Case in point: for my latest campaign, I experimented with the “sponsored stories” option. You know what those are, Facebook shows you an ad saying “your friends X, Y and Z like this page”. Because of the credibility that you associate with your friends (usually), you are much more likely to check out the link than if it was a regular ad.

I started advertising a month ago in Seattle, in preparation for the conference coming up there. When I created the sponsored story, Facebook showed me my target market: a grand total of 920 people. Makes sense, most of the people who “like” the Startup Conference are in Silicon Valley.

Over a month, I have let those ads run by themselves. For some reason they are very cheap compared to regular ads. Over this month, most of the traffic came from those.

I just found the time to check out the stats for the ad and guess what: the ad now reaches 125,000 people. And that’s for Seattle alone! This is viral at its best (granted, it’s paid).

So all the buzz today is Facebook’s S1 filing for its IPO. Some people are concerned that Facebook has peaked, or that mobile will eat their lunch. I don’t see it that way. My personal experience, once more, shows me that it’s Google I’d be worried about. The high value marketing spending is leaving Google and moving to Facebook. Yes, if you are looking for “socks on sale”, Google remains your friend. But more and more, it’s Facebook that has the key to the customer’s mind.

10 startups will be selected to give a 1-minute elevator pitch in front of the whole audience. The best 3 pitches will then qualify to give a longer pitch to a panel of investors at the end of the day.

Even if you don’t win the competition, you will receive instant, valuable feedback from an expert entrepreneur. Join us! The pitch competition is consistenytly the most enjoyable and memorable part of the Startup Conference.

To apply, simply contact the organizer with a paragraph about your idea: pitches at thestartupconference (dot) com and a 3-slide powerpoint deck. Make sure to read this advice on what a great pitch looks like.

DEADLINE to apply: Feb 10.

Let’s be honest for a second and look objectively at the facts:

1) Yes, Samsung did copy Apple’s iPad design. From the power adaptor to the shape and the looks.

2) But no, Apple does not deserve patent protection for the iPad design. The patent system is broken.

This is your typical catch-22: if I argue the first point, then everyone online comes after me with the argument that patents are bad. If I argue the second point, Apple fanboys downvote me because I’m obviously refusing to admit the uncanny resemblance between the two products.

Oh, and what about patent reform? I stand by the column I wrote for VentureBeat two years ago: patents are useful, but at a minimum, we need to shorten their lifespan.

Do you know what a world without patent protection would look like? It would be filled with exact clones of the most successful products. There would be no Android. There would just be Apple’s iPad, and pixel-perfect copies from China. Having no patents would actually result in less innovation for the average consumer — but greater freedom for the geeks.

Hint: it’s not about having the best idea.

Case in point. What if Facebook pitched itself as a “social network where friends could see activity from other friends and share pictures. Oh, and users could also post pages they like.” Would you be interested?

What if they shortened the pitch to: “we are the #1 web destination in the world with 800 million monthly unique active users.”

Which version gets your attention? In the second version, they don’t even have to talk about features. All they talk about is social proof: our company is useful to people, and the proof is not because we think it should be, but because normal people find it useful. And there are lots of them. And we have proof.

In VC terms, the important part of the pitch is the execution, not the idea.

Of course, it’s unfair to compare your brand new startup with today’s Facebook. But understand that even if you were today’s Facebook, you would still lose a pitch competition if you only described your idea.

Let’s go back in time a little bit. This would be a more realistic pitch from Facebook, circa 2004: “we are a social network where friends could see activity from other friends. We are targeting college students and are selectively opening the service one campus at a time, to create an aura of exclusivity. Our pilot at Harvard was wildly successful, with 1000 people signing up the first day.”

What is good about this pitch? It shows you have a strategy to acquire customers. The strategy has already started working. You look like someone who has a plan, and you make your plan sound believable. You are clearly beyond the “idea stage”, you faced real customers and it looks like customers like your product.

That’s how you win.

But that’s not the right approach to understand how valuation works for your startup. It’s really not that complicated. Does this drawing makes more sense?

Before you get all ballistic and mathematical on me, allow me to explain in layman terms.

As you launch your startup, you are really shooting for a certain trajectory. How fast has your revenue been growing so far? Will the growth slow down, and when? You don’t need mathematical models, all you need is your intuition to tell you where your startup is likely to land.

Having this picture in your head is a good way to understand why, when your pitch deck says you intend to grow at 20% per year, VCs get much less excited than when your objective is to double in size every 6 months. If you aim low, you won’t land far.

And then there is the eventual flattening of your growth. All companies eventually slow down. That happens when they become #1 in their market, because then, they can only grow as fast the market, there is no more competitors to take market share from. That’s why the question about market size is such an important factor in investment decisions as well. When companies don’t grow much at all, their valuation is pretty easy to figure out, based on their actual profits, or as a fixed multiple of their revenue.

What about the dotted line? You can make up any projection you want, but your past history is the only thing you cannot change. Have you been growing like crazy so far? In Silicon Valley terms, it’s called having traction. It’s the #1 argument for a high valuation.

So really, it’s all there in this simple graph: your total market size, your targeted growth and your traction. Rather than tweaking equations you don’t understand, focus on making that graph more attractive to investors, and you’ll get a better valuation.

10 startups will be selected to give a 1-minute elevator pitch in front of the whole audience. The best 3 pitches will then qualify to give a longer pitch to a panel of investors at the end of the day.

Even if you don’t win the competition, you will receive instant, valuable feedback from an expert entrepreneur. Join us! The pitch competition is consistenytly the most enjoyable and memorable part of the Startup Conference.

To apply, simply contact the organizer with a paragraph about your idea: pitches at thestartupconference (dot) com.

DEADLINE to apply: Nov 1st.

In my 20 years of programming, I may have spent 1% of my time on complexity. Maybe 10% on optimizations. Granted, I don’t work at Google where scaling is a critical issue, so your mileage may vary. But.

I can think of a few good reasons:

Best-in-class libraries: it used to be that every programmer needed to know how to code bubble sort, insertion sort, quick sort… Now I just call list.sort() and I’m done. Even a beginner can do that, and their code will run faster than what I can usually code by hand. Remember how assembly code used to be faster than high-level languages and then compilers started beating humans? It works the same way for libraries. Have you looked at the source code for strcpy?

Optimizations often have nothing to do with complexity: sometimes you just have to iterate over all the structures. But you can offload the processing to a queue and make it asynchronous, or compute intermediate values. In my experience, actual complexity is way less important that a bag of optimization tricks. Thrashing your cache? Triple the speed of your code by running multiple steps on one element before processing the next neighbor.

There are also non-technical reasons why O(n) may not be that important. In a startup, finding product market-fit is more important than premature scaling. If one developer takes two months to code a feature and writes beautifully optimized code everytime, and another developer can do a quick and dirty job in half the time, guess which one will allow me to iterate faster?

Also guess which code is easier to maintain, the highly optimized one with an esoteric data structure, or the obvious (but slower) implementation. If you have any experience with software engineering, you know that code maintenance is often as important as everything else combined.

To all: I pledge that I will never again ask an O(n) question in an interview. It’s just not that important to my companies.

Instead, I want to find out if you have good coding habits, don’t get lost in complex data structures and can spot obvious typos in your code.

You can follow the discussion on Hacker News here.

Allow me a video game analogy: coding is like playing Tetris; verification is World of Warcraft. There is simply no comparison, it’s way more fun to verify than to code. Why?

I did become a design engineer for a while. I discovered that coding is a very mechanical process. Given a set of goals (an architecture document), you will eventually generate a working solution (the code). Yes, there is some excitement when you use a particularly clever structure or find an elegant solution. But overall, the glamour of being a designer wore off very quickly. Like Tetris: you get to solve complex puzzles, but everything stays within a clean little box.

Verification is an uncharted world. There, you are looking for human errors, and everyone knows that there is an infinite supply of those. You don’t know what could go wrong, all you know is that something probably went wrong, but there is no telling what or where. How do you go about finding it?

It’s this human dimension, the discovery and the uncertainty that make playing the verification game interesting and challenging. Just like World of Warcraft.

There is only one problem with preferring verification over design: management doesn’t agree. There is a very good reason for that: a bad designer will fail to produce a working design. That’s immediately noticeable and will prevent companies from shipping products. A bad tester? Looks just like a good one: “I’m done, I didn’t find any bugs.” Only time can tell, and by then it’s usually too late.

So professionally, verification is not rewarded as much. But it’s a lot of fun anyways, which is why I keep doing it.

And then there is Steve Jobs. He is the one counter-example I use each time someone asks me about out of control executive salaries. When most CEOs happily cruise over a 5% growth of their stock — at best — netting them millions in option and stock grants, when it’s not even clear that they have any control over the performance of the company they supposedly manage, there is this one exception: Apple.

The difference between Apple with and without Steve Jobs is so clear-cut, you can’t help but agree that one man made all the difference. And it’s not just about “shareholder value.” Steve Jobs’ Apple impacted millions of people’s lives, changing three industries forever (music, phones and computing).

How do you place a value on such a contribution?

Here’s one wild idea: the CEO stock options should be compared to a reference index built with its main competitors. So it wouldn’t be good enough to just cruise at 5% with the overall rising market, you’d actually have to outperform your competition to get any significant bonuses. Guess what! If you were a PC manufacturer using such a formula, your reference index would include Apple. And it would be obvious that you’re doing a lousy job in comparison.

Klout is a site that computes your influence, based on how popular and influential you are online. If your Klout is below 10, you are basically a low-life and not worth all the bits you are sending through the Internet pipes. If your score is closer to 90, your mailbox is probably inundated with requests from complete strangers asking you for all kinds of weird favors.

So what is wrong with this picture? Klout works great for the active Twitter crowd. That’s about 0.15% of the real world.

Really influential people — emphasis on really — have better things to do than worry about how many followers retweet their links. Actually, they usually don’t have time to tweet at all. Or don’t care.

And that’s why Klout’s idea is flawed: data from social media is a tiny fragment of what constitutes influence. Call it self-selection bias if you will (Klout only measures the influence of people who care about their Klout score).

It’s just not the world I live in.

So think twice before adding your Klout score to your resume. That, or we’ll all turn into Bob Garlick. I’d rather not.