Pages

Saturday, December 22, 2018

Free Shipping: Over-Valued, but not Purely a Gimmick

NPR had a good story "There's Always A Cost Associated With Free Shipping". The story partly suggests that free shipping is ultimately of a gimmick or come-on--at the end of the day, there is no free lunch, total product cost is what matters. I agree, marketing theory says, find some product cost (or other attribute) that consumers over-value (or over-dislike), and play that up for an asymmetric benefit. "Free Shipping" fits that profile.

I have a few counter-arguments for this specific case, though. First, historically merchants have sometimes treated "shipping and handling" as a profit center. I.e., advertise a low price, and only once the consumer is deep into checkout, reveal an inflated S&H charge. It is form of bait-and-switch.

Second, even if merchants are playing fair, and treating shipping as only a pass-through cost, it can be hard for consumers to predict what the shipping cost will be. Resulting in added cognitive burden, and the risk of spending time shopping and going through checkout, only to get a big surprise at the end.

So free shipping takes both of the above issues off the table.

(There are significant environmental issues associated with free 2-day shipping, I thought this was a good summary. It would be nice if Amazon led the way by pushing the green angle on not taking free 2-day if you don't need it--they already offer some credits, which is a small start.)



Saturday, December 15, 2018

A Different Kind of Nationalism for the USA?

09/15/22 UPDATE: Rep Sean Casten has a kindred proposal

12/18/22 UPDATE: This seminar organized by Bruce Schneier is one of the few (only?) times I have come across a similar-in-spirit suggestion that maybe it no longer makes sense to ""organize our representative units by geography". 

*************
"Nationalism" is in the news lately, in the US and elsewhere. I view myself as an American nationalist, but in a much different sense than the term is usually used. I am a "nationalist" in that my allegiance to country is far, far ahead of allegiance to state or locality. So naturally I would like to see a Congress that is constituted to reflect national interests more, and local interests less. I propose a thought experiment in that regard.

What if half of Congressional representation were de-coupled from geography? The goal being that these congresspeople would represent the interests of the entire country, not a particular state or geographical district. How could that work, and would it be a good thing?

The concept that I propose is for a national congressional district to be defined by randomly assigning voters to said district. So instead of districts that average 710,000 people within a given state, assign 710,000 voters from anywhere in the US randomly to the district. If I am right, the national congressperson would be far more inclined to take a national view, rather than stubbornly advocating for local interests--even when those interests diverge from their core beliefs.

There are of course tons of details to work out. A critical one is--who is permitted to run to represent a given district? I think that is straightforward--a person can run for the district to which they have been randomly assigned. That is their district(!)

So that speaks to some of the how. Would it be a good thing? I think it could be. Our current political system is broken in so many ways. An underlying flaw that gets way too little discussion, is the fact that 50 states, of widely varying populations, is a broken and outdated system. First, there is the somewhat widely-acknowledged flaw of size--where voters in small states have vastly disproportionate voice. There is also the dysfunction of seniority--if a given states' congressperson happens to be senior, they will be in a position to cater to the interests of that state--completely contrary to the national interest. Finally, I am pretty sure that no other democracy has gone 100 years--let alone 250--without revising its administrative districts (our U.S. states). It is miracle that this kind of man-made boundary drawing worked okay as long as it did, but it is way past time for a re-do.

I suspect a major, hidden benefit is that the national Congressional reps would be the moral exemplars. Their views would strongly signal what is best for the country, not a parochial interest.

DISCLAIMER. This is a thought-experiment. I don't expect it to happen. I certainly haven't through through all the details. But clearly the status quo is not working. The two, related things I would hope to get people thinking about are: 1) We would be well-served to look for ways to put nation before locality; 2) Closely related, the default assumption that our 250-year-old legacy of states is the ideal or natural way of doing things should be vigorously challenged.


Sunday, November 25, 2018

Improved Gameplay for Codenames

I really like the board game Codenames. Especially good for 6-10 players. But as is usually the case, I have ideas for improving the rules.

There aren't enough lead changes. It usually comes down to whichever team misses one opportunity loses. Here is how I would like it to work:

  • A 6x6 grid instead of 5x5. Increases the intensity, and the opportunity for multi-matches.
  • (2) Spy cards. In playing ~30 games, I have never seen the Spy decide the game. That is mostly good, but I would like to make it a bit edgier. 
  • No 1-card clues--2-card minimum. Makes it much more challenging, especially with the above adaptation. (One mitigating strategy, if desperate, would be to include a neutral card in the clue.)
(Obviously, the pre-filled 5x5 maps the game comes with won't work. But it would not be hard to create your own 6x6 grids in Excel and print them out.)

Saturday, November 24, 2018

Anti-Pattern: Part 1 of Series Does not Part 2

Often when using the web for technical research, I will come on Part 1 in a series of articles. If Part 1 is useful, the next thing I typically want to do is read Part 2. Wouldn't a link be helpful, to let me click on through to Part 2??

I understand there is a design--challenge: often Part 2 doesn't exist when Part 1 is create. The obvious solution: when creating Part 2, take the extra time to go back edit Part 1 to add the link. The even better solution (not always possible) is to have a structure so that you know in advance where Part 2 will be posted. In that case, the editor can add a note "When available, Part 2 will be posted here".

(The article that pushed me over the edge, providing the impetus for this post, was this otherwise excellent article at O'Reilly, who really out to know better.

Monday, November 19, 2018

The NORAD-Santa Thing Sounds Like an Urban Legend


I've known of the Norad Tracks Santa thing for decades, probably before I ever heard the term urban legend. But if I were hearing of it today for the first time, I am pretty sure I would conclude "must be a UL". So, mark that down in my UL-detection record as a (fairly rare) false positive.

Of course, even true stories that become legends morph. Per Wikipedia:
Over the following years, the legend of how the annual event originated began to change. By 1961, Shoup's version of the story was that he had not been gruff with the child, but instead had identified himself as Santa Claus when he spoke to the child on the phone. Shoup and his family later modified the story further, adding that the child had dialed the "red telephone"—an impossibility, because the hotline was connected with the Strategic Air Command by an enclosed cable, and no one could dial into from the outside—rather than the regular phone on Shoup's desk, that it was a misprint in an advertisement that led the child to call him rather than the child misdialing the number, and that a flood of calls had come in from children on Christmas Eve 1955 rather than from just one child on November 30.

Sunday, November 18, 2018

What if Everyone Had the Risk Tolerance to Invest in Equities?

I invest 100% in equities. I have no plans to change this, even in retirement (still years away). Both theory and empirical evidence indicates that, over a reasonably long time horizon, equities provide a much better rate of return than bonds or, heaven forbid, CDs. So my question is, what would happen to the economy if all savers had a risk tolerance for equities? Set aside the transition effects, obviously it would be disruptive if it happened overnight. But assume over the course of a generation, everyone wises up and develops the risk tolerance for equities. What would happen?

1. Would return on equities go down, since more capital is available?
2. Would economies become more productive, since middlemen are being cut out, and risk capital is available?
3. Something else entirely?

Sunday, October 21, 2018

Navigation Apps: Better Historical Data to Extrapolate ETA

Modern navigation apps, such as Google Maps, are amazing. Refinements such as telling me which of the two left-turn lanes to take are incredibly impressive. But based on my limited experience--I am a full-time telecommuter who only occasionally drives the 30 miles across the metro to go into the office--there is one area I see a lot of room for improvement.

If I start navigation early into rush-hour, for my 30-mile cross-metro commute, I might see ETA 42 minutes. But that will typically elongate as I drive, peaking around 51 minutes. So the missing ingredient, particularly relevant for long, rush-hour drives, is the degree to which congestion builds. I have to think Google has more than enough data to build this in to their algorithm.

App Idea: Recycling

People suck at recycling. Primarily because they make mistakes of commission, motivated by good intentions. Placing a contaminated item in recycling is ~10X as bad as not recycling the same item, when not contaminated.

Ironically, I was having this very discussion with a colleague, who themselves started the discussion. I brought up my favorite example, the cardboard pizza box. Recyclable if never used, but certainly not once full of grease! To my astonishment, this was news to said colleague!

This suggests to me a mobile or Alexa/voice app idea. I envision a universal app (within the USA--apparently one already exists for the small but beloved country of Belgium) that provides directions, based on zip code and potentially other factors, such as waste-hauler, regarding what can and cannot be recycled. It would work really, really well with voice input.

(It is possible to go overboard the other way--it is wasteful of both time and water to wash containers very lightly "contaminated" with their contents--empty soda bottles, etc.)

Tuesday, October 16, 2018

Rules for Napping


I am a big fan of short, rejuvenating mid-late afternoon naps. Just enough to take off the edge of unproductive drowsiness, without plunging into a coma-like state that makes re-awakening excruciating, or messing up one's sleep-wake cycle.

Set an alarm. This is the first commandment. Never but never nap without an alarm.

Keep it short. Everyone will have a different sweet spot, but in general, it seems like 30 minutes is the max. For me, assuming I fall asleep very quickly, 18 minutes is the ideal.

The rest of the tips relate to not getting toooo comfortable. First is: not in your bed. Don't let you body conflate napping with prolonged sleeping.

Not too warm--it is much easier to rouse one's self if not enveloped in a womblike state of torpor.

Not in the dark. The very worst thing is to start a nap at dusk, and wake up after dark.

Sunday, October 07, 2018

Feature Set and Rationale for Night Suite App (Part 1)


From the day I got a smartphone, it was obvious to me that the standalone alarm clock was obsolete. Nevertheless, in the ensuing 10 years, I have continued to search for the perfect alarm clock app (Android). I still haven't found one. Alarm Clock Plus was close, until it quit being maintained in 2014. I have switched to Alarm Clock Xtreme, which is good, but not great.

What has been on my mind recently, though, is that even a perfect clock app did appear, I still wouldn't be satisfied. I want a Night Suite app. The alarm clock would anchor the night suite, but there is a lot of important, complementary functionality that I want to be fully integrated with my alarm clock.

In my view, a compelling night suite would integrate an excellent Alarm Clock with the following key functions: Do Not Disturb with Whitelist, Night Clock with super-dimming, low power warning and super-reliability.

Do Not Disturb with Whitelist

This is pretty basic, and there are good apps for it, but I want it to be integrated. If my family members need to get hold of me in the middle of a night, I want Do Not Disturb to be overridden. Not if they send me a text, mind you, but if they call my phone at 2:00am, I do want it to ring.

Bonus feature: option for whitelist or non-whitelist members to receive an auto-response when they text during Do Not Disturb hours. Customizable, here is the default: "John Doe is in Do Not Disturb mode. If this is an emergency, you can try calling to bypass Do Not Disturb mode".

Night Clock

This one is stupidly simple, but shockingly hard to find. I want a very, very dim bedside clock. Similar to an LED alarm clock--bright enough to see, dim enough not to act like a light source and cast light on my sleeping eyeballs. Currently I use a combination of an app that dims but not enough, and a filter that imposes a semi-transparent black overlay to further dim the phone.

This should be matched with sunrise functionality, that decreases the dimming toward dawn. This would need to be flexible, to handle varying degrees of room shading. This could be handled by a percentage function for Dawn Adjustment.

Low Power Warning

If power goes out during the night, and my phone is in danger of running out of battery before the alarm goes off, I would want to be woken up. Or, as is more often the case, when I don't plug it in successfully.

I think the basic functionality here would be to sound an alarm if the margin of error for power dips below a certain configurable level. Default would be 1 hour.

Super-Reliability

I am not nearly technical enough to know how to achieve this, but a world-class alarm clock, of any kind, needs to be super-reliable. So anything that could be done to ensure an alarm that is less than 10 hours away does indeed go off would be critical. The thing that comes to mind for me is that I think alarm apps can either have system alarms, or their own alarms. I believe I have experienced alarm app crashes where the app didn't re-start. Maybe one way to deal with this is for the alarm app to have its own alarm, but also set a system alarm for 2 minutes after its own alarm. When its own alarm triggers, it would shut off the system alarm. Thus, if for some reason the app alarm didn't trigger, the system alarm 2 minutes later would be the backstop.

Suite Functions

Single-push widget to go into Night Mode, as well as options for scheduled times for Night Mode. Scheduling functionality would include "X hours before alarm", where default is X=8.

Alarm Notification Icon

The typical Android alarm clock app seems to put an alarm icon in the Notification bar if one is set within the next 24 hours. As a user of pre-set and recurring alarms, I don't find this very useful. Most of the time it tells me what I already know. In the rare event I forgot to set an alarm, I don't want to rely on noticing the usual icon is not present; I want a positive notification when there is no alarm set, within X hours of my standard morning. Default X=8.

Advanced Features

This is my basic, must-have-list. See Part 2 for advanced and nice-to-have features.

Feature Set and Rationale for Night Suite App (Part 2)

Recap

From the day I got a smartphone, it was obvious to me that the standalone alarm clock was obsolete. Nevertheless, in the ensuing 10 years, I have continued to search for the perfect alarm clock app (Android). I still haven't found one. Alarm Clock Plus was close, until it quit being maintained in 2014. I have switched to Alarm Clock Xtreme, which is good, but not great.

What has been on my mind recently, though, is that even a perfect clock app did appear, I still wouldn't be satisfied. I want a Night Suite app. The alarm clock would anchor the night suite, but there is a lot of important, complementary functionality that I want to be fully integrated with my alarm clock.

In my view, a compelling night suite would integrate an excellent Alarm Clock with the following key functions: Do Not Disturb with Whitelist, Night Clock with super-dimming, low power warning and super-reliability.

Part 1 covering the rational and must-have features is here. This Part 2 covers bonus and nice-to-have features.

Core Alarm

On-the-fly snooze (a feature from Alarm Clock Plus). So when the alarm goes off, in addition to Snooze, and Dismiss, there is a third option: Snooze Until. When you select Snooze Until, it defaults to the alarm's existing snooze duration, but you can dial it up or down.

Skip until X date. My current alarm app, Alarm Clock Xtreme, has a "skip next". This is super-useful, but an advanced feature where I can skip until X date would be great for vacations.

Do Not Disturb with Whitelist

Option to reply with a text first time the number calls. Second call within period triggers a ring. Useful for travel, or relatives who aren't good at keeping track of time zones.

Low-key visual indicator if you received a call from a non-whitelist party. Possibly useful if you wake up in the middle of the night and notice and think the call could be important.

Timer

The on-the-fly snooze can be even more useful for timers.

Friday, August 03, 2018

Low Unemployment Yet No Wage Gains

I have heard so many news reports over the past few years about how tight the labor market is, yet employers are not raising wages. Here is my theory for one cause; I call it "Wages and Employment are Sticky, but Employees Do Tend to Expect Internal Equity".

Let's say you have 100 widget builders, their hourly rate is $12/hour. Demand is high, you need 15 more widget builders to meet demand. However, the labor market is tight, you aren't finding candidates at $12/hour. Based on some experimentation, you conclude that in order to attract those incremental employees, you need to offer a 10% premium to the current wage, or $13.20/hour. Given strong demand, you have pricing power (no discounts) and equipment utilization is excellent, so even at $13.20/hour, those incremental employees will be profitable.

Except--what about your 100 existing employees? They are experienced and loyal. If you are taking people in off the street at $13.20 per hour, don't the existing employees deserve at least that much?

And therein lies the problem. "Buying" incremental labor is not like buying incremental raw materials. Raw materials don't expect internal equity. In the above case, the incremental hourly cost, idealized to ignore internal eequity, would increase by 16.5%. But if we factor in the need for internal equity, it goes up to 26.5%--making additional employees much more expensive, and exerting far more pressure on profitability.

I'm not a trained economist, but that is my theory. I don't claim that it is a complete theory. For one thing, the same principle would apply over the ages--hardly unique to our time. Still, it seems I am surprised that I never hear this argument explored, in the innumerable news reports I have heard on this topic.

Saturday, June 23, 2018

Republican strategist Steve Schmidt renounces the Republic party and calls for the election of Democrats

(Now former) Republican strategist Steve Schmidt renounces the Republic party and calls for the election of Democrats:
this Independent voter will be aligned with the only party left in America that stands for what is right and decent and remains fidelitous to our Republic, objective truth, the rule of law and our Allies. That party is the Democratic Party.
...
Season of renewal in our land is the absolute and utter repudiation of Trump and his vile enablers in the 2018 election by electing Democratic majorities. I do not say this as an advocate of a progressive agenda [my ital].  I say it as someone who retains belief in DEMOCRACY and decency.
Steve Schmidt has done what so few other traditional conservatives have done. He has publicly articulated a clear line, differentiating between transient matters of policy and mainstream American political philosophy, versus the existential, anti-democratic perils of the Trump Party. Even classical conservatives who I generally appreciate, even when I don't fully agree with them, such as Mona Charen and Jay Nordlinger, have not been able to bring themselves to state this obvious truth: the most important thing the electorate can do, the pre-requisite for moving toward health and sanity of the nation, is to banish Trump, his party, and the 98% of his party who are enablers and worse.

Given our current two-party sclerosis, that means electing Democrats. Period. The next project after that can be starting a third, center-right, party grounded in values of the Enlightenment. But first we have to overthrow the Trumpists.



Saturday, April 21, 2018

Tax Simplification Opening for the Democrats

"Tax simplification" is typically a Republican issue. You know--eliminate all those confusing brackets. Or--we will make taxes so simple you can file your return on a postcard.

The Republicans had their go at tax reform, and I'm pretty sure it did not include too much tax simplification.

The thing is, there IS a huge opportunity here. In many countries, for the 80% of people with straightforward tax situations, it is even better than file on a postcard. The government agency sends you a pro-forma tax return. If you think it is correct, you sign and send back. That simple.

This could be done in the USA as well, but the tax-prep lobby stands in the way.

I think this would be a KILLER campaign theme for Democrats.

Jargonwatch: Cover Off, On

In my corporate workplace, I am starting to hear the phrase "cover off on" to mean, well, "cover".  Examples:
  • Can I call you to cover off on a few things?  
  • Do we have anything else to cover off on before we adjourn?
So not one, but two, superfluous prepositions.

I did a quick search, was surprised to find nothing on this usage. What I did find were complaints (1, 2) about "cover off"--which, I agree, is itself objectionable. One of the proposed etymologies was a mental combining of "can we cover this topic" [so] "I can check it off my list". I think that is a stretch.

Saturday, January 13, 2018

Home Circuit Breaker Labeling

Just spent 40 minutes updating and correcting the labeling on my home's electrical panel. Vagueness, unclear abbrevs, messy writing, omissions and outright mistakes. Some of that is inevitable human sloppiness, but it occurred to me that a few things could be done to improve it.

Instead of a label stuck to the panel, make it an insert in a sleeve that is stuck to the panel. Advantages:

1. Can remove it, so making it easier to write legibly.
2. Can replace cleanly with an updated copy--no cross-outs.

For bonus points, industry should standardize the template, so it can readily be printed off from the web. Here is a link to the simple template I made.

For super bonus points, include a schematic of the house, with rooms numbered, and cross reference breaker numbers to room numbers.

Wednesday, January 03, 2018

Amazon Feature Idea: Delayed Review Reminders

Amazon reviews are useful, but the best reviews come from people who have used the product for months. However, it seems like most reviews posted are from people who have used the product for days. I think human nature is either to be excited about a new purchase and review it shortly upon receipt, or never.

So my idea for a minor, optional Amazon feature would be to let buyers opt-in to be gently prompted to review things weeks/months after purchase. The exact duration of the prompt could be algorithmic, taking into account the nature of the product, and any gaps in the existing distribution of reviews.