20 comments

[ 2.8 ms ] story [ 57.8 ms ] thread
I wonder what drives this, is it just poor decision making ability.
Or it could be people with unusual needs or lifestyles.

Someone observing a kosher diet will buy different food from the mainstream. Someone with limited use of their hands will buy kitchen appliances that most people won't need.

Incidentally, many "As Seen On TV" products were created for people with disabilities. They get marketed to mainstream audiences when the manufacturer realises that there's not a large enough market of disabled people to sustain the product.

Probably the same thing that makes people more likely to root for underdogs.

Avoiding the mainstream, etc.

These people arent as broken as this article may make it seem.

Its interesting how these zip codes are housing bubble resistant for example! Id wager these people arent trying to get rich, just trying to live a decent life and be nice to each other.

The ultimate crime
Not to mention they're telling us that bird of a feather like to flock together, and giving it a name such as “harbingers of failure” ...

Modern Marketing came out of WWII Psychwarfare efforts, gotta keep that in mind.

It's not the people who fail, but the products.

I suspect it's some kind of anti-aura effect; those look like the groups that are less likely to be "influencers". The Apple/Tesla strategy also used by lots of fashion products is to sell at high prices to market segments that the rest of the public identify as "cool".

(comment deleted)
Coefficients of demographic characteristics of "Harbringers of failure" zip codes:

Age -.2971

High school -.1710

Home value .2865

Bachelor’s degree .1197

Income .0646

White -.3919

Single family .4047

African American .2232

Multifamily .4132

Asian .3426

Distance -.3740

Hispanic .4854

Comp. distance -.4156

Coupon discount -.2333

Nbr. households .3160

Coupon frequency -.3741

Urban .4190

Unit price paid -.0523

Urban clusters -.1603

Meaning of these variables is here: https://docs.google.com/document/d/1d8O9ja_p08w6cSkOPZ-L1sDZ...

Not only do your numbers not appear in the document you provide, one has to wade through pages and reams of stats ... And your comment still makes no sense to me.
Effectively what these mean is that, for example, if your product is popular in places where a lot of people are young white people, in a single family household, in the countryside, without a university degree it is not likely to be successful.

If your product is popular in places with a lot of older hispanic people, with a university degree, who live in the city, in an apartment building, it's more likely to be successful

Can't edit it now, but noticed the negative sign was missing off one of them, "single family" should be a negative value.

From the paper:

  These values effectively state the correlation between the odds 
  of the product succeeding, and people who buy the product having 
  that characteristic.

  Age                    Average age of head of household 
  Home Value                    Estimated home value 
  Income                    Est. household income in the zip code
  Single Family                    % of households in single family homes
  Multi-family                    % of households in multi-family homes 
  Distance                    Distance to nearest MassStore store 
  Comp. Distance                    Distance to nearest competitors’ store
  Nbr Households                    Number of households in the zip code
  % Urbananized                    % households classified as “Urbanized” 
  % Urban Clusters                    % households in “Urban Clusters”
  High School                    % of households whose highest educational attainment is graduating high sch
  Bachelors                     % of households whose highest educational attainment is a bachelor degree
  White                    % of households that are identified as “White”
  African American        % of households that are identified as “African American”
  Asian                    % of households that are identified as “Asian”
  Hispanic                    % of households that are identified as “Hispanic or Latino (of any race)”
I feel like a software harbinger

Whenever I decide to use some software, it vanishes in insignificance. QBasic, Delphi, FreePascal/Lazarus, Micrografx Picture Publisher, Conitec 3D Game Studio, Java Applets, Mercurial hosted on SourceForge or Bitbucket, QCodeEdit, XQuery

That's funny. Sad, too. Maybe you are just different enough from others that you pick differently.
Anybody else feel like they are on both sides of this? Do these people REALLY only end up supporting the products that end up failing?

The first time I tried a 1st gen iPod, I was amazed and had to have it. Then bought a Zune HD when it came out, and kept it long into when everyone else had moved on.

Windows Phone, yup, I was a supporter. But was also doing yoga, WHM, intermittent fasting, etc. I was "barefoot" running before Born to Run came out, I rode a single speed mountain bike which never really became a thing.

I didn't read the entire report because I don't have an account, but are the people who are supposedly the "Harbingers of Failure" just more likely to be early adopters who are more likely trying new things and therefore they end up liking and supporting things that end up failing?

I started reading this article, then found myself skimming more and more as it became flakier and flakier. Now, because I just skimmed most of it, I may be missing some important points -- but my impression is that the authors are so bound up in the details & statistical tests that they can't see any big picture.

For example, in discussing household moves they talk about which zipcodes people choose to live in, as if they had perfect freedom of choice. But I don't see where they account for median income. How are we to know if people in a "harbinger" zipcode can even afford to move to a "Group 1" zipcode? Or if it would be a convenient distance from their employment?

I simply can't see putting much or any faith in these "results".