VisualDNA Shops Launched

Wooohoooooo!! After 2 months hard work we are finally live with VisualDNA Shops!!

You can go to http://shops.visualdna.com and check it out or just watch the Stephen Fry video here:





As far as we know, it's the first time Stephen Fry has endorsed a web product. It's really exciting to have such a prolific blogger, twitter and all round super chap as part of the team.

My Little Brother "Mike" Wants To Be An Internet Entrepreneur

Imagini at the Social Media Cafe

We are sponsoring the Social Media Cafe this morning. Please do come along from about 11am if you would like to support us.

Revolutions Around the Sun?

My favourite type of interview question at Google was the "Analytical" question the goal of which is to test a candidate Product Manager's ability to carefully analyse a problem and come up with an answer or solution. One afternoon I found myself pondering the following question:

"How many times does the Earth rotate on its axis in
one complete revolution around the sun?"

And after thinking about it for about an hour (it's a hard question) concluded that I had the right answer and thought that it would make a great analytical question ... it is a question that anyone can understand, and puts almost everyone on the same level playing field as they won't have heard or thought about it before.

So what is the answer?

9 out of 10 candidates will quicly say the answer is 365

This is not the right answer.

Half of these then quickly adjust their answer to include a leap year and suggest that the answer is 365.25

This is not the right answer either.

(actually there is no leap year every 100 years and so the answer would be 365.24, but that this is not the right answer either).

It does not actually matter that they got the question wrong originally if they then manage to think through the problem and demonstrate their analytic ability, but only 1 out of 10 people realise that the obvious answer may not be the right answer and start thinking about the underlying assumptions before answering too quickly. This is the important skill that I look for.

The usual assumptions made are:

1. There are 365 days in a year
2. The Earth rotates once on its axis every day
   

And it turns out that one of these assumptions is incorrect. 1 out of 10 candidates work out which one it is (the second assumption is wrong).

For those that can't progress, I suggest they consider an imaginary Universe where the Earth does not rotate at all on its axis as it rotates around the Sun. In this case the Earth would rotate zero times yet someone standing on the surface of the Earth would still percieve one day. The conclusion is thus that the number of percieved days is not necessarily equal to the number of times the Earth rotates on its axis.

9 out of 10 candidates then conclude that the Earth rotates one less time than the number of percieved days and give the answer as 364.

This is still not the right answer.

I ask them to consider the imaginary Universe where the Earth rotates on its axis exactly once as it rotates around the Sun. The insight here is that the number of days perceived is either zero or two depending on whether or not the Earth rotates on its own axis in the same direction as it rotates around the Sun.

It turns out that the directions of rotation are the same and thus the number of days perceived is one less than the number of times the Earth rotates on it's own axis and the final answer is in fact 366.24

The question here is really about candidates thinking through the problem carefully, stating and validatings their assumptions and then not jumping to incorrect conclusions. Some people think that this constitutes a "trick question". In some ways it is, but it does not matter if someone does not get the right answer immediately and the question has enough meat and subtley to get a very good feel for a candidate's ability to analyse problems.

In many real situations the obvious solution is not the right solution. For example, the obvious solution to increasing the monetization of Google would be to put a big, big advert on the Google home page. It would indeed create a lot of value for Google .... in the short term. Unfortunately in the long term users would find the experience poor and stop using Google. With the benefit of hindsight, the correctness of this decision is fairly obvious, but pre 2000 there was great pressure on the Google founders to monetize the Google home page. Many, if not all, of Google's competitors went down this route and the rest, as they say, is history.

How to be an awesome Google Product Manager

I remember back in 2005, shortly before my first interview for a Product Manager role at Google (the first of 14 interviews!), searching online and finding this brilliant article by Ken Norton titled How to Hire A Product Manager. This was very fortunate, as at the time I little idea what a product manager actually did and even less idea what questions I was likely to be asked at interview!

After four years as a product manager at Google and having interviewed over 400 candidates for Google positions, mostly Product Managers, I feel that I know a little more ... and I thought I would share what I know.

Ken's article is actually pretty much on the money, but in retrospect, he describes what I now consider to be the hygiene factors for a great product manager. Namely they need to be:

1. Super smart
2. Technically strong
3. Have a "spidey sense" for great product
4. Able to earn leadership
5. Able to channel multiple points of view

The product manager has responsibility without authority and has to be able to lead and a cross-functional team to deliver complex and innovative products (it is an amazing and fulfilling role)!

However, having these skills, though necessary for success, is not sufficient. Sure, with the above skill set you will do a competent job and deliver great product, and there are many PM roles in the company where this is all that is needed, but you won't become a superstar.

To become a superstar you need one more magic ingredient ... and that is discipline. Discipline is the skill of working hard at all those things that are dull and boring and take time, but that need to be done well in order for a project to succeed.

Discipline is about taking the time to communicate so that everyone knows what is going on and people can fix problems before they get out of control.

Discipline is about writing Product Requirement Documents so that everyone knows exactly what is being built.

Discipline is about holding regular meetings to keep projects moving forward at a swift pace and sharing well written minutes.

Discipline is about replying quickly to emails and not putting off important but otherwise dull tasks that need doing.

Discipline is about creating plans and clearly communicating strategy and vision.

Discipline sounds easy, but is actually very hard.

For one thing, being disciplined is pretty dull. For another, being disciplined takes time. As a product manager you always have far more things to do than you have time to do them in and it is tempting to put off the less enjoyable tasks. Poor time management skills will make it impossible for you to be disciplined, so if you find it hard to be disciplined, start by looking at how you manage your time.

So if you think that you have all the skills of a great product manager, but for some reason are not succeeding as well as you think you should be, ask yourself if you can be more disciplined.

(In my next post I will explore the difference between "hot" and "cold" product managers).

So long and thanks for all the fish!

Tomorrow is my last day in the office at Google after four amazing years and I will soon be starting my new job as MD at Imagini. As a result I have decided to delete all my old blog posts and start this blog afresh with news, thoughts and opinions relating to my new venture. Feel free to subscribe to this blog if you would like to be kept updated on progress!

Note that his blog will soon be moving to the smug.ly domain (as soon as I can work out how I can change the CNAME on the libyan spider network to point to ghs.google.com).