How can we ever hope to produce
effective Expertise Finders when we can't even get people in our own
organizations to keep their personal information up to date? That's a
question many professional services organizations ask constantly -- the
simple internal process of putting together a business proposal,
solving a problem or assembling a project team is often, nightmarishly,
- inefficient (takes too long),
- ineffective (often doesn't identify the best people for the project/problem),
- unduly subjective (people pick people they like to work with over people who are better suited),
- arbitrary (people may be selected because they are
under-assigned or located physically close to the customer, even if
they are really inappropriate choices for the task), and
- unreliable (not only is the information on which the
selection is made usually outdated and incomplete, it's often
inaccurate, self-aggrandizing and unverified).
How even more hopeless, then, is the dream of developing an Expertise
Finder that will find the best experts in the context of a particular
project need outside the organization, where the data is even less structured, the content even less complete and less verifiable, and the internal tools don't work.
A decade ago I read a prediction that, by today, the Internet would
have spontaneously (by a self-managed process) developed a database of
every consultant in the world and a verification system to go along
with it, so the big consultancies would all collapse, and customers
would essentially pick their own consultant teams person-by-person, not
limiting themselves to the employees of any one consulting
organization. This hasn't really happened, because normally the
customer picks only a project leader,
a consultant (usually in a big consultancy but sometimes an internal
person or even an outsourcer) who they then trust to assemble the rest
of the project team. If the work's done well, the consultant will be
rewarded for his/her choices, but no one really second-guesses those
choices or the deeply flawed, sub-optimal way he/she makes them. We use
similar processes to assemble project teams of other types of experts:
We pick our GP but rely on him/her to refer us to specialists, and we
pick a general contractor and usually rely on him/her to pick the
subcontractors, for example. The process is fraught with the same
suboptimization described in the bullets above.
The traditional IT approach to building such a database doesn't work.
It entails designing a form, a template of all the data elements about
each expert that might possibly apply, and then forcing people to fill
in and keep up-to-date all the relevant fields. That's essentially how
most social software works, too, and it's proven terribly
unsatisfactory.
Last year I envisioned an Expertise Finder that would work by crawling
people's blog content, penetrating corporate firewalls to find the best
people in the world who had the desired expertise and creating a 'map'
showing the most direct network path to those people (see sketch above)
and how much their expertise costs. I expected that the technology
gurus and Googles of the world would be able to build such a 'search'
tool quite easily, and the real challenge would be getting the content,
getting people to 'buy in' and post information about their expertise,
and getting corporations to allow outside customers access to this
information from their internal systems (or put a mirror copy on the
public Internet). But so far all we have are Ryze and LinkedIn and
eCademy (with its well-intentioned 'b2b Marketplace' and Google's
Orkut, and they don't work that way at all -- they take the traditional
'form-filling' approach, and are better suited to finding work
colleagues (or dates!) than either suppliers or experts.
The groups hoping and trying to develop such tools are sanguine of these challenges. Designers appreciate that information needs to be captured in (or converted to) a format useful to the expertise-seeker,
which is not necessarily the same format in which the expert normally
posts, or finds easiest to post, his or her expertise. And everyone appreciates that trustworthiness of the content and the tool are paramount.
.
What do you think?
- About the expertise finder design process:
If someone were to just put up a large empty 'space' and encourage a
large-enough group of experts and expertise-seekers to work together,
in time would the right solution evolve organically? Or would this just
produce a lowest common denominator solution that would satisfy no one
and not be used?
- About verifying expertise: Is the pathway to the expert,
the n degrees of separation between the expertise-seeker and an
identified expert important, so that the expertise-seeker can 'qualify'
the expert through the intermediary contacts he/she trusts? How else
can the degree of expertise of an individual in a particular subject be
intelligently and objectively verified, short of wading through long
recommendation letters?
- About making the system trustworthy:
Can we ever hope to supplant the tedious but effective process of
picking up the phone and asking someone you trust "Who do you know
who's an expert in X"? Can a computerized system be designed to mimic this person-to-person process?
- About building in expertise selection trade-offs: How do you factor in the availability and cost of experts along with the congruence between their expertise and what the expertise-seeker needs?
- About the role of blogs and other documented expert knowledge:
In what situations does it make sense to show expertise-seekers samples
of the work done by experts, both to qualify them and (in some cases)
to obviate or reduce the need to talk to them directly? Can you foresee
people ever paying money for documented knowledge without actually
conversing with the expert directly?
I continue to believe that there is a tremendous need for a
high-quality expertise finder, a new and very different type of search
tool from the tools that merely search data. And I believe that both
the technical and cultural challenges can be solved. But I no longer
believe that the development of expertise finders is inevitable, nor
that they can be developed in the 'laboratory'. They're going to need,
I think, a lot of bright minds asking a lot of 'what if' questions,
working together iteratively and allowing the design to evolve. And
they're going to need a lot more out-of-the-box thinking, radically
innovative thinking, if they hope to meet users' needs and
expectations. But the payback for success could be enormous.
|
MADE IN CANADA
trust your
instincts 
