2014
-
Data Leakage – Part 2
Feb 15
When you click a link on a page, your browser takes that page’s URL and sends it along with you as the “Referer” header. The site you end up on can then track exactly which sites are sending traffic. If your site is private (i.e. an internal company blog) and you link to an external…
-
Server-side Analytics through Google
Jan 4
Some of the products I build for the web don’t come with a website. But I can still use Google Analytics to track their usage. Thanks to a handy PHP library I found on GitHub, I now have the luxury of tracking the execution of server-side scripts in the same analytics dashboard I use to…
2010
-
Environmental Modeling – Part 2
Jul 2
Pair Neale Martin’s habitual and executive minds with the mental model we each carry of the world, and product design gets simpler: don’t move the power button. Make your product feel different only where it’s better.
-
Environmental Modeling
Jun 25
We build mental models of the world from our senses, then use them to predict what happens next. A theory on how programs can do the same, and how I’ve started using it in JavaScript.
-
Traces
Apr 19
Following a sea otter’s trail along the beach at Fort Stevens got me thinking: every product, like every creature, leaves traces long after it passes. Even the iPad will shape buying behavior down the road.
2009
-
Bayesian Statistics – Conclusion
Dec 12
Most marketing decisions come down to numbers. Revisit yesterday’s case study without assuming everyone rated every product, and Bayesian averages flip the result: the product that looked weakest turns out to be the highest rated.
-
Bayesian Statistics – Part IV
Dec 11
Simple averages say drop Product C. Weighted averages say drop Product A. A Bayesian average says drop Product B. A case study in comparing product ratings across market segments, and why the method you choose changes the answer.
-
Bayesian Statistics – Part III
Dec 10
A simple average breaks down when the room holds both small children and basketball players. How a Bayesian average makes product ratings comparable no matter how many people rated them — no statistics course required.
-
Bayesian Statistics – Part II
Dec 9
The Monty Hall problem drove me crazy through high school and college. Then a colleague introduced me to Bayesian statistics, which accounts for what the host knows, and why you should always switch doors.
-
Bayesian Statistics – Part I
Dec 8
When one widget gets eighty ratings and another gets five, a simple average lies to you. Thomas Bayes, an 18th-century statistician, offers a way to weight those ratings fairly — the first in a short series on putting his method to work in marketing.