Four short links: 14 October 2010

Google Price Index, The High Cost of Freemium, Literate Programming, Results Clustering

  1. Google Creates New Inflation Measure (The Guardian) — The Google Price Index will be based on the cost of goods sold online and could use real-time search data to forecast official figures. Clever use of unique data, but can the GPI findings be reproduced by another agency? I do like the idea of moving national statistical measures into real-time.
  2. How To Break The Trust of Your Customers In Just One Day — some horrifying revelations about how freemium worked for Chargify and their customers: Over the past year, we discovered that the customer that never paid had the highest support load. [...] Everyone’s always talking about freemium, but very few people actually use it, and we discovered this in looking at our customers for the past year. The reality was that less than 0.4% of customers had any sizeable number of free customers on their accounts. (via Hacker News)
  3. Annotated Backbone.js — very readable literate programming. (via Simon Willison)
  4. Carrot2 — open source results clustering engine.
tags: , , , , , , ,
  • Alex Tolley

    The US government uses hedonic pricing to adjust prices for quality, and substitution where possible.

    At best, Google’s data might capture spikes early, but it is unlikely to capture anything close to the government figures. Maybe it’s value will be in early detection of changes that will then be reflected in the financial markets.

    OTOH, there is suspicion that government figures do not reflect the inflation citizens actually experience, so possibly Google’s data might offer another take on inflation beyond the CPI and PPI data.