But in order to make that an interesting hypothesis, they didn't just say,
that will be helpful, they said that we'll actually get more value by tapping into
this internal network, than if we actually go to some sort of external network,
beyond the boundaries of Roche.
On the basis that those employees should be
hypothetically more keen to help their colleagues and they would
understand the context of Roche's specific problems better than others.
So they designed this experiment whereby they identified the set of
scientific problems and
they asked their internal network to send out emails to all people working in R&D.
Follow it up with some, some, some, some reminders and
then they also posted that very same question same questions on Innocentive
which is this external technology platform that we've spoken about earlier in
the course where people post problems and the external network.
The scientific community have the opportunity to solve those problems.
Now, let's be very clear.
In a sense of get a feel for this, I mean, they have to pay money to an incentive for
the right to post this question and
to the extent that a scientist comes up with an answer on Innocentive,
they will also receive some sort of fee in return for that work.
Nonetheless, by setting it up very expressively with the same problems in
these two communities it was a proper experiment.
What was the outcome?
Well the experiment worked out, but
not in the way that they perhaps originally expected.
It turned out that posting the question on Innocentive gave them significantly more
birth in terms of quality and quantity, responses, than by posting it internally.
And in fact, I think that Roche is no different than most other large
companies in that respect.
It's very easy for people, employees in other parts of the company, to basically
just do their own thing and ignore requests from other parts of the company.
The point though is that by designing it as an experiment they learned
something really useful.
They learned something very valuable about how to ask specific questions and
to get replies to that.
And I'm going to come back to that actually in the next segment.
The second experiment I want to brief acknowledge again in Roche was very,
very different.
It was a group of people who said look we're spending all this money
on expense claims processing.
You know people do a flight, people stay in a hotel, and first of all we've
got to get permission to fly and you've got to make sure that you, you know,
flying the appropriate class or whatever it is.
And then you've got to get the expense claim signed off by your boss or
your boss's boss or whatever it is.
And Roche is a company that spent hundreds of millions of dollars a year
on these sort of travel and hotel expenses.
So, they said to themselves, this is a crazy system.
Let us actually trust in people's desire to do the right thing.
Let's, in the language of this course,
let's move from a bureaucratic process where the expense claims are checked off
by supervisors to an emergent process whereby we use essentially transparency,
visibility as the mechanism for ensuring people spend their money wisely.
And so they simply set up a, a,
a portal on the website whereby people would post what expense money they spent.
When I went on a flight, I put the money there, and the, the, the evidence for
spending the money there and
people simply looked at it to verify that I was spending money wisely.
Julian BirkinshawProfessor of Strategy & Entrepreneurship
