The Big Friday Effect

The Big Friday effect is what I have called my observation that more mail is delivered on Friday than on any other day of the week.  It became apparent very quickly as I began to analyze my USPS mail and has remained through the entire year that I’ve been conducting this mail-counting experiment. You can see the Big Friday effect in the figure below, which plots the total number of pieces of mail I’ve received by weekday.  It is a curious effect that has an interesting cause.

The quantity of mail, by category, with the day of the week it was delivered.

The quantity of mail, by category, by the day of the week it was delivered. Notice that Friday is significantly higher than the other weekdays.

I wanted to look more deeply into the weekday distribution to understand what is behind Big Friday. I analyzed the mail from individual senders to see how it was distributed throughout the week, restricting the analysis to the top 15 largest senders (see below).  This limits me to senders with mail volumes of about 1 item per month or more.  Any sender with less mail volume than that won’t be able to have much of an impact on any given day.

Mail totals for the top 15 senders of mail (to me), broken down into categories.

Mail totals for the top 15 senders of mail (to me), broken down into categories.

Plotting each sender’s mail into weekdays is revealing.  Most of the them have mail delivery distributed throughout the week, which is what you would expect for a mostly random sending process.  There are two notable exceptions — they are the senders with the majority of their mail deliveries concentrated into one weekday.

The weekday mail totals for the top 15 senders.  Note that the y-axes all have independent scales.  Note the scale for Redplum.

The weekday mail totals for the top 15 senders. Note that the y-axes all have independent scales. Note the scale for Redplum.

The Amherst Citizen is a local newspaper which is generally delivered on Wednesdays.  This does give Wednesday a boost in the top figure, but it isn’t a huge contributor because it doesn’t come out every week.  It is also not that reliably delivered on Wednesday, with Thursday deliveries being about 1/3 as many as Wednesday’s.  But look at Redplum (lower left), the well known junk mail merchant.  With 44 deliveries on Friday alone it dominates the weekday totals.  Thursday and Saturday have two each.  Considering that there are 54 weeks of mail deliveries in these numbers, Redplum would seem to be very effective at getting advertisements and coupons delivered to households just in time for weekend shopping plans.

To illustrate just how strongly Redplum impacts the numbers, we can look the mail by weekday for these top 15:

Mail received by the listed senders by the weekday it was received.  Note the large contribution from Redplum on Friday.

Mail received by the listed senders by the weekday it was received. Note the large contribution from Redplum on Friday.

And then without Redplum’s contribution’s.  And so we find that Big Friday is all about one junk mail marketer being very precise with their product’s delivery.

Mail received by the listed senders by the weekday it was received.  Redplum was removed to illustrate it's effect.

Mail received by the listed senders by the weekday it was received. Redplum was removed to illustrate it’s effect.

Advertisements

The USPS and You

Every day but Sunday, a government employee comes to that place you call home and leaves you with any number of items.  Packages perhaps, but certainly letters, bills, advertisements, or magazines.  Most of these are sent to you by complete strangers.  Is there something interesting or valuable that can be learned by paying attention to what arrives in the mailbox?

Questions we might want to ask:  “How much mail do I get?”, “Who sends me mail?”, “How often do they send it?”, and “What kinds of mail do I get?”  Advertisers certainly have each one of us in their databases.  I’m sort of curious to know something about what they think they know about me.  But I’m also eager to explore what can be learned by simply paying attention to something that goes on around me with a high degree of regularity.

I’ve mentioned this before, but my methods here are to record the sender and the category of each piece of mail I receive daily.  This is for mail specifically directed to me, or not specifically directed to anyone (i.e. “Resident”).  I’ve been doing this since the end of July 2014, so I have a fair amount of data now.

Let’s start with quantity.  On average I’m getting about 100 pieces of mail per month.  This is pretty consistent over 8 months, but note that things picked up at the beginning of November and then dropped back in January.  The rate (i.e., slope) didn’t really change, there was just a shift in the baseline.  The November shift is undoubtedly from election related mail.  The January shift is the post-Christmas dropoff that we’ll see later.

Cumulative amount of delivered mail.

Cumulative amount of delivered mail.

One of the more interesting observations is the breakdown of the mail by category.  It should come as no surprise these days that the majority of mail is advertising.  If you include political adverting (a category I break out separately), this overall advertising category accounts for more than half of the mail I get in my letterbox.  Considering that the USPS’s own numbers suggest about 52% of the mail was advertising in 2014, it looks like my dataset is representative.  Interestingly, the percentage of mail that was advertising in 2005 was only about 47%, so the percentage of mail that is advertising is on the increase.  This is not unexpected.  The NY Times published a piece in 2012 indicating that the Postal Service had announced their plan for addressing the huge decreases first class mail.  It was to focus on increasing the amount of advertising mail that they carry.  The Wall Street Journal has a piece from 2011 showing that the advertising percentage was only about 25% in 1980 and has been increasing steadily ever since.  Mission accomplished.

Categorical percentages of delivered mail.

Categorical percentages of delivered mail.

The next largest category, “Information”, is communications from people that I know or businesses that I deal with.  In other words, mail I want or care about in some fashion.  This is about 22% of the total.  Bills are a separate category as I think they are different enough to track separately.  Yes I still get magazines.  No I don’t wish to convert to a digital subscription.  But thank you for asking.

I find it interesting to look at the breakdown of the composition of the mail over time.  Judging from the sharp changes in color in the largest category (bottom bar), you can probably guess when the last state primary and general election took place.  But note that in general, each week is dominated by advertisements.  Notable times that this is not true are the week leading up to an election, when political advertisements dominate (note that these are still advertisements), and the weeks leading up to Christmas.  This last week shows an increase in “Information” mail largely because of Christmas Cards.

Weekly mail by category.  Note that 2015 began mid-week.

Weekly mail by category. Note that 2015 began mid-week.

Let’s look more closely at the advertisement mail numbers all by themselves.  October was the peak month, which is somewhat surprising given the frenzy over the Black Friday shopping.  Predictably, direct mail fell off in January after the end of the Christmas shopping season.  But somewhat surprisingly it climbs back without too much delay.

Amount of advertising mail received each month.

Amount of advertising mail received each month.

So who exactly is it that sends me so much junk mail?  Good question.  Redplum is the biggest of them all by far.  Also known as Valassis Communications, Inc., they provide media and marketing services internationally, and they are one of the largest coupon distributors in the world.  In other words, they’re a junk mail vendor.  You can count on them, as I’m sure the USPS does, for a weekly delivery of a collection of ads contained inside of their overwrap.  After that I have Citibank, Bank of America, SiriusXM, and Geico, in that order.  I would not have expected Geico to show up this high on the list, but there they are.

The amount of advertising mail sorted by sender.

The amount of advertising mail sorted by sender, restricted to those with 2 or more pieces of mail being delivered.

Another question to consider is when does all this mail come?  We looked before at the monthly advertisement mailings numbers, but we can dig a little deeper and look at how mail deliveries vary by weekday.  If we look at raw numbers, we notice that Friday is by far the biggest mail day in terms of the number of items received.  This has been consistently true for the entire time I have been analyzing my mail.  I don’t have a good explanation for this observation.

The quantity of mail, by category, with the day of the week it was delivered.

The quantity of mail, by category, with the day of the week it was delivered.

But there’s more to it than just that.  We don’t get mail every weekday.  Lots of federal holidays fall on Mondays where there is no mail delivery.  What we really want to do is to look at how much mail we get for every day that mail was actually delivered.  This lets us compensate for an uneven amount of delivery weekdays.  When we do this, we find things even out quite a bit.  Big Friday is still the king, but the other days even out quite nicely.  Understanding what is going on with Friday deliveries is something I’m interested in.

Mail by category each weekday, normalized to the number of days mail was delivered each weekday.

Mail by category each weekday, normalized to the number of days mail was delivered each weekday.

What you can see from all this is that you are (or I am, in any case) more likely to get certain types of mail on some days than on others.  This is somewhat easier to see if we plot each category by itself.  I find it remarkable to see that I basically don’t get bills on Wednesdays.  Credit card applications come primarily Saturdays.  Charities don’t ask me for money on Mondays.  And political ads come on Thursdays and Fridays.  I’ll bet that if I further broke down the advertisement category into senders that more weekday specificity would emerge.

Normalized daily mail categories per weekday.

Normalized daily mail categories per weekday.

In the interest of completeness, we finish up by looking at the statistics of the daily mail delivery.  That is, how often do we get some particular number of pieces of mail in the letterbox?  Here we don’t concern ourselves with the category, only the quantity and how many times that quantity shows up.  We can see from the plot that we most often find three pieces of mail and have never found more than thirteen.  This distribution in quantities approximately follows what is known as a Poisson distribution.  It has nothing to do with fish, but rather was named after a French mathematician Siméon Denis Poisson.  The red line fit is a scaled Poisson distribution with the average (lambda) equal to 3.5.  This indicates that, on average, I get 3.5 pieces of mail daily.  This is slightly lower than the mean value from the plots above of 3.9, but they’re calculated in slightly different ways and have somewhat different meanings.

The distribution of mail quantities follows a Poisson distribution.

The distribution of mail quantities follows a Poisson distribution.

The most unexpected things that I have observed are the Big Friday effect, and the amount of regularity in the weekly of delivery of some specific types of mail.  As they have endured over eight months of data collection, I am inclined to think they are real, but it will be interesting to watch and see if they exist after an entire year of mail collection.  It is also interesting that the Wikipedia article on the Poisson distribution specifically mentions that it should apply to mail, seemingly appropriately, but I can find no record anywhere that anyone has actually done this experiment.

Eleven Weeks of Counting the Mail

For no good reason outside of my own curiosity I have taken on a project of analyzing what shows up in my mailbox.  After eleven weeks of collecting information, there are some interesting things that I’ve observed.   How useful this project is, or will be, remains questionable.  But I can find no signs online of anyone doing this sort of project, so perhaps it will reveal something valuable.  Even with eleven weeks of data, the numbers are still relatively small, so we should be careful not to draw any concrete conclusions just yet.

First, let’s start by looking at what has shown up in the mailbox each week.  As much as some folks rant against stacked bar charts, I think they can be useful in some situations.  Situations like this, for example.  Here we can see the weekly volume of mail that my local carrier delivered to the box, how that number changed over time, and what the majority category was for each week.  Each week’s bar is sorted by quantity, with the largest quantity category placed at the bottom.  This makes large changes easy to spot.  It should be easy, for example, to grasp when the last election took place from the sudden rise (and subsequent disappearance) in the purple bars.

Eleven weeks of USPS mail, by category.

Eleven weeks of USPS mail, by category.

In brief, over these eleven weeks I received anywhere from 13 to 33 pieces of mail each week.  The average was 22 pieces of weekly mail with a standard deviation of about 6.5.  I’m not sure that a Gaussian model is the right one for this type of data (Poisson is probably more reasonable), but it gives a reasonable summary for a smallish data set.  Examining the distribution statistics is something I intend to do once the dataset grows a bit larger.

The relative breakdown between the categories is not surprising.  Advertisements dominate the composition.  I trust the political ads will be competitive for second place as the November election gets closer.  What I think this shows pretty clearly is how the Post Office gets paid.  Fully half of the mail I have received has been advertising, either the normal (34.5%) or the political flavor (17.8%).  I don’t think I would have guessed it to be that high.  If there is ever legislation passed to limit junk mail, the USPS will be in bigger financial trouble than they’re in already.

Mail_totals_percent_by_category (11 weeks)

What I find terribly interesting, though, is the breakdown by weekday. Meaning, the day of the week that the piece of mail was received.  Eleven weeks is not the biggest sample size, and so this may not hold true in the long run.  But Friday is absolutely a dominant day for the quantity of mail.  And not just for a single category.  The three major categories all have big mail days on Friday.

Mail totals by day of the week.

Mail totals by day of the week.

We can see this more clearly by looking at the categories separately.  I’ve separated out the weekly totals in actual pieces of mail for these categories.  First are “Advertisements,” otherwise known as “junk mail.”  Friday stands above the crowd with 22 above Wednesday’s 19.  Monday, Tuesday, and Saturday all come in at 10, so this difference appears to be significant.

Advertisement_by_Weekday (11 weeks)

Then we have “Information.”  This is mail from people or businesses that I have a relationship with.  Doctors’ offices, schools, banks, etc.  This is a very big Friday category with 18.  The next largest day is Saturday with 9.

Information_by_Weekday (11 weeks)

And then “Political Advertisements.”  I’m somewhat conflicted on how valid this category is. There was an election held on a Tuesday, and the direct mail spike happened on the Thursday (7 pieces) and Friday (6 pieces) of the week before.  Since the election I have received at most one political advertisement on any given day.  So this contributes to the “big Friday” effect, to be sure, but it was just one week out of eleven, so its impact is mitigated by its relative rarity.  But note that eliminating the category altogether would still result in Friday being the dominant weekday for mail delivery.

Political Advertisement_by_Weekday (11 weeks)

So this “big Friday” effect seems like it might be real.  But the raw numbers are still relatively small and the time relatively short.  So I’m cautious about drawing too many conclusions just yet, but it certainly is an interesting, if preliminary, result.

Pre-Primary Political Mail

It is the day before the Primary election, so it is time for a political mail update.  Not quite the inundation at the mailbox today that I was expecting, just a single piece of political mail.  I’m disappointed, actually.  Saturday was a big drop from Thursday, which was the biggest day.  I would have expected the peak volume to be a little closer to election day, but then again, I don’t run campaigns.

Political advertisements, by date received.

Political advertisements, by date received.

Political mail by sender plot is updated below.  Still, no direct mail from any of our U.S Senate campaigners, although Mayday.us is pushing hard for Jim Rubens.  I remain baffled by the lack of direct mail ads from this contest.  Dan Hynes, running for State Senate in my district has the most direct mailings for the primary at 5.  Marilinda Garcia and Steve Hattamer close behind at 4.  Lots of folks with just one mailing, including Haverstein, who is running for governor.  I will be very interested in tomorrow’s results.

Political advertisements, sorted by sender.

Political advertisements, sorted by sender.

Analyzing the mail

Being a data person makes me do unusual things at times.  For example, I’ve been recording the contents of my mailbox each day in an attempt to understand what advertisers know, or think they know anyway, about me.  By mailbox I mean the container on a wooden post on the side of the street in front of my house, and not the electronic sort which contains promises of Nigerian wealth and deep discounts on Viagra.  I note the date, the sender, and the category of each item of mail not specifically addressed to anyone else in the house.  So anything generically sent to my address without a name, or sent to me specifically makes the cut.

Things like this require much patience from my wife.

Knowing that this is an election year might make this an interesting project.  Given that we have a primary election here next Tuesday, the political mailers are probably the most interesting story.  But first, let’s have a look at six weeks of mail, broken down into categories.  The bar graph is sorted so that the largest category is at the bottom for each day.  This is on the busy side, to be sure, but from the colors alone you can see that straight up “Advertisements” (in red) are the largest category, with “Political Advertisements” (purple) a close second (the two are maintained as separate categories).  You can see the last two days were a bit bigger than usual for volume.

Categorized mail by date received.

Categorized mail by date received.

This is clearer by looking at totals by category instead of by date.

Six weeks of mail, sorted by category.

Six weeks of mail, sorted by category.

That’s all fine and good. Nothing unexpected there.  Mostly advertisements, some bills, 401(k) statements, a few magazines.  What is remarkable though, is how the “Political Advertisements” category is growing.  That first graph is somewhat difficult to read carefully because it has all the categories in there together.  So let’s just see political ads on their own.  

Political advertisements, by date received.

Political advertisements, by date received.

Now this is interesting.  Given that the last few days have contained half the political ads I’ve received in six weeks, I can only imagine what the next few days will bring.  

Something that is interesting can be seen in the totals by “sender” graph (political mail sorted by by who sent it, plotted below).  Not from what is there, but from what is missing.  In the US Senate race, the biggest race in the state, four days before the primary, and I have received nothing at all in my mailbox from any of the US Senate candidates.  Nothing at all from Scott Brown or Bob Smith, but do note that Mayday.us is running ads in favor of Jim Rubens, though they claim not to be authorized by any candidate.  (It seems my naming convention for this category get picked up by the ad blockers, so if you don’t see some of these graphs, pause your ad blocker for the moment.)

Political advertisements, sorted by sender.

Political advertisements, sorted by sender.

I’m not sure what that means right now, but it does lend some insight into their respective strategies.   And as I wrote that my phone rang with a call from Scott Brown’s campaign…