Insights from Nebo

Scroll Down
August 4, 2009

TD Gammon—the most amazing computer you've never heard of.

The only way to improve is to focus on what you're doing wrong. This simple concept turned an unassuming computer program into the most dominant backgammon player in history. The computer was called TD Gammon, and it changed the way we think about the world.

Gerald Tesauro created TD Gammon in 1992 to test his theories on artificial intelligence. The goal was to make TD Gammon the best backgammon player in the world. In the beginning, TD Gammon picked each move randomly. It wasn't programmed with good moves, or bad moves.

Unlike Deep Blue, the famous chess playing computer who predicted millions of moves in advance, it didn't rely on the brute force of predictive logic, it simply measured its rate of failure. In fact, TD Gammon could only search 3 moves in advance during it's 15 second turn. TD Gammon could only improve by practicing its craft. Each move it made was recorded and the outcome logged. If it lost a game, it focused on identifying the moves that caused it to lose. TD Gammon analyzed only its short comings, the errors that it made, and how to resolve them.

Over time, TD Gammon improved to an astounding degree. Within a few years it was the world's best backgammon player. Kit Woolsley, a previous backgammon champion, said "There is no question in my mind that its positional judgment is far better than mine."

TD Gammon could care less about success. It had no ego. It didn't care if it made a brilliant move, or how many games it actually won. It only cared about the errors it made. It was programmed to do nothing, but continually improve. And that single-minded focus allowed it to dominate its field.

July 30, 2009

Experimental Media Intakes

Experimenting is a part of working in the online space. The media, search, and social landscapes are constantly in flux. Standards come and go, and the daily feed of information is closer to real time than it ever has been. To be any good at interactive marketing you have to experiment. You have to question the assumptions about the way search engine's function, the way information can be taken in, and the value of various marketing tactics.

One of the ways I choose to experiment is by periodically changing how I keep up with information online. There have always been a few staple blogs that I read to keep up with discussions about particular niches, but for the most part I keep my media diet in a changing state to keep up with the media's own rapid rate of fluctuation. Sometimes I'm subscribed to hundreds of blogs; sometimes I'm subscribed to five. Sometimes all I do is scan Twitter and other times all I do is Tumble. Most of the time I mix and match, and I think that's probably true for most people whose work relates to online media.

PSFK made an interesting comment in a post recently that they are trying to become The Economist of ideas -- not only with their blog, but with their Twitter feed. One reader said that they hardly read anything now besides the PSFK Twitter feed which links to more than just what gets posted on PSFK. I'm considering trying this out for awhile. I think it will be an interesting test not only of PSFK's trends coverage, but also of Twitter as an information source. I've never subscribed to a twitter feed via RSS before so I'm looking forward to seeing what it's like.

What do you think? What's your favorite way to keep up with information online? What platforms are your favorite and how do you experiment?

p.s. Google Reader is coming up strong with the like/share options.

July 27, 2009

Leading Users To Dead Ends

Besides books, I don't read anything on print anymore, and so I don't see print ads very often either. But, last Thursday while I waited for my car to be fixed at Walmart I took the opportunity to immerse myself in a fishing magazine and its advertisements. Print ads don't have to connect a reader to a website, but there were few ads that didn't try in some way. Most of the ads would append a url to the end of the body copy, and leave it at that. But, some of them tried a little harder. Intriguingly, the two ads that I think did the best at connecting readers to their websites had completely different strategies for converting users who actually got to the website: one built a dead end website, the other built a busy street.

July 22, 2009

There's a difference between hiring people and recruiting talent.

As an interactive agency, our biggest expense is our people. Hiring a new employee is a big investment, and potentially a risky one as well. This is made more difficult because the interactive industry is growing, and is highly competitive. There are lot of good companies vying for a limited number of exceptional people. Over time our approach to hiring has evolved based on a rather unlikely template. We've decided to approach hiring new employees more like college basketball programs approach recruiting talented players.

Why? College basketball is highly competitive. In fact, it's more competitive than most industries out there. Plus, everyone has to play by the same rules. This means that the biggest differentiator a team can have versus it's competition is the talent-level of their players.

So how do the great teams go about getting more talented players than their competition? College basketball teams don't post ads on websites and wait for the applications to flood in. They scout, identify and recruit talent that fits their needs. Great companies used to do the same thing with corporate recruiters. But, with the advent of tools like linkedin, it's easier than ever to proactively scout for talent. Companies can identify potential employees with nothing more than a few minutes searching.

But, changing from a hiring mindset to a recruiting mindset also shifts the way you approach the interview process. Each touchpoint with a potential hire is an interaction with your brand. It's the business equivalent of a high school athlete making a campus visit. If the applicant has to answer a series of generic pre-interview questions by email, is blown off by a distracted team member when they come for the in person interview, and has to sit uncomfortably in the lobby for an extended period of time; then the chances are the applicant is going to judge your company based on those interactions. And if those interactions are negative then you're at a disadvantage from the start. Talented people aren't really "hired" by companies, they "choose" to work there. If you give them a reason to say no, chances are they'll take it.

July 20, 2009

Business Principles Learned from Board Games


1. There are rules.

2. Changing the rules can make the game more interesting.

3. Context determines strategy. There is a time and place to build hotels on Baltic Avenue.

4. There is a point, when the game has stretched into the early hours of the morning, when winning is more about not giving up than it is about having the right strategy.

5. There are always surprises: jails to go to, fees to pay, and (if you're lucky) jackpots to collect.