White Paper

Trust and the Sharing Economy: A New Business Model

by Charles H. Green

 

 

About Charles H. Green

Charles H. Green is co-author of The Trusted Advisor and The Trusted Advisor Fieldbook, author of Trust-based Selling, and founder of Trusted Advisor Associates.  Based on the trust equation introduced in 2000 in the first book, the Trust Quotient (TQ) is a self-assessment test that has been taken by over 25,000 people, validated by University of Chicago PhD psychometricians, and field-tested over the years in work by clients like Accenture and Deloitte. It is aimed at describing the level of trustworthiness of individuals.

Preface: After many years consulting on trust to major corporations and other organizations, I was asked by an angel investor to look into the sharing economy. More specifically, I was asked by angel investor Miles Spencer and his firm, Vaux les Ventures, to look into the bottlenecks in the sharing economy.

Obviously, Spencer had connected Trust and Sharing long before I had ever heard of the developments in Web 2.0 that were empowering that new eco-system. What followed was a deep dive into many of the participants, and eventually a summary of my findings in this white paper. Full disclosure: I have been engaged by, and have a deep belief in the benefits of TrustCloud, mentioned below. TrustCloud is a portfolio company of Vaux les Ventures.

I. The Economic Context of Trust
“Trust” plays many roles in society, one of them being economic.  As the Sharing Economy has emerged, so have attempts to address the prime friction in the market: how can one take real-world risks with real human assets, based solely upon information found online?

This paper will examine ways to solve that problem from several angles, and propose a path forward for the market.

The Sharing Economy, in my view, stands at the edge of some fundamental economic shifts, with Trust at the center. I start by articulating those shifts, and then describe implications for business models.

A. Trust and Economic Progress

Over the past few centuries, the bulk of economic progress has come from the combination of scale economies and division of labor, relentlessly driving down the cost of everything.

In the 1930s, economist Ronald Coase defined another part of the equation: transaction costs. These are costs which are, by themselves, unproductive, but which enable scale economies: contracts, cost accounting, safety inspectors, for example. As long as transaction costs are exceeded by the savings they create, they’re worthwhile.

By 2005, one study estimated that transaction costs had grown considerably over the 20th century to a level of “from 35 to 60%” of total costs.

Also in 2005, the Boston Consulting Group’s Philip Evans and Bob Wolf wrote an HBR article titled Collaboration Rules, one of the first mainstream business publications to highlight the critical role of trust in lowering transaction costs.

Time-share condominiums lowered the transaction costs of taking vacations – they replaced buying your own property. They are not a new idea, but the transaction costs required to make them work are still large: sales staff, legal contracting, marketing.

By contrast, the next level of transaction cost reduction, Airbnb, lets everyone get into the time-share business, without the transaction costs once deemed necessary.

Rental cars have also been around for decades, but the transaction costs of renting with Hertz are enormous.  Conversely, ZipCar stripped costs radically, making for far greater car utilization.

In both cases, Trust is at the heart of the matter.

B. Trust and the Sharing Economy

The early stages of a future Trust Economy are being seen in today’s Sharing Economy, where early adopters meet low-hanging fruit, offering up massive economic improvements.

The Sharing Economy matches under-utilized assets with potential users. As McDonalds realized in the 70s that it could increase its real estate productivity by 50% just by adding breakfast , so have ZipCar and Airbnb also transformed idle assets (inventory) into economic value – with one difference.

McDonald’s didn’t need to drastically change social behavior or replace transaction costs in order to open for breakfast. But in the Sharing Economy, those are precisely the changes that must take place: people must become comfortable enough trusting others that they will forego the expense of insurance contracts, police, lawyers, security systems and even private ownership, to enjoy the benefits of sharing assets.

As trust mechanisms develop, the first “transaction costs” to disappear will be the redundant ownership of assets. As the mechanisms develop further, other opportunities will arise.

The Sharing Economy is itself a play in a much grander fundamental shift. It is a shift from an infrastructure that protects people from each other, to an infrastructure that helps people trust each other. Getting the fundamentals of trust right will make it easier to navigate a series of reinforcing positive economic waves.

C. Trust Beyond the Sharing Economy

Airbnb and asset rental companies are, in many ways, still working off the paradigm of scale economies, the low-hanging economic fruit of our time.

Trust metrics will offer even greater economic value down the road, when massive economic gains can be reaped by applying them to things like:

  • Improving actuarial science for auto insurance companies
  • Improving management control systems for corporate organizations
  • Changing parole policies for sentencing guidelines in the prison system
  • Altering bonding requirements for positions of responsibility
  • Lowering default rates on credit cards and mortgage lending

At the same time, the Sharing Economy represents a significant and exploding business opportunity here and now, apart from any future developments. 

II. Modeling Trust
The Sharing Economy involves the most accessible examples of transaction cost reductions – situations where players have actually replicated assets in order to prevent the risk of working with others (Airbnb, RelayRides), or where an individual is to be given access to owned assets (TaskRabbit).

In such cases, the transformative event is from sole to shared ownership of (or access to) a previously private asset.

In trust terms, this is a straightforward case. There is a clear trustor, a clear trustee, and the burden of trust rests on the trustee to show trustworthiness. A successful trust model will address this clear archetypal trust opportunity, but with an eye at the same time to future applications of Trust that require more nuance.

A. Near-Misses in the Trust Game

There have been some attempts to quantify social relationships, but these have not precisely addressed Trust.  For example, Reputation.com is really about a person’s bona fides, about what others are saying.  But what people say about someone is at least one step removed the attributes of that person, and at best, only a part of what people consider when deciding to trust someone.

Klout, another popular tool, measures a person’s influence. Being trustworthy may increase someone’s influence – but so can notoriety, publicity, popularity or visibility. None of these are precisely Trust.

Neither reputation nor influence account for all the subtleties that accumulate into trustworthiness, or that help the decision to trust. Trust, reputation and influence are not subcategories of each other, though they all overlap.

There have also been several attempts to create trust-based business, but these have failed. While I can’t comment on the technical merits of those ventures, it’s clear they were doomed on substantive grounds.  Along with the financing and technical, a successful business model must get the Trust aspect right or it will fail.

Here are four potential pitfalls for a Trust business model:

  1. Trust is only weakly transitive. If A trusts B and B trusts C, it follows only weakly that A trusts C. Business models that focus on testimonials (e.g. early RapLeaf) are inherently weak. Behavioral indicators are far superior.
  1. Trust requires context. It makes no sense that say that “Joe is trustworthy” without specifying “to do what?” I may trust my dog with my life – but not with my ham sandwich. Business models that treat Trust as a single attribute will be unable to migrate across multiple business cases.
  2. It takes two to trust. Trust happens when one person (the trustor) trusts, and another (the trustee) is trusted. To say “trust in banks is down” doesn’t tell you whether banks have become less trustworthy, or that customers have become less inclined to trust. Business models that do not clearly distinguish metrics for each role will be erratic.
  3. Trust is risk mitigation. Ronald Reagan famously said, “Trust but verify,” but that’s an obfuscation. Trusting means you don’t verify. Trust itself is risk mitigation. Risk is borne by the trustor, and is largely mitigated by the trustee. Business models that do not clearly identify the risks and the risk-takers will be vague and confusing.   

B. Getting the Trust Model Right

In the Sharing Economy, a good Trust model must take the following points into account:

  • It is the trustor (the owner of the apartment to be rented, the owner of the tools to be shared) that bears the risk of an asset-sharing transaction;
  • It is behavioral data about the trustworthiness of the trustee that is needed to mitigate risk to the trustor;
  • Sharing Economy trust metrics will be context-meaningful. The traits of reliability and integrity to deliver on a commitment will be more relevant for rental businesses; there will be situations, however, where other attributes, like credibility or familiarity, will be more desirable.

C. Examining TrustCloud’s Trust Model

There are many models of Trust. For example, the General Social Survey has massive amounts of data gathered over time, but it’s largely data about the propensity to trust – not about trustworthiness.  In his excellent book, The Decision to Trust , Fordham professor Bob Hurley lists 10 factors affecting trust, ranging from trustworthiness to trusting-ness to environmental factors.

For reasons articulated above, the Sharing Economy needs a clearly defined model of personal (not organizational or environmental) trustworthiness (not trusting-ness). It needs to be behavioral in nature, and offer enough flexibility to be contextually relevant.

This model has been at work in my book, The Trusted Advisor, and within my firm, Trusted Advisor Associates.  Since the 2000 publication of The Trusted Advisor, we have articulated the Trust Equation: a four-factor model of personal trustworthiness. An online version of the equation called the Trust Quotient has been taken by over 25,000 people, validated by University of Chicago PhD psychometricians, and field-tested over the years by clients including Accenture and Deloitte.

In its original form, the model is:
T = (Credibility) + (Reliability) + (Intimacy)

(Self-Orientation)

Where:
T = Trustworthiness
C = Credibility: truthfulness, competence, credentials, transparency
R = Reliability: dependability, track record, integrity, predictability
I = Intimacy: others feel secure sharing information, empathetic, discreet
S = Self-Orientation: focus of attention (self vs. others), selfishness, self-absorption, self-preoccupation

To make the equation more accessible to retail audiences, I suggest restating the equation as four positively-stated variables, with less psychological jargon.

Credibility + Reliability + Familiarity + Sociability
Between these four concepts, nearly all common forms of ‘trustworthiness’ are covered.

D. Trust Players in the Sharing Economy

In the capacity-shifting examples of the Sharing Economy, one person or entity is the owner of the asset – that person is the trustor. The other person is the potential renter/user of the asset – that person is the trustee.

The Trust Question for the trustee is, “How likely is it that the asset I rent will meet my expectations?” It is not a high-risk question, and it can be answered through conventional marketing and media channels.

The Trust Question for the trustor is far more important:  “How likely am I to get my asset returned in good condition?” The inability to answer that question is what’s kept people from inviting strangers into their homes, from lending tools to others, and in general placing their personal belongings and properties at risk to people they don’t know.

The ability to answer this question is critical, and TrustCloud™ is an early leader in the space. To serve the Sharing Economy, data must be provided to the trustor about the trustee. Using the model above, the ideal data will be:

  • About the trustee, not the trustor
  • About the individual’s behavior in context of a transaction
  • Behavioral, not just reputational
  • Indicators that mitigate risk of property damage: primarily reliability and sociability (credibility and familiarity skills are less relevant)

D. TrustCloud’s Value Proposition

Trust metrics productively scale far beyond the capability of any one application. For Airbnb, TaskRabbit and RelayRides to source their own trustworthiness data makes no more sense than for them to build their own tablet computers.

TrustCloud aims to be the FICO Scores of Trust, offering custom-designed metrics to fit a range of trust-improvable segments.

This value proposition is straightforward within the Sharing Economy. It’s also valid for the broader Trust Economy of lowered transaction costs on the horizon – though the particular data relevant for an auto insurance company may well be different from those needed by Airbnb.

TrustCloud’s intent is to build a data collection and translation model capable of scaling through both economies; this offers the flexibility to offer customized Trust metrics to customers with differing needs, within the Sharing Economy and beyond.

III. Managing Two Parties: Trustors and Trustees
Some businesses only have one set of customers to worry about. Others – real estate brokers, executive search firms – must manage relationships with both parties in a transaction. TrustCloud falls into this group.

Selling Trust metrics to Sharing Economy buyers is, like selling market research, not complex. By contrast, the trustee side – the people about whom trustworthiness data will be collected – presents some unique and challenging issues.

A. Selling Data to Trustor Communities

Trustor communities are “in business.” They may be companies looking to rent owned assets; they may be broker businesses looking to facilitate transactions between retail customers; or they may be individuals, acting not as companies but still as economically self-interested parties.

It’s not conceptually difficult to sell to such businesses. It boils down to a simple statement of value: how much economic improvement will the trustor get from the ability to use trustworthiness data about buyers, and what would it cost the trustor to get that data on their own. It may or may not be a simple sale, but the concept is not complex.

A good example of this is the Trust-E website product and how it got acceptance from website owners who desired their “stamp of approval.”

B. Getting Data from Trustee Communities

Getting data from the trustee community is a different matter. The communities are acting almost always not as “businesses,” but as individual consumers. Individual consumers may be economically motivated, but that’s far from the whole story.
Some trustworthiness data is available from publicly available sources. Much, however, requires the tacit, or explicit, permission of those whose behavior is being described. The issues here fall into two types: privacy and motivation.

Privacy Issues.
The highly-public nature of the social web means that plenty of data is available for analysis – and people know that.  It’s critical, however, to respect the privacy settings of sites like Google, LinkedIn and Twitter.  Working within the privacy boundaries of source networks for data and then emphasizing strict privacy safeguards for both the process and the information gleaned should provide some reassurance in this area.

Motivational issues.

 

How will people come to willingly give up information about their own personal behaviors? There are three possible answers:

  1. Financial. The Sharing Economy, and the broader Trust Economy outside it, are based on fundamental economic value shifts, and typically customers as well as sellers will benefit from such savings. As the Sharing Economy develops, trustor businesses can offer discounts or coupons for customers with TrustCloud-certified information. However, that’s a position that must be earned.
  2. Status. Klout has done well following this strategy. It’s a uniquely valid strategy for “influence,” but not for Trust. While there may be status associated with credit FICO scores, it’s the kind of status that is talked about quietly. A status strategy a la platinum credit cards is ultimately not a good fit with Trust; people will want to feel they can improve their trust ratings, not be focused on the appearance of being ‘low-trust, low-status.’ The Status strategy doesn’t feel appropriate.
  3. Gamification. FourSquare and others have succeeded with gamification. Trust actually fits well with game-playing, because of the collaborative aspects of both; hence there is a psychological appropriateness of this strategy. Also, Sharing Economy trustees will be made up heavily of early adopter, techno-savvy people. Given that badges and the like are well-known to the market, and given that this strategy requires the least capital, it’s the most sensible starting gate strategy.

Yahoo’s Developer Network has a useful classification scheme that outlines differing types of motivation.

IV. TrustCloud: the Data
As per the opening of this White Paper, TrustCloud should approach Trust data strategically: mindful of the nature of trust, and the nature of the issues the communities are facing.

TrustCloud’s data collection effort follows three principles:

  1. High volumes of data in low-risk situations, to get started
  2. A wide range of data elements
  3. Clearly defined taxonomies of trust to organize that data.

A. High Volumes of Data for Low Risk Situations
The alternative to high volumes of data is to rely on very precise definitions of data. Yet this approach has two problems. First, the nature of Trust is so situational that it’s better to custom-construct trust definitions to represent specific client situations than to ‘bake in’ a particular definition.

Secondly, becoming a market leader in trust metrics is best accomplished through sheer volume of data collection; in the data business, more is better.

The recommended strategy is to start with “low-risk” transactional situations to build up a track record, and to ensure against highly motivated data attacks. (An example of the opposite is FICO scores, which are aimed at a very few, high-risk transactions, like defaulting on mortgages).

An example of high-volume, low-risk situations is the track record someone develops online in setting appointments, and then meeting them. A track record of meeting stated commitments over time demonstrates a pattern of reliability. For someone to “hack” such metrics would require them to, in effect, change their life, to a life of greater reliability. In such a case, the metrics could not be gamed without “gaming” the life itself – and it would no longer be ‘gaming.’

B. Taxonomy of Data Elements: Internal and External Uses

TrustCloud should use an over-arching taxonomy of Credibility, Reliability, Familiarity and Sociability as a guide to organizing, defining and collecting data.  All of the data collection would fall under one or more of those categories, or combinations of those categories.

Trust Metrics to Trustor Communities.

Trustor communities vary. The trust metrics that Airbnb requires will differ from those that TaskRabbit would require. A vendor stuck in a data policy based on tight definitions, or on reputational rather than behavioral metrics, is unable to respond appropriately.

TrustCloud should have an ability to fine-tune metrics to fit almost any customer need. As a practical matter, it could offer basic “packages” of data which can be further refined.

For example, to an Airbnb-like customer, an “Integrity” package of trust metrics might fit.  It would draw heavily from the Reliability and Sociability metrics.

Besides the ability to develop customized packages, this approach to data allows for recursive learning. One can track performance at the individual indicator level, and tweak definitions to suit the learnings from the actual data.

Such a recursive capability will be difficult to emulate by competitors who focus on fixed definitions, fewer datapoints, and rely more heavily on transitive or reputational metrics.

Trust Metrics to Trustee Communities.

The situation is entirely different for trustee communities. Given this gamification strategy, the creative use of badges is critical. Unlike the trustor community, the trustees have little if any interest in understanding the underlying taxonomy.

As the market matures, the trustor and trustee communities may come into alignment around definitions, a la the FICO score metaphor, but I do not see this happening at the outset.  The industry should evolve by generating infectious enthusiasm among users, while educating them about the economic value of their Trust scores at the same time.

V. Conclusions: Trust Infrastructure to the Sharing Economy
A highly viable potential market exists for the player(s) who can offer a powerful  “trust infrastructure” to the sharing economy. There are three rationales for that statement – economic, competitive, and strategic.

Economic. There exist significant opportunities to release business value that has been trapped by excess capacity and unnecessary transaction costs.  This can be freed by matching trustors and trustees through valid data. Those opportunities exist in cases of dedicated-usage real estate, and of single-owner properties that are used infrequently. Cars, bikes and apartments are only the beginning.

Competitive. Trust metrics, like most data businesses, are highly scalable; it doesn’t make sense to have more than a very few competitors. The rush for scale and market share will be fast, and heated.

Strategic. Defining Trust metrics correctly will be critical. Players who confuse Trust with reputation (thus basing their approaches too heavily on testimonials) will not have a trustable product. Players without contextual definitions of trust will not have a usable product.

The player who can articulate contextual definitions of Trust and back them up with behavioral data will drive the market.

Layers. While a fragmented market is not likely (due to scale issues noted above), there may be room for several layers of players in the market. Someone could play a standards role, or a privacy role (think ICANN, or Trustee); these do not have to be the same player as the provider of trust data.  It would be in the best interests of the trust data provider to collaborate strongly with other such layers, as they can make trust data more viable and widespread.

Long Term. As I argued earlier, the Sharing Economy may be an early phase in a longer-term, and much larger, Trust Economy.  If so, it will be through evolution, not revolution.  The trust data collection efforts of the Sharing Economy are not going to be rendered outdated by a new technology.

To the contrary, success built in the Sharing Economy will probably form the foundation of broader applications of Trust data.  The winner here-now has a much greater chance at longer-term success as well.