Andrew McAfee's Blog: Is craigslist or eHarmony the Right Model for Enterprise 2.0?
Sameer
· 10 months ago
I think paying attention to the audience type in a discussion such as this is really important. For those in the enterprise that consider "browsing" and "discovery" to be an important ingredient (Marketing, R&D, Development) to work success, a free form approach makes a lot of sense so you can make your own judgments about quality and credibility . On the other hand, for a sales rep, standardization of profiles around predefined contexts work much better. So a Twitter or CL model might be too open ended and requires a lot of attention. PP might afford the right balance if it could be fused with predefined tasks ("who's credible around a solution, market, customer, etc). Which leads to me to think Friendfeed has the beginnings of a more promising model than Twitter or CL.
Chris McGrath
· 10 months ago
I think the best method depends on the seriousness of the relationship.
For most business contacts, the Craigslist metaphor is great. It's fun. It's interesting. I don't need to know everything about you.
But if we're going into business together, I like the eHarmony metaphor. I need the specifics, the particulars. I want to make sure I don't miss anything about you.
Interestingly, 236 eHarmony users are married every day in the United States, on average. That accounts for about 2 percent of all marriages in the U.S. Craigslist doesn't keep stats on that kind of thing, but I presume that if they did, they wouldn't really compare in the "seriousness" department.
So can we have both in the enterprise? Start with Craigslist, progress to eHarmony when ready?
(Disclaimer: eHarmony is a customer of my company.)
Chris McGrath
· 10 months ago
I think the best method depends on the seriousness of the relationship.
For most business contacts, the Craigslist metaphor is great. It's fun. It's interesting. I don't need to know everything about you.
But if we're going into business together, I like the eHarmony metaphor. I need the specifics, the particulars. I want to make sure I don't miss anything about you.
Interestingly, 236 eHarmony users are married every day in the United States, on average. That accounts for about 2 percent of all marriages in the U.S. Craigslist doesn't keep stats on that kind of thing, but I presume that if they did, they wouldn't really compare in the "seriousness" department.
So can we have both in the enterprise? Start with Craigslist, progress to eHarmony when ready?
(Disclaimer: eHarmony is a customer of my company.)
Joe Schueller
· 10 months ago
I'm firmly in the CL camp. We've tried too long and too hard to create taxonomies that guess people's needs and intents and miss every time. I now subscribe to the minimal amount of systemic categorization and maximizing people's use of search to let them get what they want when they want it.
This is largely an unpopular stance in IT organizations that tend to have leadership that skews to the more structured/workflow process types.
AndySummers
· 10 months ago
I think there is a third option that lies between these two, but it depends on the feature set of the platform. If the platform delivers a combination of social networking as well as project and content collaboration, the system itself can provide you with a pretty good picture of the person.
For example, suppose I have identified a person as a possible resource for an new initiative. I have visibility to the profile information he has posted, but I also have visibility into the work he has done; projects he has worked, comments he has posted (similar to the DISQUS, for example) , work product he has produced, and people with whom he has worked. If after reviewing these data, he seems to have the expertise I want, I can also see if any of my close connections have worked with him, and I can reach out directly to them for an assessment.
A system like this would not require individuals to enter a bunch of profile information and keep it up to date, but rather because the system is the platform through which they do much of their work and collaboration/communication, the information about that person grows as a natural consequence of their daily activities.
Any system that relies on individuals to provide and keep current information about themselves will ultimately be limited by people failing to provide enough information, gaming the system by spinning what they do provide, and not keeping their information current.
Luc Gendron
· 10 months ago
Refering to your blog's title, what if the best result would come up from a combined approach? A more complete profile datas inspired by eHarmony's model and, while facts reveal more than words, a period of time to study someone actions in Web environments.
Like quantic physics, I rather prefer base my analysis on people's movement than static infos only. I am wondering if eHarmony's algorythm could take care of business interests and caracteristics of potential partners to provide a "perfect" business match?
whoisrtp
· 10 months ago
I don't believe the CL model works well in enterprises, in part because there is just too much information to make enterprise-level search effective and efficient. The productivity loss due to ineffective search is incredible.
I concur with many of Andy S's points. Explicit information provided in corporate profiles/directories is not updated frequently enough (unless made mandatory, which I'm guessing is pretty rare). An employee's social capital provides a great deal of valuable information, including expertise level, communication skills and style, etc. Insight gained from SNA will reveal a more accurate picture of how work really gets done and who the real experts are. Leveraging this is critical going forward as global competition necessitates a stronger focus on HCM, KM, and HR decisions. (Hopefully prediction market data yields this type of insight as well.)
Overall, context is key. The eHarmony model is great if your profile doesn't change, which I assume is the norm in the context of long-term, personal relationships. Answering hundreds of questions might seem excessive, but given eHarmony's success rate, the return on that investment of time is invaluable.
Two of the biggest problems in the enterprise are expertise location and search. The former needs some combination of explicit and implicit information. The latter does as well, however implicit info should carry more weight. Both are contextual activities, but search is more immediate and situational. Looking at the user's search history, web 'footprints', search history of the user's direct connections, etc., will provide much better search results.
The CL approach is a disaster; not only is the user starting from scratch for each search, but the user's results are dependent on the CL poster's ability to enter an accurate description, use keywords and/or acronyms the user understands, etc. This approach is part of the reason that today's enterprise-level search tools are ineffective.
Bob P
Ashwin
· 10 months ago
Hi Andrew,
Thought provoking article. To use a cliche, "Actions speak louder than words". I would rather judge someone by what they do online than they say online. Having said that, there are two hurdles to that approach.
One, actions are generally spaced over time. i.e. What books I read, what blogs I write, what movies I see ... needs to have history before I can collect all that information and make some judgments based on that.
Two, I may not get all this information from someone again because either they don't share that explicitly or there are still technological hurdles that don't let us integration every online activity into every other.
So collecting information about someone is easier, faster and possibly the only way to do it today. But I think as more people start to share what they do online and technology allows us to do that, I see that explicit information capture would not be as necessary.
Gregg Petersmeyer
· 10 months ago
Andy -
Very interesting discussion. There are a few things we at Personal Pathways would like to contribute to the conversation.
In actual practice our profile is somewhere in between CL and eHarmony (we agree with several of the comments that a combination of the two is actually the preferred approach). In contrast to eHarmony, we do not require users to fill out every portion of the profile but rather allow them to fill out as much as they would like when they would like and in the form they would like. In contrast to Craigslist, we do provide more structure to the information in the profile to both 1) help users who are not comfortable starting with a blank text box present information about themselves and 2) solicit information in a form that the system can more readily identify connections among individuals. By combining these approaches, a variety of insights about individuals can be achieved.
Ideally, insights about a person reveal a variety of non-random dimensions to them – dimensions of life and work that matter to another person and can affect the chemistry and effectiveness of a relationship. Therefore, the challenge is to reveal glimpses of a person of a nature and substance that increase the probability of triggering other people or another person to think about them differently than they would have absent this information. Research and experience support the idea that the level of trust and confidence that underlies collaboration that achieves differentiated results depends on knowing one another beyond a superficial level. While that non-superficial level might be gained by either the CL or eHarmony approach, Personal Pathways' approach helps surface meaningful insights about colleagues of sufficient variety to support the most effective collaboration.
We appreciate the framing of the question from you and the comments thus far. We also look forward to more discussion on this topic moving forward and seeing how this topic plays out in the enterprise.
Kelli_ShuterCessna
· 9 months ago
I think Andy S says it well. The answer definitely falls between the two and sounds like PP are doing something along those lines. In my company we've been more toward the CL model - not with any Web 2.0 attempts at social networking - but knowledge management, communities and the collaboration that should come from encouraging communities of practice. I believe you really lose a lot of your valuable information and possible connections when there's little or no structure. Speaking as an eHarmony success story, I can say the structure of the questionaire helped go way beyond the superficial. It definitely required some seriousness of intent. The fact that I didn't have to look at anyone elses 200+ answers was important-that would have lost me immediately, just like weeding though a bunch of search results that had no relevancy or algorithm behind them. The computerized compatibility matching gave me a starting place that would have otherwise taken a lot longer to get to. Then you have the platform to do the free-form interaction. In some cases a little bit of freeform info was enough to decide "no match" and in other cases, just the opposite.
My company recently did an internal network analysis to evaluate the quantity and quality of linkages between employees and we're now looking at the results for trends (e.g. people in knowledge communities tend to network more with other people in communities and we still have a lot of geographic and organizational walls). Next we'll need to come up with some action items to improve the picture. I think one action has to do with giving people more ways to connect using technology since we are so globally dispersed. But I believe that's a big cultural change and a very up-hill battle.
Christoph Schmaltz
· 9 months ago
I agree with Ashwin, that actions are more important than words. We advise organizations that want to build a corporate social network to merge structured and unstructured data in people's profiles. That means, any corporate information, e.g. role, location, contact details should be pulled in from the HR directory. The profile should also include some free-form fields for interests, expertise knowledge etc.
The most part of a profile however, should be dynamic. It should pull in all the actions that the user performs on the corporate systems, i.e. tagging, commenting, bookmarking, favouring, rating, adding information to the CRM, checking in a document into the DMS, making changes on the wiki etc. These actions truly reflect what a person is about, rather than relying on the words that a person uses to describe himself.
Experience shows that people can't be bothered filling in their profile data and even less to keep it up to date. That's why I can't imagine people answering a large number of questions unless it is mandatory. But even then, the role and work of people change. Do they then have to answer those questions again?
I would be very interested to know who the BETA company testers are. I assume you are not allowed to disclose that information, are you?
Justin Cooke Customer Service
· 7 months ago
Very interesting article.
I think that I tend to appreciate the Craigslist approach over eHarmony and I agree with you that standardized questions tend to elicit standardized answers. In thinking about it, I would imagine the time that goes into determining the questions and then how to categorize people based on those answers would be quite intensive. While I think some level of categorizing based on standardized questions may be useful, I'd like to see something that adapts based on a more long-term interaction that isn't quite so black and white.
RJ
· 7 months ago
Having met my wife on eHarmony, I have to say that the algorithms eHarmony came up with do a phenomenal job of matching people based on their personalities. After using numerous dating sites, I finally hit upon eHarmony and found my right match. What eHarmony does is analyze your personality based on the standardized questionnaire. The report they issue is itself very insightful. They illuminate what your most important needs are, and I was surprised to find that through my 14 years of dating I had not learned nearly as much as I did from the results of my eHarmony questionnaire.
The questionnaire gave me new insight into what I needed in order to prosper in a long term relationship. The "other" dating sites allow you to browse through different profiles, but you are subject to your conscience desires, and not driven by your subconscious needs. You may go for a person who is very attractive, but does not have the deep-seated personality traits that make for a long lasting relationship.
My vote is for eHarmony all the way. I don't know if they have divorce statistics yet, but I would bet my savings that the rate is much lower than the alternative means of meeting your true love.
All Comments
For most business contacts, the Craigslist metaphor is great. It's fun. It's interesting. I don't need to know everything about you.
But if we're going into business together, I like the eHarmony metaphor. I need the specifics, the particulars. I want to make sure I don't miss anything about you.
Interestingly, 236 eHarmony users are married every day in the United States, on average. That accounts for about 2 percent of all marriages in the U.S. Craigslist doesn't keep stats on that kind of thing, but I presume that if they did, they wouldn't really compare in the "seriousness" department.
So can we have both in the enterprise? Start with Craigslist, progress to eHarmony when ready?
(Disclaimer: eHarmony is a customer of my company.)
For most business contacts, the Craigslist metaphor is great. It's fun. It's interesting. I don't need to know everything about you.
But if we're going into business together, I like the eHarmony metaphor. I need the specifics, the particulars. I want to make sure I don't miss anything about you.
Interestingly, 236 eHarmony users are married every day in the United States, on average. That accounts for about 2 percent of all marriages in the U.S. Craigslist doesn't keep stats on that kind of thing, but I presume that if they did, they wouldn't really compare in the "seriousness" department.
So can we have both in the enterprise? Start with Craigslist, progress to eHarmony when ready?
(Disclaimer: eHarmony is a customer of my company.)
This is largely an unpopular stance in IT organizations that tend to have leadership that skews to the more structured/workflow process types.
For example, suppose I have identified a person as a possible resource for an new initiative. I have visibility to the profile information he has posted, but I also have visibility into the work he has done; projects he has worked, comments he has posted (similar to the DISQUS, for example) , work product he has produced, and people with whom he has worked. If after reviewing these data, he seems to have the expertise I want, I can also see if any of my close connections have worked with him, and I can reach out directly to them for an assessment.
A system like this would not require individuals to enter a bunch of profile information and keep it up to date, but rather because the system is the platform through which they do much of their work and collaboration/communication, the information about that person grows as a natural consequence of their daily activities.
Any system that relies on individuals to provide and keep current information about themselves will ultimately be limited by people failing to provide enough information, gaming the system by spinning what they do provide, and not keeping their information current.
Like quantic physics, I rather prefer base my analysis on people's movement than static infos only. I am wondering if eHarmony's algorythm could take care of business interests and caracteristics of potential partners to provide a "perfect" business match?
I concur with many of Andy S's points. Explicit information provided in corporate profiles/directories is not updated frequently enough (unless made mandatory, which I'm guessing is pretty rare). An employee's social capital provides a great deal of valuable information, including expertise level, communication skills and style, etc. Insight gained from SNA will reveal a more accurate picture of how work really gets done and who the real experts are. Leveraging this is critical going forward as global competition necessitates a stronger focus on HCM, KM, and HR decisions. (Hopefully prediction market data yields this type of insight as well.)
Overall, context is key. The eHarmony model is great if your profile doesn't change, which I assume is the norm in the context of long-term, personal relationships. Answering hundreds of questions might seem excessive, but given eHarmony's success rate, the return on that investment of time is invaluable.
Two of the biggest problems in the enterprise are expertise location and search. The former needs some combination of explicit and implicit information. The latter does as well, however implicit info should carry more weight. Both are contextual activities, but search is more immediate and situational. Looking at the user's search history, web 'footprints', search history of the user's direct connections, etc., will provide much better search results.
The CL approach is a disaster; not only is the user starting from scratch for each search, but the user's results are dependent on the CL poster's ability to enter an accurate description, use keywords and/or acronyms the user understands, etc. This approach is part of the reason that today's enterprise-level search tools are ineffective.
Bob P
Thought provoking article. To use a cliche, "Actions speak louder than words". I would rather judge someone by what they do online than they say online. Having said that, there are two hurdles to that approach.
One, actions are generally spaced over time. i.e. What books I read, what blogs I write, what movies I see ... needs to have history before I can collect all that information and make some judgments based on that.
Two, I may not get all this information from someone again because either they don't share that explicitly or there are still technological hurdles that don't let us integration every online activity into every other.
So collecting information about someone is easier, faster and possibly the only way to do it today. But I think as more people start to share what they do online and technology allows us to do that, I see that explicit information capture would not be as necessary.
Very interesting discussion. There are a few things we at Personal Pathways would like to contribute to the conversation.
In actual practice our profile is somewhere in between CL and eHarmony (we agree with several of the comments that a combination of the two is actually the preferred approach). In contrast to eHarmony, we do not require users to fill out every portion of the profile but rather allow them to fill out as much as they would like when they would like and in the form they would like. In contrast to Craigslist, we do provide more structure to the information in the profile to both 1) help users who are not comfortable starting with a blank text box present information about themselves and 2) solicit information in a form that the system can more readily identify connections among individuals. By combining these approaches, a variety of insights about individuals can be achieved.
Ideally, insights about a person reveal a variety of non-random dimensions to them – dimensions of life and work that matter to another person and can affect the chemistry and effectiveness of a relationship. Therefore, the challenge is to reveal glimpses of a person of a nature and substance that increase the probability of triggering other people or another person to think about them differently than they would have absent this information. Research and experience support the idea that the level of trust and confidence that underlies collaboration that achieves differentiated results depends on knowing one another beyond a superficial level. While that non-superficial level might be gained by either the CL or eHarmony approach, Personal Pathways' approach helps surface meaningful insights about colleagues of sufficient variety to support the most effective collaboration.
We appreciate the framing of the question from you and the comments thus far. We also look forward to more discussion on this topic moving forward and seeing how this topic plays out in the enterprise.
My company recently did an internal network analysis to evaluate the quantity and quality of linkages between employees and we're now looking at the results for trends (e.g. people in knowledge communities tend to network more with other people in communities and we still have a lot of geographic and organizational walls). Next we'll need to come up with some action items to improve the picture. I think one action has to do with giving people more ways to connect using technology since we are so globally dispersed. But I believe that's a big cultural change and a very up-hill battle.
The most part of a profile however, should be dynamic. It should pull in all the actions that the user performs on the corporate systems, i.e. tagging, commenting, bookmarking, favouring, rating, adding information to the CRM, checking in a document into the DMS, making changes on the wiki etc. These actions truly reflect what a person is about, rather than relying on the words that a person uses to describe himself.
Experience shows that people can't be bothered filling in their profile data and even less to keep it up to date. That's why I can't imagine people answering a large number of questions unless it is mandatory. But even then, the role and work of people change. Do they then have to answer those questions again?
I would be very interested to know who the BETA company testers are. I assume you are not allowed to disclose that information, are you?
I think that I tend to appreciate the Craigslist approach over eHarmony and I agree with you that standardized questions tend to elicit standardized answers. In thinking about it, I would imagine the time that goes into determining the questions and then how to categorize people based on those answers would be quite intensive. While I think some level of categorizing based on standardized questions may be useful, I'd like to see something that adapts based on a more long-term interaction that isn't quite so black and white.
The questionnaire gave me new insight into what I needed in order to prosper in a long term relationship. The "other" dating sites allow you to browse through different profiles, but you are subject to your conscience desires, and not driven by your subconscious needs. You may go for a person who is very attractive, but does not have the deep-seated personality traits that make for a long lasting relationship.
My vote is for eHarmony all the way. I don't know if they have divorce statistics yet, but I would bet my savings that the rate is much lower than the alternative means of meeting your true love.