Wednesday, May 14, 2008

ThinkTomorrow

I am in Half Moon Bay for the Think Tomorrow Private Capital and Venture Capital Summit. 

Think Equity, a boutique investment bank, organized the conference to introduce new comapnies and categories to the investment community.

Speakers included John Doerr, Chad Hurley, Mark Andreessen, Bill Campbell and others. 
 
I spoke on a panel titled, "Monetizing Social Media, Can It Be Done?," along with guys from Rockyou, Flixster, Gigya, and Lotame.  Surprisingly, the panel believes that it CAN be done:)

The day started with Bill Campbell, the chairman of INTU and an AAPL board member, interviewing Google's SVP of Product Marketing and Management, Jonathan Rosenberg.  The topic, innovation, is at the heart of all we do and the insights were valuable.

Jonathan started off by comparing and contrasting his training in product management with the current model employed by Google. 

Jonathan grew up in the MRD -> PRD -> tech spec -> waterfall gant chart method of product management that for many years defined software project management.  The documents passed serially from one department (marketing, product management, engineering, qa) to the next, the time frames were measured in months, and engineers were told what to build and when to have it ready.  The net result was often dissatisfaction across all constituents - product management failed to get what they thought they asked for, engineers felt like second class citizens/code monkeys, and the business suffered as a result.

At Google, Jonathan found several conditions held true, most of which all Internet companies can relate to.  First, the market moved extremely quickly and the product cycles needed to match the market's velocity. Second, the pace of change made requirements analysis challenging and a traditional product development process ripe for failure, and finally the needs of the market were tacit, not overt, and engineers were often best suited to understand the technology/user need base case and find the implementation and feature set that best met that tacit need. The business stack ranked problems and priorities and let engineers solve the problem - a respect and appreciation for the ability of engineers to find elegant answers to the business need permeates the company.

Given change is a constant, Google also focused on hiring generalists. Hiring for a discrete need often failed as the need would change to another priority the day after the specialist started. Finally, he found he needed to free himself from the baggage of his training and open his mind to new approaches - where management becomes a vehicle for encouraging collaboration of very bright people rather than prescribing discrete tasks.

Finally, he argued as the market changes winners and losers emerge (i.e. features) - Google feeds the winners and culls the losers quickly and works to let the users guide them as to what deserves funding and resource.

At Widgetbox, we recognize that the collaboration trumps serial and siloed product development processes. We recognize that engineers are often best suited to think through how best to implement a feature. Accordingly, we are working to be agile, collaborative rather than declarative, and have settled on Scrum as the convention for optimizing the challenge of fast moving markets and a need to organize effectively and efficiently to best respond to users and the market.

Can you be nimble without avoiding chaos? Can you have engineers play a key role in feature implementation without missing the market? 

The challenge is to find a convention, a framework, that provides common semantics, rhythms, deliverables, and a method by which the company can rank priorities.  We are experimenting with Scrum as our convention, as daily stand-ups, weekly releases, and bi-weekly sprints as our rhythm and deliverables, and are working to create institutional capacity for flexibility and collaboration across functions.

Each sprint team is composed of UE, product, eng, and qa resources and, as in traditional Scrum, sprint teams pull off the highest priority tasks from the product backlog. We are working hard to ensure we gang tackle business problems and let cross-functional teams develop the optimal implementation/solution.

Priorities. Rhythm. Data; three goals we are working to master. Rosenberg's talk proved quite inspirational and a lesson that new markets require agile methods of software development and that management is often tasked more with creating conditions for great people to develop great answers rather than in prescribing the answers to employees expected to merely implement them. 

No comments:

Post a Comment