Monday, January 24, 2011

Top-down vs. Bottom-up product development

We were recently discussing a new feature for my new startup TheIceBreak that involves building functionality that has never been done before. In order to design it, we relied on a few assumptions that are difficult to validate without actually building the product.

This reminded me of the fundamental differences between dogfood-style (bottom-up) vs vision driven (top-down) product development, and the companies that follow them: Google vs Apple, and how this approach defines how products are built.

Google has a strong bottom-up culture and believes in being extensively data driven. All the products that are built at google go through extensive number-crunching and analysis before (except the 20% projects). It is very difficult for someone to justify a brand new product as there might not exist enough existing data to validate it. Also, someone who is a small contributor might not see the 'big picture' and help move the company in that direction.

Apple (or Facebook), on the other hand, are driven by vision. There is, of course, a lot of data analysis that happens to get to the vision, but they repeatedly build new products which create a brand new market that never existed before. They have changed the company focus multiple times in a major way that it affects more than 50% of their revenue or users. It usually involves the high level teams define a clear product vision for the company, and everyone works towards executing on that path.

Creating something that is truly groundbreaking is extremely difficult to validate using existing data, so it relies on someone who has a clear foresight on what is going to be useful. It is very difficult to create something using iterative, data driven techniques that will change people's behavior in a major way. It is, however, a great way to do incremental improvements to an existing product and get big results and can work quite well until someone 'changes the game'. Top-down, vision driven strategy can refute the existing mindset to create something truly revolutionary, but it relies on a 'leader' to be able to analyze the data they have and define this clear 'vision'.

Having a clear overall vision for a company also helps the teams to know what's good and what's bad, and they have a clear path that they can execute on and be highly successful. This vision has to be broad enough to cover global trends, but also sharp enough so it can actually be followed, and it is absolutely the most critical thing for the long term success of a company.

One can also argue that the difference is similar to a democracy vs dictatorship. On paper, under the ideal conditions, dictatorship based governance can be more efficient. However, its more prone to 'rouge dictators' which leads us to the belief that democracy is better in the long term.

In the end, getting the right vision is extremely difficult (and a lot of people actually find it boring to try writing one down), but might be the biggest factor in determining long-term success of a company!



Thursday, January 06, 2011

Presentation Zen

Just finished reading the book 'Presentation Zen' by Garr Reynolds.

Its a really good book if you want to do an actual presentation, when you are the speaker on a podium and everyone else is listening.

The concepts in this book revolve strongly around keeping the 'deck' as simple as possible with heavy use of images, etc. The author also suggests that instead of adding more details on the presentation, there should be a separate handout.

This approach makes sense for one-way presentations to a large group of people, and there are some really good points made.

I find that this approach is not very useful while presentating to a smaller group of people whome you are more involved with, like a product review with the team. Such presentations involve a lot of discussion, and its good to have all the data cleanly presented. There are lots of charts and tables (though, less bullet points), which does not blend well with the style in this book. People also expect to be able to just 'read' the deck and understand the key messages.

There needs to be a balance between keeping the deck at a high level vs adding details. It needs to have enough details that people can recall the presentation by looking at the deck, but not so much that it overwhelms them.

Monday, January 03, 2011

Pivoting + Startups

Pivot is one of the most overused terms in the startup world in 2010. The general thinking is that Pivoting is a good thing and founders are very proud to say that they have 'pivoted' X number of times for their current startup. I have seen funding pitches which even contain a slide on 'possible pivots', where the founders talk about how its easy for them to change their vision.

Having 'pivoted' my own startup (Pickv -> TheIceBreak), I feel this is definitely something that is important in a startup's lifecycle. However, it is not all that fun or glamorous, and there are very strong reasons to have a plan that does not involve pivoting as the strategy.

Pivoting usually involves changing the vision of the company. From what I have seen, there are 2 kinds of 'pivots':
  1. Pivot to a bigger opportunity: You have a generally decent product getting tiny amount of traction, but see a bigger opportunity so you switch over to that.
  2. Pivot to a new product: What you have done is not doing well, or has very little hope of getting traction, so you pivot to a new product that might or might not have anything to do with the original idea.
In both cases,
  • Pivoting is extremely painful: You have spent a ton of time and energy really optimizing the service for a particular use case. Now you are dumping that use case.
  • Pivoting doesn't really solve the problem: It shifts the problem somewhere else, or results in a different problem. 
  • Too many Pivots = lack of vision: If you have a clear long term vision, there might be changes required in the execution where you take 'diversions' to get to the final goal, but there is not usually a full 180 degree pivot. If there are too many pivots, the long term vision might get lost and the company might flounder and stutter without getting anywhere.
  • Pivoting means something has failed, and its time to try something new. This applies to the first kind of pivot that I described earlier. While its good to accept failure early and move on, there should be enough time and resources spent on each iteration so it can reach its potential.
In general, I am still a supporter for pivoting, but has to be done thoughtfully. There is absolutely nothing glamorous about doing so and it should generally be the last resort.