Expertise Finders: Illumio, Privacy, and Better Ways to Find People

Illumio is based on Tacit’s data mining tool. It tries to achieve a delicate balance between (a) unearthing relevant documents and e-mail relationships by scouring subscribers’ hard drives, and (b) protecting subscribers’ privacy. Here’s how it works:

The system is reciprocal: You can make requests of those in your Groups, and they can make requests of you. On the surface, it seems innocent enough. The Desktop search tools it uses can be user-controlled to limit the search to only selected folders, so if you have confidential information it won’t be visible to Illumio. And while it looks through your e-mail archive and address book for names and other business-card information, it doesn’t retrieve entire e-mails. However…

One of the investors in Tacit is the CIA, through a little-known government spy tech company called In-Q-Tel. This is more than a supplier-customer relationship, and it’s a bit disturbing that the NYT didn’t talk about it in their article. The fact that the Illumio logo is an eyeball staring through a keyhole doesn’t help either. Given the Bush administration’s propensity for illegal wiretapping, and (we find out today) preemptively exempting itself from over 750 laws, why should we believe Tacit president David Gilmour’s assurances that Illumio won’t be used to scour our hard drives for anything the government deems ‘significant’? The short answer is: We probably shouldn’t. It’s a shame that Tacit chose to get into bed with the devil, because even if this software is clean, it is now suspect. 

So the question becomes, if a company without dubious government connections were to introduce a tool like Illumio, (a) would people trust it, and (b) would it do the job of finding needed “know-who”?

My answer to the first question is probably not. Software like this inevitably runs into the conundrum of reciprocity: We all want to access other people’s information, but we don’t want others (especially people we don’t know well) to access ours (without our specific, one-off approval, anyway). File sharers have gotten over this. After initially turning off uploading so they can download files without reciprocating, they get a twinge of guilt, double-check their anti-virus software, and turn on uploading. When nothing bad happens, they breathe a sigh of relief and leave the door open, confident that only that folder labeled Shared can be accessed by others.

Instant messaging involves less trust than file-sharing, but it too makes some people nervous. Some people just prefer to have an ‘unlisted number’: They can dial out, but no one can dial in. Within corporate firewalls, where it’s only the company data at risk, and where the IT department can be blamed if anything goes wrong, Tacit’s tools have been used for years. But even in such controlled situations, many companies have outright rejected any tool that would open portals to anything on personal hard drives. Of the six large companies I know best, one is using Tacit and the other five looked at it and rejected it on security/privacy grounds.

Even cookies, those undecipherable pieces of code that allow websites you access to recognize you automatically without having to re-key ID and passwords, are viewed with great suspicion by many.

I suggested in a recent post that, starting with file-sharers and cookie-users and IMers, we could establish an open peer-to-peer expertise finder by opening up access to our Address Books alone. This would involve conscripting certain fields in the Address Book record to be used to rate our own and others’ expertise according to our own personal folksonomy. Then, companies with search expertise (Google, Yahoo etc.) could mine that information (and only that information) to create “know-who” search tools. At the time I believed this made more sense than a top-down ‘managed’ social networking app, no matter how brilliantly thought out, that required us to maintain information in some central place, and to learn to use a fairly sophisticated tool. We’ve tried that, and it doesn’t work. Tragedy of the Commons, too busy being inefficient to learn how to be efficient, and all that.

Let’s go back to the problem we’re trying to solve. Say I’m looking for:

1. Local Suppliers: Someone to supply and install (or teach me to install) a solar swimming pool heater;
2. Business Partners: A group of people to form an enterprise that would teach people, organizations and communities how to live more simply, sustainably, and self-sufficiently;
3. Customers: Some customers who need or would value the above enterprise;
4. Life Partners: Members for a new intentional community to be located in place X; and
5. Virtual Suppliers: Someone to migrate my blog over to a new platform.

If I managed to get all 1850 people in my address book into appropriate Groups, how would an Illumio-type product handle these “know-who” searches?

The obvious thing about these real-life examples is that finding an ‘expert’ is all about establishing a relationship, and that requires some ice-breaking, nurturing, and evaluation that is as much emotional as rational. I’ve won clients and assignments that I wasn’t the most expert at, simply because I already had a pre-existing relationship, or because the initial chemistry was good. And vice versa. It’s all about trust.

Illumio attempts to use my Personal Groups as the judge of that chemistry: If it ‘decides’ that Jo Smith is a candidate for one of the five searches listed above, it will first ping the people in my Groups where that name turned up, asking them to give me an introduction to Jo (and, by implication, to confirm that, in their trusted judgement, Jo would in fact be a good candidate).

My guess is that, because search #1 requires someone local and quite specialized (and few people in my address book are local), Illumio would turn up no Jo for #1. An e-mail to any of three of the Yahoo/Google groups I belong to would probably be faster, more effective, and easier for everyone.

And because #2, #3 and #4 are so broad, anything Illumio came up with for these searches would, I suspect, be useless, little better than random picks from all the FOAFs of the people in my address book. At best, the request might pique the curiosity of the people in my address book (or at least the 150 I know well ;-) sufficiently to get them to recommend someone, drawing on vastly more information than anything that could be found on their hard drives. And I could get the same thing by just sending an e-mail to those 150 people.

Illumio would probably come up with a bumper crop of candidates for #5, but then so did my mere mention of this possibility on my blog. Until we get much further into a Gift Economy, the need for tools like this to find peer assistance will not be great.

I can see more value for Illumio in business organizations that need to draw on expertise outside their organizations (because the expertise is not present or is not immediately available) inside. These would be one-off situations with short time horizons like the need for a facilitator or subject matter expert. So for organizations already using Tacit’s product internally, I can see the value of them encouraging people in their employees’ external networks to sign up for Illumio. That, I think, is the real market for this product.

I confess that the address book-based expertise finder I proposed in my earlier article wouldn’t do any better than Illumio for the five “know-who” searches above. It suffers from the same limitations — too impersonal and too little context. And it’s too far ahead of the peer production / peer assistance curve.

So what would work? What’s the best online solution for each of these five types of “know-who” search? Here’s my guess:

1. and 3. For finding local suppliers, we need Local Consumer Networks: This is what craigslist and its imitators (and the classified ads and yellow pages) are trying to do. Right now what all these tools lack is trust. What we need are tools that enable us to assess both the quality of local suppliers and the reputation and credibility of recommenders in these networks. Kinda like a ‘local’ of the Consumers’ Union. There are some betas out there trying to offer that, but they’re hopelessly underpopulated. Who knows, these could even evolve into reciprocal Gift Economy and Peer Production networks. And as for finding customers, in this brave new world, you don’t find customers. You make it easy for them to find you. Marketing and advertising are just for the big guys, and their influence and advantage are fading. The key is to do good work, and then get your customers to recommend you, both person-to-person and through these Local Consumer Networks. It’s all about reputation.

2. and 4. For finding business or life partners, we need Open Space Invitations: The Internet gives us access to millions of potential partners, and Open Space gives us a mechanism for inviting people to meet and explore partnership opportunities. All we need is a platform for offering these invitations, and a lot of help learning to craft them properly.

5. Finding virtual suppliers is different from finding local ones. For a start, there are a lot more to choose from. As oil prices rise, it will make more sense to find local suppliers of physical goods. But as technology continues to improve, it will make more sense to use virtual suppliers for anything that can be supplied as bits instead of atoms. We have seen something of a revolution in the supply of goods that can be supplied as bits (information and entertainment especially). But when it comes to service (as anyone who has struggled with a ‘help desk’ will confirm), virtual is still synonymous with awkward and suboptimal. To make virtual service comfortable and effective, we need (tired of me saying this yet?) Simple Virtual Presence.

Five different types of “know-who” needs, three different solutions. And none of them, alas, is provided by existing or imminent social networking tools and methods. But we’ll get there. Unlike the prevailing oligopoly markets, the new Internet-driven markets are truly responsive to need. I predict that within a year we’ll have powerful models for all three solutions. And getting “know-who” will become awhole lot easier.