I recently had an impromptu conversation on the nature of trust in my tumblog with my fellow ‘social catalysts’ (more info coming soon). I’m fairly new to exploring this topic in this way so really appreciated the depth of responses.
The insights I’m left with are that there are essentially 3 dimensions to trust in facilitating a connection. Each dimension is a range that might be expressed as follows.
- The person (from “they’re not a sociopathic stalker” to “they walk on water and you’ll want to be their best friend”)
- The subject (from “i don’t think it will hurt anyone” to “they’re going to save the world and I’m devoting my life to their work”)
- The fit (from “it won’t be a total waste of your xyz” to “it’ll be the best investment of xyz you’ve ever made”)
Of course the scale might be worded a little differently but I think you get the idea.
I wonder, if I received a LinkedIn request with 3 ‘sliders’ on each of these dimensions how would it change the process for me? Would I be able to make decisions much quicker? What if those sliders also came with histories from that recommending party – e.g. where do these slider positions relate to the average positions of their past recommendations? Where does this rank in relation to overall value/or value on each slider compared to other connections they’ve recommended? etc..
And how would it change in what went into making the connection before it got to me? Would there have been more thought? Would that help filter connections? Would I lose some valuable connections somehow?
It’s an interesting area that I’m looking forward to exploring and testing in practice through a project that I hope to announce soon. In the meantime – the more perspectives the merrier.
![Reblog this post [with Zemanta]](http://img.zemanta.com/reblog_c.png?x-id=64478ac3-e3e3-49f9-a8c5-24d08f9625a7)
Trust is a dynamic that can not be quantitatively measured, I expect. Qualitatively it is a subjective thing that also includes how you are feeling at the moment, all of your past experience whether related to the person asking for the connection or not.
I am generally a person who does not like to be bombarded with requests. A guy from the USA is thinking of emigrating to Canada. We met at church. A month or so later he is asking me to help him find housing. I nonchalantly replied yes, but notice that I have not done anything on the request. May well be just not enough trust.
Quantitatively, one contact in this case did not equal trust. In other cases one contact does equal a certain trust. To relate it to your 3 scales – the person created a minor connection; the subject of the request would make some difference to him as an individual but not much else; the fit was pretty minor as I honestly did not know what I would do with the request. I think the person is what makes the difference since I did respond positively to a similar request from someone I knew well.
Interesting thinking.
DH
Interesting – in this case the benefit derives from you interest in
seeing that person's search for housing met. Person you know well –
you are more invested in/interested in seeing their needs met,
therefore your motivation and benefit from acting on the request is
higher. I was trying to get at this in the fit dimension – perhaps it
needs tweaking or maybe there needs to be another dimension?
I certainly agree about the quantification – that's a big reason I
like 'sliders' – helps indicate degrees/relativity/intensity…
I am not sure the 3 dimensions you look at describe trust. What they describe is attention, and the level of importance of the connection. But trust is only “forwarded” through the first one (the person). Then trust is invoved because who made the assessment will contribute to the level of attention you will pay to the connection.
If I make the connection I will tell you that you should pay attention because:
– this is a person I trust
– I think you will be interested and there is a fit
As a result, you may or may not pay attention depending on the context in which the connection was received.
And this will impact the level of trust you will credit me with for the future.
If you thought it was a good connection and my timing was right, then I have gained trust
If it is a good connection but your context has changed, I have not lost your trust
If the connection was not good, then I have lost a few points based on how annoyed you will feel about the whole thing.
Trust includes some level of emotional. If you agree at the rational level that I am a good guy, but the chemistry is not here, then the level of trust is less.
This is why trust does not go too far from one person to the other. If I trust you and you trust Joe, it does not mean I will trust Joe. He will have some level of credit (minimum trust capital), but then he will have to build on it or loose it.
A good place to explore these issues is http://trustmojo.com/ – they did a lot of studying on the issue. I like what they say at the end: architecting for trust is architecting for participation. Getting the architecture right is hard.
Related to this, one tool I like to track reputation (which is an indication of the level of trust that you have within the community) is Venyo.org.
I have started experimenting with their “Vindex” as a way to establish a benchmark among entrepreneurs, I would be interested in your feedback on the idea…
Marc, I hear your disctinctions – great contribution. I'm taking a look through trustmojo – some great stuff there.
One of the key things in this for me is the 'rating' aspect being completed by the person making or transfering the request. This request benefits from addressing those 3 dimensions. Next is the rating of the first and 2nd dimension (person and subject) independent of me. This gets to the independent rating you mention below.
This is interesting too – I will try it out. I'm hesitant about independent indexes – feel a little academic, impersonal, and require another level of verification. Somehow I think indexes should emerge naturally from the context – e.g. ideally if there were a way to be able to 'summon' an index dynamically based on a context that is relevant to the current situation – some sort of semantic based search and trust algorithm. But that's fraught with all sorts of other issues.
If you are interested in the concept of rating a referal, you may want to look at Inquisix (inquisix.com), they are built around something like this. So you get rated on the quality of the intro you make.
interesting. what i don't want is a new place to have to go. i want
something that works with context more deeply as well. are any of
these services getting significant traction as far as you know?
Michael I agree with Marc that taken together your 3 dimensions are related to attention rather than trust. “The fit” is especially about attention. There are lots of people I trust whose interests may not be a perfect fit.
Personal trust (related most to your first dimension) is one of several important factors of attention. And the case of accepting a request is mostly about a decision to give someone and their ideas my attention. E.g., in Duncan Holmes’ case, even if he trusted the requester (for any number of reasons) the request still didn’t seem important enough, primarily because the relationship wasn’t strong enough yet.
But then the semantics can get messier, because you can also “trust” that something or someone will be worth your time to investigate. But in that case, the “trust” is not personal, i.e., about trusting a person; it’s a more logical or intuitive trust that X is a good course to take.
But you can also “trust” a friend or colleague’s judgment about other people and about how well they really understand your interests. So I may trust Bob’s competence and integrity, but still feel that Bob doesn’t understand what I’m doing well enough to recommend that I talk to Harry, whom I’ve never met.
So now, I’m saying a lot of obvious things. It’s complex.
Trusted referral tools like LinkedIn can benefit from some extra meta data to reduce the work involved. But if you try to get it too complex, too few will use it, and those who do use it will use it in potentially confusingly different ways. The secret is to simplify the meta data and the system for using it just enough. I am certain that it can be done and heaven and my friends know I want to try.
In general, I like Michael’s first and third sliders best for helping me decide to accept a referral. The first is about personal trust (or reputation) and the third is about fit. I didn’t completely understand the 2nd dimension, but it seems to also be about fit.
I would want to know those two things from someone forwarding a request: How well do you *know and trust* the requester, and how good a *fit* do you think this is for me. Of those two, the most important for me would be the “know and trust” part, and I would then use my own quick judgment about the fit. I would also assume that my friend wouldn’t have referred something that he thought wasn’t a good fit. And if my friend forwarded something that *wasn’t* a good fit, I would want to let him know more what I’m interested in. (I might also temporarily downgrade the attention I give to his referrals.)
Marc, thanks for the tip about trustmojo which I’ll take a look at. I just recently looked at venyo and it’s interesting.
Not taking a deep look yet, I have questions about how Venyo will make the reputation ratings meaningful and protect them from gaming, or “trust farming” – e.g., getting lots of high ratings from my friends, and rating my friends highly in return. They must be thinking about that and I see that they have a senior research and algorithm guy. Do you know anything about how the algorithms work?
Thanks again to everyone for this in depth conversation.
Let me try and clarify the second dimension (subject). If for
example, someone asks me to pass on an investment opportunity to an
investor I know that investor is also looking to me for my assessment
of the actual investment opportunity. This is my assessment of the
subject which is separate from my assessment of the potential fit of
the opportunity with the investor which is also separate from the
investor's know/trust assessment of me and mine of the prospective
investee.
Thanks. That's helpful. “Good fit” includes multiple factors, including: a) personal fit (including values/style/chemistry), b) viability of the recommended person's idea or venture (related to your “subject”?), c) good fit (relevance) with the target person's interests. However, for usability, I wouldn't try to parse those out as different parameters to rate — or would do so as an option only and not as the default. Or – I would ideally test them with a large group of users.
Great – I like it, simplfying to two with 'subject' being captured in
'fit'.
Marc, I hear your disctinctions – great contribution. I'm taking a look through trustmojo – some great stuff there.
One of the key things in this for me is the 'rating' aspect being completed by the person making or transfering the request. This request benefits from addressing those 3 dimensions. Next is the rating of the first and 2nd dimension (person and subject) independent of me. This gets to the independent rating you mention below.
This is interesting too – I will try it out. I'm hesitant about independent indexes – feel a little academic, impersonal, and require another level of verification. Somehow I think indexes should emerge naturally from the context – e.g. ideally if there were a way to be able to 'summon' an index dynamically based on a context that is relevant to the current situation – some sort of semantic based search and trust algorithm. But that's fraught with all sorts of other issues.
If you are interested in the concept of rating a referal, you may want to look at Inquisix (inquisix.com), they are built around something like this. So you get rated on the quality of the intro you make.
interesting. what i don't want is a new place to have to go. i want
something that works with context more deeply as well. are any of
these services getting significant traction as far as you know?
Michael I agree with Marc that taken together your 3 dimensions are related to attention rather than trust. “The fit” is especially about attention. There are lots of people I trust whose interests may not be a perfect fit.
Personal trust (related most to your first dimension) is one of several important factors of attention. And the case of accepting a request is mostly about a decision to give someone and their ideas my attention. E.g., in Duncan Holmes’ case, even if he trusted the requester (for any number of reasons) the request still didn’t seem important enough, primarily because the relationship wasn’t strong enough yet.
But then the semantics can get messier, because you can also “trust” that something or someone will be worth your time to investigate. But in that case, the “trust” is not personal, i.e., about trusting a person; it’s a more logical or intuitive trust that X is a good course to take.
But you can also “trust” a friend or colleague’s judgment about other people and about how well they really understand your interests. So I may trust Bob’s competence and integrity, but still feel that Bob doesn’t understand what I’m doing well enough to recommend that I talk to Harry, whom I’ve never met.
So now, I’m saying a lot of obvious things. It’s complex.
Trusted referral tools like LinkedIn can benefit from some extra meta data to reduce the work involved. But if you try to get it too complex, too few will use it, and those who do use it will use it in potentially confusingly different ways. The secret is to simplify the meta data and the system for using it just enough. I am certain that it can be done and heaven and my friends know I want to try.
In general, I like Michael’s first and third sliders best for helping me decide to accept a referral. The first is about personal trust (or reputation) and the third is about fit. I didn’t completely understand the 2nd dimension, but it seems to also be about fit.
I would want to know those two things from someone forwarding a request: How well do you *know and trust* the requester, and how good a *fit* do you think this is for me. Of those two, the most important for me would be the “know and trust” part, and I would then use my own quick judgment about the fit. I would also assume that my friend wouldn’t have referred something that he thought wasn’t a good fit. And if my friend forwarded something that *wasn’t* a good fit, I would want to let him know more what I’m interested in. (I might also temporarily downgrade the attention I give to his referrals.)
Marc, thanks for the tip about trustmojo which I’ll take a look at. I just recently looked at venyo and it’s interesting.
Not taking a deep look yet, I have questions about how Venyo will make the reputation ratings meaningful and protect them from gaming, or “trust farming” – e.g., getting lots of high ratings from my friends, and rating my friends highly in return. They must be thinking about that and I see that they have a senior research and algorithm guy. Do you know anything about how the algorithms work?
Thanks again to everyone for this in depth conversation.
Let me try and clarify the second dimension (subject). If for
example, someone asks me to pass on an investment opportunity to an
investor I know that investor is also looking to me for my assessment
of the actual investment opportunity. This is my assessment of the
subject which is separate from my assessment of the potential fit of
the opportunity with the investor which is also separate from the
investor's know/trust assessment of me and mine of the prospective
investee.
Thanks. That's helpful. “Good fit” includes multiple factors, including: a) personal fit (including values/style/chemistry), b) viability of the recommended person's idea or venture (related to your “subject”?), c) good fit (relevance) with the target person's interests. However, for usability, I wouldn't try to parse those out as different parameters to rate — or would do so as an option only and not as the default. Or – I would ideally test them with a large group of users.
Great – I like it, simplfying to two with 'subject' being captured in
'fit'.