Tuesday, June 23, 2015

LS1: John Vriezen

Intro: The Lean Startup is a book about 21st century management.  How does building a FRC robot and running a FRC team run into the same challenges Eric Ries identified in modern businesses?  How is our FRC team a group of entrepreneurs?

Answer: Modern businesses have to be able to thrive in fast changing envirnments, due to the pace of technology.  FRC robot building is similar in that there are only six weeks to design and build.   During that time, new insights about the game, game strategy, and technology available can be learned as you go-- this can come from team members, or from other resources such as Robot In Three Days, Chief Delphi discussions, and videos from other teams.  In additional, prior year robots (even years before your team existed) can help inform your design decisions.  For example, the ball used in 2014 was very similar to that used in 2008 and in both cases, throwing the ball over an overhead structure earned points.

In areas other than bot building, FRC can be a little more 'stable' in that some things can be more directly repeated with less 'startup' design.   Summer camps can be run similar to how they've been run in the past, but of course improvements can always be made.   Prior sponsors can be solicited again, but there are always opportunities for new sponsors and grant opportunities.  As the team's accomplishment accumulate, the story you tell these donors will need to evolve.

Chapter 1: What is productivity?  When building a FRC robot, what specifically is productivity?
Based on this definition, was our team productive during the last build season?

Answer: Productivity is not just a measure of how much work is being done.   The important measurement is how much useful work is being done.   We could organize all our tools and parts five times (in five different ways) and we'll have done a lot of work, but not be very productive.   For building a robot, being productive means different things at different stages of the build season.  Early on, being productive means quickly assessing and answering the big question items that will have significant influence on later activities.  For example, which robot capabilities are the most valuable and which are we capable of accomplishing given our resources?   What type of drive train is the right choice? Is there a unique way our robot can provide a key capability to a top alliance captain-- in 2015, this was the ability to grab RCs from the step during autonomous (at least this was the case for the very strong regionals.)

Was our team productive?   Definitely much more so than in prior years based on my limited knowledge of prior year efforts.   The decision to build a practice robot meant we were able to achieve a very large amount of driver practice, which is huge.   The extensive CAD efforts paid off in that we were able to iterate on our design and assemble our near final design from the ground up in a day and a half.  Areas where more productivity could occur is avoiding the lost time trying to find parts and tools, due to lack of organization of these.

Chapter 3: What did IMVU assume to be true when they designed their product?  How did customers actually behave?  Was there a faster and cheaper way to learn the lesson they learned?

IMVU assumed customers would want to extend their current IM systems and friends in those systems to add the IMVU features and invite their existing friends to do the same.   Customers were hesitant to do so, but were very willing to use the new system to meet 'strangers' and form a new friend group. 

IMVU could have figured this out more easily by experimenting with different approaches and delivery methods and contexts

Chapter 3: What is something that we were unsure of last build season that we experimentally validated?  Was there a faster way to learn what we learned?. 

We spent a lot of time refining the cam finger geometry so that they would lift both totes and RCs effectively.  Rather than doing this with McNeilus turnaround, we likely could have tried many variations using some hardwood and the shop's jigsaw and drill press and quickly arrived at a good shape. 

Chapter 4: Choose the Zappos, HP, Kodak, or Proctor & Gamble case study.  What assumptions did the Zappos founders make when they started their business?  How did they test their assumptions more efficiently than the IMVU team?

The HP case study was interesting, in light of the fact that my Company, Vision Solutions has been doing some similar things regarding charity and volunteerism.   The book discusses things HP might do, as it sounds like as of the writing, things were just getting started.    At Vision, they periodically have fund-raisers for natural disasters (which makes sense in a very real way, because our product provides protection for customers in the face of disasters like fire, flood, hurricane, earthquake, etc.)   More recently, they started the site charity double match for employee donations, and finally they have now requested that employees log volunteer time in a separate time tracking system (over and above the tracking we do for regular vacation and project time tracking.)   This, I think is not being well accepted-- more busy work so that the company can 'brag' about how much volunteer time is done by their employees, and 'recognize' the big contributors-- not sure what form this recognition will take, however.

Getting back to the question--- All the companies started with small efforts to assess feasibility and customer behaviors, so that they could direct their efforts most efficiently. 


So far this book is proving useful.   I've heard about various web services doing similar experiments where different user interfaces or price offerings are made to different users to gauge response behavior.  It does, in some respects seem unethical at times, where certain incentives or prices are offered 'randomly' to some site visitors and not to others. 

The basic premise does seem to be most applicable to cases where you have a reasonable large customer pool (or potential customer pool) so that you can have statistically significant results. And the product that is not 'mission critical' also helps.   I can't see it working at my company, where our customers spend very big bucks and bet their business on our software working whenever they call on it, otherwise, they are unable to run their business until the problem is resolved. 

With respect to CyBears, in addition to possible applications to building the bot, we could use some of these techniques for fundraising.   We could develop alternative ways of approaching donors (whether single persons, or companies) and measure effectiveness in terms of donations, volunteering, etc.) 

1 comment:

  1. I agree that companies started with small efforts so it would let them know if the public did like that product or not. But when they learn that it works it can be built faster witch is good and efficient.