Category Archives: Uncategorized

Consumer Price Index

Somebody asked where he could  get Consumer Price Index data, in particular food and energy data. Curious, I made these two charts.  Same data; just different vertical axis: linear and log.

I don’t really have anything to say about these.  Just thought I’d share … well  … I guess we learn from this that most of the cost of “other goods and services” is passed thru health care costs.

Money Under the Mattress

Update: As pointed out in the comments I was confused by what that chart was showing… ignore the following.

When you put your money in the bank that bank promises to keep your money safe and then give it back later.  When I was boy they also gave you a bit of interest, but recently they scoff if ask about that.  In fact these days they usually manage to charge you some fees so that when you get your money back it’s actually less than you put in.  In this situation it starts to seem reasonable to just keep your money in tidy bundles at home.

When Henry Ford’s son took over the family business he enquired into the firms cash reserves: it was around $700 million, all of it in the vault down in the basement.  Henry didn’t like banks.

There are lots of places to park your money.  Some of them are more certain than others.  Right now you can currently park your money with the US treasury and they will promise to return it to you later, they will give you some interest – negative interest.  I find that hard to think about.  You are apparently better off storing your money in a vault in the basement.

Unique Organs in Cloud Computing

A long time back I had what I felt was a brilliant insight which ran as follows.  Cloud computers and their operating systems are a different and unique species.  The be unlike anything we have seen before.  This will discomfort observers who think they understand what is important in an operating system or a computer.  The niche this species lives in has opportunities that will cause rapid evolution of new things and drive a lot of old things to become vistigal.

Further I felt it was obvious that a cloud computing offering from vendor A would be very very different than the offering from vendor B.  Not for the usual reasons either.  The usual reasons is standardization is hard and it is work for the early movers dismiss for many reasons.  In this case the diversity would arise because vendors would have existing hard to reproduce assets which they could bundle into their offering.  Amazon, for example, could bundle their supply chain.  Google could bundle their cache of the web and the search tools upon that.

These predictions haven’t turned out to be as accurate as I expected them too.  The evolution has been fast, but not as fast and disruptive as I expected.  And the second prediction, e.g. that we would see interesting complex vendor specific APIs emerging in the cloud operating systems has  occurred  far less than I expected it to.

But this real time search API on Google’s App Engine looks interesting.

when in doubt shout … or not

The walled garden around most academic publishing is such a pain.  I get left with sound bites.  It is like those fear driven calls to action where if you ask for proof about the fearful thing your informed that the evidence is all top secret.

Here is a fun little example. This is the abstract for this paper by some marketing B school guys.

When in Doubt, Shout!
Paradoxical Influences of Doubt on Proselytizing
David Gal and
Derek D. Rucker
Author Affiliations:  Kellogg School of Management, Northwestern University …
Abstract

A seminal case study by Festinger found, paradoxically, that evidence that disconfirmed religious beliefs increased individuals’ tendency to proselytize to others. Although this finding is renowned, surprisingly, it has never been subjected to experimental scrutiny and is open to multiple interpretations. We examined a general form of the question first posed by Festinger, namely, how does shaken confidence influence advocacy? Across three experiments, people whose confidence in closely held beliefs was undermined engaged in more advocacy of their beliefs (as measured by both advocacy effort and intention to advocate) than did people whose confidence was not undermined. The effect was attenuated when individuals affirmed their beliefs, and was moderated by both importance of the belief and open-mindedness of a message recipient. These findings not only have implications for the results of Festinger’s seminal study, but also offer new insights into people’s motives for advocating their beliefs.

Fetinger’s work is a classic.  Members of a doomsday cult have a stark problem when the day of doom comes and goes.  What do they do then?  What Fetinger observed is that they redouble their efforts, they become even more committed.  And interestingly instead of disbanding they run out and recruit more members.  There are useful life lessons in that.

So what I’m curious.  What does this paper say?  Do doubtful people tend to shout?

If you can’t get a copy from inside the garden wall sometimes the info leaks out.  The worse case scenario is when the article’s blurb echos around and all contact is lost with the actual research.  This research  rattling around in the media etc., but this article is the only one I’ve found were the author actually appears to have read the research.

They conducted three experiments. In one, 151 people recruited from an online database were surveyed about their eating habits. Their confidence level was manipulated by asking them to describe either two situations in which they felt considerable uncertainty, or two situations in which they felt a great deal of certainty.

They were then asked whether they were a vegan, vegetarian or meat eater and instructed to indicate on a seven-point scale how important their choice of diet was to them.

Finally, the participants were instructed to imagine they were discussing their food choices with someone who followed a different type of diet, “and to write what they would say to convince that person of the advantages of that diet.”

The result: Participants induced to feel doubt wrote longer messages and spent more time writing them than participants induced to feel comfortable. This effect was particularly strong among those who viewed their dietary preference as very important to them; it disappeared altogether among those who considered diet unimportant.

In both this and a second experiment, “individuals induced to feel doubt about their beliefs exerted more effort toward advocating beliefs,” the researchers write. A third study, which looked at Mac and PC users, found the doubtful also “expressed a greater likelihood to attempt to persuade other people of their beliefs.”

So it appears that this research is not reproducing Fetinger’s result.  I’d expected an experiment were fans of product A would be presented with compelling evidence that product B is better.  It appears that all we are seeing here is a proof that if you provide a person with some reason to doubt his world view he will respond by devoting some effort to the maintenance of that world view.  That doesn’t seem dysfunctional.  In fact it seems entirely appropriate.

That is a shame because.  It would be really fun to have a study that showed that people engage in irrational over the top arguments when they become doubtful.  This research promised me shouting!  It doesn’t deliver.

Emeritus at Apache

Long time coming, but I am finally converting my membership in the Apache Software Foundation to Emeritus.

The last straw was the foundations inability to form an effective and muscular response to the frog boiling taking place in the Java community down stream from Oracle’s acquisition of Sun. Which is perverse, since I don’t have a dog in that fight. And while I have a half a dozen issues with how the foundation has evolved over the last decade none, including the above, are actually particularly serious.

So I am not stepping back as a matter of principle. Sure my principles and concerns enter into it, but even if you sum them up and multiple by a small integer they do not amount to a much.

And certainly they are nothing compared to how much I admire, and how proud I am of what we have built.

My interests have shifted.

Oh, and a heads up for Planet Apache blog readers; I will remove this blog from that aggregated feed shortly.

Customer Service Department

Today’s mail brings a two page letter of fine print. The letter was mailed from Wilmington Delaware and has a PO Box and zip code for it’s return address. This letter is obviously about a change in the terms and conditions of some credit card. But no place on the letter does it indicate what credit card. The bank it is goes entirely unmentioned. The letter has a header, it looks a bit like a email. It informs me that it’s from “Customer Service Department” and that it’s subject is “Account Information.”

No wait! It also has the an account number, humm. Yes it’s a credit card number, in fact the entire number. Mystery solved, it’s Bank of America. Don’t we love the Bank of America? Hm, I wonder if any of my neighbors got this letter?

Which reminds me, why am I unaware of a service I can subscribe to that will send me diffs of the changes to various terms and conditions. Since, I certainly can’t find what changed in this letter.

GroupOn – Pricing Games

I don’t find anything surprising in this report out of Rice University about the experience of businesses that tried GroupOn.

If you are unfamiliar with GroupOn, they are a hugely successful internet business.  A two sided network bridging between small local businesses and local consumers.  They have an opt-in email list, sorted by city, of consumers.  Once a day they mail out an offer to buy a coupon for a local businesses; typically 30$ worth of goods for $15.  They often sell a few hundred of these each day, and GroupON keeps, I’ve been told, 30%.  The business gets the money right away.  A large percentage of the coupons are never redeemed.  So it is a gamble for the business and the consumer.

The buisness may or may not get a long term customer out of this transaction.  The business may achieve higher utilization using this to reach cheap-skate customers.  Given that most businesses have very low marginal costs those customers can be quite profitable.  The deal might damage the reputation of the business.

Pretty far down the hearsay chain here, but the article says of the report says:

  • 66% of the deals reported to be profitable
  • 42% would do it again
  • 20% of those who were profitable would not repeat the deal
  • restaurant deals are less profitable than spa deals
  • customers using the deal are: poor tippers, “entitled”, not likely to repeat
  • employee satisfaction is a good predictor of profitability on these deals
  • other predictors of profit include:  overage on the bill and repeat customers.

This quote by the study’s author “… there is disillusionment with the extreme price sensitive nature and transactional orientation of these consumers among many study respondents.”  irritates but doesn’t surprise me.  Small businesses tend to be naive about how to play pricing games effectively.  So they don’t comprehend what is necessary to tap into the higher utilization rates these games enable.  If you are going to play those games then yes you are going to make contact with more one time customers who are cheap skates.  For many small businesses a good customer relationship is a part of how they are compensated.  Maintaining that when playing pricing games is exceptionally difficult.

Over time large firms have learned how to vacuum up a larger and larger share of the business that was once done locally.  They have been quite calculating about that.  One reason I enjoy being subscribed to GroupOn is that it provides me with a stream of examples of businesses that remain small and local.  I guess I hope that these businesses can learn how to play pricing games effectively.

First Time Right

I had not heard this term before: “First Time Right”

It describes that situation where switching costs are so extremely high you had better get it right the first time you choose.

I saw the term used in a deck describing a problem faced by modern device makers.  More and more devices are being made that absolutely must have connectivity to the cloud.  Examples of these devices include: power meters, security systems, eBook readers, navigation systems.  Quickly this is becoming everything.  Typically do this using the cellular networks.  The problem for the device maker is that once they pick the cellular vendor to use they can not switch.  Even if they pick a GSM vendor they can’t switch to another vendor unless they swap the SIM cards in all the devices, devices they probably do not have access to.

I like this term and it’s fun to enumerate things that feature a strong element of “first time right.”  Your native language, marriage, where you settle after college, where you site your factory, the programming language you become fluent in.  More often than not these are things with strong network effects.

keeping my confirmation bias happy

I must thank Andrew Gelman for drawing my attention to this flawed but yet interesting article about a “recent breakthru.”  No doubt my regular readers will appreciate that I have hinted that this is clearly the case a few times already.  But is is always nice to have one’s intuitions validated.  No doubt anyone familiar with power-law distributions will find the results unsurprising.