REPLY Society Strategy Development (SD7130)

[SDMAIL Jack Harich]

Posted by  Jack Harich <register at thwink.org>

SDMAIL Kim Warren wrote:
> Posted by  "Kim Warren" <Kim at strategydynamics.com>
>
> Thanks Jack - I know there are great processes for brainstorming with groups,
and were it possible both time-wise and physically to do it, they would be
great.

3 months is plenty of time. Even 2 weeks is plenty of time, at one week per
cycle. Businesses do this rapidly all the time via email, as well as physical
meetings.

Before the conference, I expect we at least have time for a convergent phase,
another ask/improvement phase, and a convergent phase.

> However, the challenges I mentioned before remain. I have asked 3 times now
for
inputs to the 'divergent' phase of your process - what would be good aims for SD
- with little response. 

Something still hasn't clicked. That's the whole point of iterative refinement.
You DON'T keep asking for more idea input. Instead, you switch to convergent
mode
and ask for judgment on the ideas you have. This improves them slightly and,
more
importantly, educates people on what it is we are looking for. Then after that,
you start another round of divergent input.

For example, right now I'm doing a business problem analysis. To build a model I
need certain cause and effect data. In round one I asked mid level managers at a
billion dollar corporation for their ideas on what is causing behavior X. I then
collected the data and we had a conference call. I explained the pros and cons
of
the data so far, and then asked for more. It came in, and one person had the
great idea to put it all into a spreadsheet. We then had another conference call
where everyone critiqued the data, such that they could see its shortcomings and
strengths. I then pointed out the areas where we needed more or better data,
modified the spreadsheet to accommodate it, and we had another round of data
collection and improvement. In this morning's email I sent out the latest
spreadsheet, asked for about a dozen blank cells to be filled in, and more
importantly, asked for about 20 cells to be improved.

This is iterative refinement. We are expecting over 20 rounds before we have
enough high quality cause and effect data to even begin modeling. Most rounds
involve no conference call. We're doing about two rounds a week, which is
impressive considering one person lives in Australia and I'm in the US.

This process would have failed if all I did was ask for more data again and
again.

The convergent phase of each cycle creates a feedback loop that drives the
process toward success. Without that phase there is no feedback loop, and people
shoot out ideas all over the place, instead of at specific targets.

Much more important than quantity of data is quality of data. If one is
collecting lots of data that cannot be used to solve a problem, then one is
doing
the wrong things right. Better is to allow the idea creators to give themselves
frequent feedback, so they are creating useful data, and the team is doing the
right things right.

Jack Homer's paper on "Why We Iterate?" shows the importance of iterative
refinement.  One never gets it right the first time.

Perhaps someone else on this list can give Kim another example?

> How effective would the sessions you experienced have been if most of the
audience had just sat in their chairs and said nothing when the facilitator
started each phase of the process?   

I've never had this happen, because I and the other managers or consultants I've
watched have always used the ask, refine, ask, refine cycle process. If instead
an ask, ask, ask process was used, I would expect input to fall off
exponentially
with each ask phase.

> The problem I raised before therefore remains - if we start your
divergent/convergent phases with a badly incomplete set of raw inputs, then we
will very, very likely converge on things most people don't buy into. 

Again, I can see something hasn't yet clicked. The whole idea of brainstorming
and iterative refinement is that you ALWAYS start off with a first round of weak
ideas. Then you improve them. This educates the group to what you are looking
for. Then you have another round of brainstorming, which from experience is
always better than the first round.

Your sentence above seems to assume there is only one brainstorming phase and
one
convergence phase. If that process is followed, then I agree. You will end up
with poor quality results.

Also, I would hesitate to call this "your divergent/convergent phases." This is
not my idea. It is a standard business best practice.

It may be that you are concerned that the goals you are about to release are
weak
and that that reflects on you as the facilitator. Don't worry. This is
groundless. This process expects them to be weak at first. A facilitator's role
is to provide the group with the assistance they need to go through several
rounds of successive refinement, until the goals are somewhat strong and ready
for the strategy committee, who will no doubt refine them even further.

I expect that in the first convergent phase lots of discussion will be on how to
objectively evaluate the different goal suggestions. Take a peek at this
approach
for ideas:
http://www.devicelink.com/mddi/archive/04/10/mddi0410p56d.jpg

So you might present the goals in a spreadsheet, in groups. But we probably need
to list the goals along the vertical axis, since we have so many. The criteria
would be along the horizontal axis with totals on the far right. That should be
a
tantalizing column.

Well, I hope this explains the concept.

Jack
Posted by  Jack Harich <register at thwink.org>
posting date  Thu, 03 Jul 2008 11:22:03 -0400