Modern Metrix

The US Economy according to Mint

I've been using Mint for over 4 months already, and I love it! Its "social" features (like comparing your expenses vs. other Mint users in your geographical area) are awesome. And I salivate just thinking about the amount of data that they collect!!! I agree that their sample is not representative, but if your research focuses on that top technologically-savvy strata of the population that spent a lot of time online, these data are perfect for insights on spendings, brand preferences, consumption of media and services, etc.

And today CEO of Mint shares some of these insights in the post on TechCrunch.

Challenges of Recommendation Systems

A great article on ReadWriteWeb about challenges that automated recommendation systems are facing today.

Collecting Data Online: Are We Missing Someone?

This is just another reminder of substantial qualitative differences between online and offline populations. These differences are often of psychographic nature, and therefore, can not be eliminated by simply rebalancing online samples on socio-demographic variables. As a result, online samples often does not include otherwise existent population segments of considerable sizes. Similar problem exists with telephone samples (due to increasing cell-phone only respondents.)

Some of these problems were covered at the New Technologies and Survey Research conference at Harvard University that I attended last year.

Subject Line Lenght and E-mail Performance

According to results of study conducted by Epsilon, shorter subject lines of promotional e-mail messages result in higher open and click rates. In particular, average correlation between the lenght of the subject line and response rates for e-mails from the consumer services companies was -.23.  Among the other factors that affect performance of the e-mail campaign are the order of words, word choice, and brand and audience awareness. The white paper (PDF) is available here.  

The Future of Measurement

Social Abacus makes predictions about the future of measurement:

Social Abacus - Future of Measurement - 2009

Publish at Scribd or explore others: Academic Work metrics Measurement

Thoughts on psychographic segmentation of media audiences

A great post from Prfkt about the future and potential of psychographic segmentations of media users. As media audiences continue to fragment, with regard to both content and delivery platform preferences, it is time for someone to come up with a media-agnostic typology of audiences based on such preferences.

Newspapers Readers Segmentation

This AdAge article reports about segmentation developed by Mediamark Research & Intelligence (MRI) that breaks down newspapers readers according to their content preferences. Nearly 12% of readers turn to the daily papers primarily to read about sports, and only 10% read newspaper from cover to cover!

I remember seeing this guy another day on the subway platform. While waiting for the train, he spent some time pulling out pages that he was not interested in, from his NYT, and placing them in the trash bin.

And that raises interesting question: could switching to the personalized newspapers (e.g., pages a la cart, or printing 12 versions of newspapers instead of one) save the newspaper industry? Consumers want to be in control of what they are reading. Why should they pay money for the pages they do not read at all? What if NYT publishes 12 versions of newspaper which are smaller, cheaper, and MORE RELEVANT to the reader?

SSN as a predictor in statistical models?

Did you know that telco and other industries often use social security number as a predictor (aparently, reliable and powerful) in statistical modeling? Aparently, your SSN includes such information as age and geographical location (state/city where you were born/originally registered.)

The Future of Advertising?

Computer games generated by computers ..

DIY TV ads from Google ..

What is next? TV, print, and online ads generated and distributed by machines without human involvement? Already Adsense ads placed on sites chosen by the algorithm perform better than those selected and placed by humans..

Consumer Segmentation and Personalization are among the Top Marketing Trends in 2009

Segmentation is one of the marketing trends/priorities in 2009:

"Customer satisfaction and customer retention remained the top two marketing concepts followed by marketing ROI, brand loyalty and segmentation, which represents a "Back to Core Principles" approach to marketing."

Add to that personalization and customization:

http://link.brightcove.com/services/link/bcpid7012076001/bctid7025748001