Learning from failure can be summarized as learning ‘what doesn’t work’.

Out of an arbitrarily large universe of possible things you could be trying (say 18,for the sake of this example), you have successfully accomplished narrowing down the scope of your next trial.

In practice, this can become problematic, because if you are looking to make an informed decision on what to try next, having only crossed one approach off your list (of the hundreds or many thousands of possible approaches to solving a problem), you are left only slightly more informed than where you started.

Learning from Success

Contrast this approach with a model of learning from success, where you have through a series of failed attempts hit upon something that is working and considered to be successful.

In this case, you well served to learn from this success and try to repeat it as much as possible, since success’s are fewer and far between.

Of course, most success come after a long string of successive failures, but in each failure there is some element of success (what actually worked) which can be far more informational than what didn’t work in terms of guiding your decision in what to try next.

Further reading:
Rewards vs Punishments by Jason Hreha of Dopamine.

Why are software development task estimations regularly off by a factor of 2-3?Edit

I have some ideas of what could be made out of them, but seeing as I have no time to develop any of them properly, I am opening them up to anyone who has a good idea to develop one of them. Have an idea and the skills for developing one of them? Get in touch.

TalkSend.com

ZipHire.com

ZonkBox.com

There are a number of reasons for this including the Technology Adoption Life-Cycle and Moore’s Law.

After a bit of tweaking, we got everything running smoothly on my brand new server, for half the price
I was paying for a comparable VPS at slicehost.

Can’t recommend these guys highly enough.

http://6sync.com/explore.

site:forrst.com/people ruby
site:dribbble.com css

In A/B testing, as is the case in the rest of economic life, there is no such thing as a free lunch.

In my experience, most split and multivariate tests are set up to maximize for a certain goal, e.g: registered users or email subs, but often fail to account for the un-intended consequences of such “optimizations”. The problem with changing your ui to try and maximize for one of these goals is that you are failing to measure the dis-utility that often gets created by cluttering or fragmenting the continuity of your interface or landing page.

For example, say I am trying to increase visits and awareness of a new feature of my product by adding a link to it on my homepage. Typically an experiment would be constructed by adding a new ui element and testing it’s copy/color etc that call attention to the new call-to-action that I am trying to optimize for. The problem with this kind of test is that by only measuring click-throughs to my new product page, I miss measuring the missed attention that was not given to another part of my interface.

Since the attention of the website visitor at any given moment is a finite resource, it is impossible to add something without detracting something at the same time. I have yet to see an a/b testing framework that accounts for this by default.

The end result is a bunch of “maximized” local optimums that don’t look or behave cohesively and often appear ugly, as I would imagine is likely the case with plenty of fish.

My answer on Quora to: I’ve noticed that when designs are refined through A/B testing, the res… http://qr.ae/MXiZ

In my experience the problem with the facebook advertising platform is that it doesn’t effectively allow for different types of advertisers to compete on an even playing field.

You could make the argument that since they allow for all different types of interest/demographic bidding, that there are many micro playing fields for different types of advertisers to compete on, but this analysis fails to capture the one most important vector that drives internet advertising, which is intent.

The beauty of Google’s adwords platform is that for any given keyword, a one-man custom boot making shop in Texas can profitably compete against internet retail giants such as Amazon because for certain keywords the intent of the searcher is more aligned with their product offering than Amazon’s.

In a generalized second-price auction such as Adwords, since advertisers are attempting to maximize their ROI on a per-keyword basis, they are not perversely incented to maximize their CTR at the expense of receiving lower CPC’s and an increase in less qualified traffic.

On the other hand, on the facebook advertising platform, different types of advertisers who will have wildly different costs for customer acquisition are competing for impressions and clicks head-to-head. This in turn causes two kinds of major distortions in the marketplace:

1) Advertisers with the highest customer acquisition value will eventually rise to the top, crowding out the smaller guys. e.g: you will see more ads for attorneys and real estate agents then you will for restaurants and dry cleaners.

2) Ad Copy / Images will trend towards becoming more and more attention grabbing in a never-ending arms race towards a higher CTR = lower CPC.

In the long run, neither of these two attributes is conducive to a robust and sustainable advertising platform for facebook.

My answer on Quora to: Is advertising via Facebook as effective as Google Ads? http://qr.ae/4GYA

Example: Compare a search on google for the term ’san francisco chinese restaurants’. Now try the same thing on yelp.

It goes without saying that the result provided by yelp is better on many levels. This is simply a matter of focus. Google clearly knows that it can’t possibly server a result to satisfy every different type of user query, so this leaves room for other sites to specialize their efforts around a particular niche. Since not every search problem is created equally, these sites have been able to focus on building domain-specific functionality and filters to solve the unique search problems in their domains. This poses an increasing threat to the core businesses of google et all as these sites gain in popularity and start becoming the place where people start their searches.

Fast forward to today (November, 2010), and for first time, google et all are starting to get very serious about neutralizing this threat of the vertical search engines. One vertical at a time they are integrating more and more domain specific search results and functionality into their search result pages. Starting with the most high value verticals (in terms of ad dollars spent): Travel, Mortgages, Local Services, they are finally starting to show the will to cut the middle man (the vertical search engine) out of the loop by integrating vertical search functionality into their search result pages.

Travel

Refinance

The “Cram-down”

Historically, the main driver of traffic to these sites has been natural search (mainly from google). As google et all begin to move in on their territory by providing extra listings, at the very top of each search result no less, scraping their content and integrating it into search results, the vertical search sites face a challenge to either build a strong enough brand, so that consumers cut out google entirely for their searches and go direct, or face loosing large amounts of their traffic due to this cram-down effect.