RoboCrush.com

It’s been an alarmingly fast past June and we blew right past Fathers’ day without a nod.

New models of aggregate trend analysis are unfolding from our initial beta testing and we’ve garnered a bit of underground media interest.

We are attempting to analyze the effects of various variable constraints when applied to model overlays across market sectors. So far the results are ambiguous and disambiguation doesn’t seem possible just through additional testing. It looks like we’ll need a radical refinement of how we approach the modeling aspects in this area.

What we initially thought were statistical outliers look as if they may not be. So we are reexamining our techniques for determining statistical outliers since obviously until we get that right it doesn’t make any sense to start any type of root cause analysis.

Some of you have asked if there are new competitors to RoboCrush.

I have a few things to say about that question.

1) Since, technically we’re not really in the marketplace yet, I guess we’re not in competition. LOL.

2) But seriously, there are companies that already have technology capable of doing components of what we’re doing and likely have taken notice of the new direction we are taking in the field of aggregate trending data  - which is actually a positive development as other companies have tried what we’re doing in the past but have not been taken seriously.

3) Competition also opens up the potential for strategic alliances - either with the competitors OR more likely with other companies with technology complimentary to the RoboCrush technology sets that we are developing.

We’ve been working on new modeling techniques in order to convert aggregate trend information into specific impact events that reflect changes in certain market segments.  While not purely a multilevel analysis approach, these new modeling techniques should be able to show different levels of trending within various sub-sections of each identified market segment as well as potentially identify the influencing factors.

Of course there are many variables involved and because of the nature of modeling aggregate trends the inherent complexities make untangling nondependent variables that are ‘wrongly coupled’ quite difficult: for example, dense clusters of “variable neighborhoods” may lead to potentially misleading identification of opportunities.

Perfecting these modeling techniques is likely to be a long uphill battle for RoboCrush, but a very important one.

Tad & PJ

Needless to say the economy is affecting a lot of businesses these days - and that includes RoboCrush.

We want to assure you that we’re in this for the long-haul and do not plan on throwing in the towel even during  these troubled financial times.

You may have seen the following news release:

“Secretary Henry M. Paulson, Jr. will discuss the economy and financial markets Monday at a meeting of the Wall Street Journal CEO Council. The meeting will bring together CEOs of major global companies for a two day meeting on the economy and the issues ahead.”

Of course there are a ton of factors to consider - not as many as in the RCFactor, though -

RCFactor table upates are complete and had very minimal impact on productivity, if any.

We’ll be looking at testing overlays of other process onto the RC Factor process to determine what other valuable information might be extracted from the current datasets we have. Part of that will be the various linking diagrams we discussed at the most recent “Future Progress Report” meeting.

A lot of what we are doing is still quite experimental - or I suppose the less “scary” way to say it would be “in it’s early stages”.

It was good to see how much people are grasping the fundamentals of the data collection and organization processes we are using to drive these future projects.