http://correlate.googlelabs.com/
Month: May 2011
If we put as much effort into understanding and optimizing the factors in a strategy as we did into tactical optimization, we’d probably be four or five times more effective. Why? Tactical testing and analysis, when done well and right, can yield returns of approximately 2% to 20%. In certain industries, like airlines and travel arbitrage, the profit margin is made up entirely on tactical testing and optimization. Revenue management, logistics, and competing on analytics is simply the norm. Compete or die. So I’m not knocking tactical analytics. In most instances, however, companies do not have to compete using analytics. That is to say, most companies do not see analytics as a key factor in their strategies. Most companies certainly[…]
ExactTarget reported in their paper, “Subscribers, Fans, and Followers: The Social Break-Up”, Feb 1, 2011, that a top reason (44% of respondants) for unliking a Facebook Brand was “The Company posted too frequently”. Among other reasons: 43% said “My wall was becoming way too crowded with marketing posts and I needed to get rid of some of them”. 38% said “The content became repetitive or boring over time”. 19% said “The content wasn’t relevant to me from the start”, 17% “The company’s posts were too chit-chatty – not focused on real value”. All of these reasons cited go directly to the concept of relevancy. When does content become too much? When it ceases to be relevant. When do you want[…]
One application of social technology to a transportation problem caught my eye today: social flights. There are two civilian air travel economies – the public and the private. Private air travel, on private jets from quasi-public air fields, is a completely different experience. For one, there is no invasive security. Delays are fewer. Service is better. The drawback is the cost. Very few people can afford a private flight. Enter the what if. What if the coordination problems among a large group of people could be reconciled? What then? Why, you’d be able to take a private jet with a number of like minded people, at a coordinated time, without delays and harassment – and probably do it for less[…]
You’re invited to #WAWTO on June 22. Success would be several analysts, developers, marketers, data miners and creatives/artists coming out to engage and connect. Check out the invite below because it’s going to be a fascinating and fun meetup: http://www.meetup.com/WebAnalyticsWednesday/events/17841071/ The thesis question for the night is: “Does anyone give a shit about your dumb idea?“, first posed by David Cancel on his Data Driven Startup blog. And to help answer that, one of the best innovators, bootstrapper, and scientists I know, Joseph Carrabis, will be on hand to present solutions, pose problems, and answer your questions. Like any collective community, we’re long on ideas (good), long on realism (okay), and a bit shorter on testing (not as good). So[…]
In a game called “Power Grid Factory Manager”, two to five players are challenged with running a factory for five turns. The three sources of randomness (exogenous shocks) are the starting bid order, the increase in the price of energy, and the three starting factory equipments available upon the very first turn. These three sources of randomness are enough to produce all the variety required. No dice here. No reliance on luck. Managers are given the same starting conditions, and have two resources – workers and money. Everybody is paid on the same schedule, where the input costs are subject to random fluctuations and capital must be balanced. There are very large degrees of freedom involved. The person with the[…]
The Economist Style Guide is a good reference for communicating insights. It’s the easiest style guide for our industry. A few favourite highlights: “Short words Use them. They are often Anglo-Saxon rather than Latin in origin. They are easy to spell and easy to understand. Thus prefer about to approximately, after to following, let to permit, but to however, use to utilise, make to manufacture, plant to facility, take part to participate, set up to establish, enough to sufficient, show to demonstrate and so on. Underdeveloped countries are often better described as poor. Substantive often means real or big. “Short words are best and the old words when short are best of all.” (Winston Churchill)“ “Active, not passive Be direct.[…]
There’s been a some talk about education bubbles as of late. Some of the evidence is compelling. A Frontline documentary on the quality and liability of for-profit post secondary education for instance. A general lamentation of the volume and cost of MBA’s being produced. And, many gripes about the affordability / accessibility of post secondary education in the US and Canada. When demand exceeds supply, the price for that supply rises. There’s a period of lag between demand shocks and supply responses. Frequently, suppliers either overestimate demand shocks, or, more likely, increases in supply follow a step function, which in turn eventually causes a price collapse. It’s this dance between supply and demand that generates the pseudo-random price of a[…]
For thirty minutes Monday night, when a federal statute prevented me from accessing information I wanted, I raged. Then I had empathy. There was an election in Canada. The federal statute is an Elections Canada law that prohibits anybody from transmitting results to regions of the country where the polls are still open. That includes all broadcasters and applies to online. For 30 minutes, between 9:30pm EST, when the polls closed in Ontario, and 10:00pm EST, when the polls finally closed in BC, I relied solely on twitter and a dashboard on CBC TV. It was horrible. That screencap shows a dashboard, populated with data from later on in the evening. You could see the seat totals for parliament, and[…]
eMetrics Toronto 2011 was excellent. A huge thank you and congratulations to Andrea Hadley, Jim Sterne, my colleagues on the advisory committee, Syncapse for putting on a great Web Analytics Wednesday, and the Rising Media team for putting it on and pulling it off. Many points of view were generated and shared. The top five that generate the very most interesting questions (for me) are: 5. Laura Callow delivered a convincing arguement in favor of testing regardless of much traffic a given site or page is doing. Traffic will eventually accumulate for significance. She also presented the relationship (dare I say correlation?) between page load time and conversion on site, and, in a delightful twist, linked it page tagging and[…]