second, the research shows how the keystone can increase the trust that organizations place on the solutions offered in its marketplace.

To be trusted, a keystone should provide tools that enable users of its marketplace to assess the ability, integrity, and benevolence of the suppliers of solutions posted in the marketplace. TNA-SL can be used to make visible the level of trust in the services and products published in the marketplace.