American Top 40 PastBlast, 10/3/81: Greg Kihn Band, “The Breakup Song (They Don’t Write ‘Em)”

As I was looking over the 10/3/81 chart to consider what I might write about this week, I noticed several songs at the beginning of the show had fallen just a few spots. That’s a little unusual for tunes on the back end of their chart run: “Really Wanna Know You” slipped from #36 to #39, and three other songs from the first hour (“The Breakup Song” at #38, “Cool Love” at #35, and “Draw of the Cards” at #34) fell six spots from the previous week. (Additionally, “Fire and Ice” had gone from #37 to #41.)  That spurred recollections of periods from my chart-keeping years of what I called “slow fallers,” where multiple songs working their way down the 40 would slow to a veritable crawl for a week or two. The time that most stands out in my mind was the fall of 76, particularly parts of October and November, and without looking deeply, I’m sorta thinking that fall of 81 also had its moments in this regard.

That led me to wonder how I might check for relative periods of low or high chart turnover. Math nerd that I am, I looked for a simple proxy for chart action, and settled on a rolling average of the number of debut songs in a week. In periods of high chart action, this number should regularly be above average; the opposite would be true during a more slow-moving time.

Between my charts and Pete Battistini’s American Top 40 with Casey Kasem (The 1980s), with occasional help from the Ultimate Music Database, I had pretty easy access to the number of debuts from 6/5/76 to 8/6/88, when Kasem stepped away from the AT40 mike.  I entered them into a spreadsheet and computed two-week, three-week, and four-week rolling averages over the whole period. (To illustrate: my four-week average for 6/26/76 is the mean of the number of debuts on 6/5, 6/12, 6/19, and 6/26; for 7/3/76, I used 6/12, 6/19, 6/26, and 7/3, etc.) Initially, I thought the three-week averages would be what I’d use here, but ultimately decided the four-week chart yielded more insight.

The mean number of debuts over the whole twelve-year period is almost exactly 4 and 1/3. It’s not too surprising that the mean of the rolling averages is virtually the same. The standard deviation of the rolling averages is a little over 0.5, so the range over which we’d ordinarily expect a rolling average of four numbers to live is roughly 3.75 to 5. I’m positing that numbers outside this range, especially if they lasted for multiple weeks, could be considered a time of unusual action.

Twelve-plus years is a lot of data. A graph for the whole period is at the top, but it’s so dense I broke the timeline into two not-quite-equal pieces so things were more readable: 6/5/76 to 12/25/82, and 1/8/83 to 8/6/88.  Let’s take a look at the first of those now, which very nearly matches the period I was keeping charts:

There’s plenty of noise, but I think we can pick out a few things. I’ll start with slow periods.

–It does indeed look like the last four months of 76 had below-average turnover overall.
–The last half of 77 looks maybe a little slower than normal.
–Action cratered both in Dec 80-Jan 81 and Dec 81-Jan 82. In between, 81 looks slower overall than 78, 79, and 80, but there’s no sustained dip around the time of this week’s show.

Now, for the higher action periods:
–The first four weeks I wrote down my charts in June 76 was the second-highest active period in this exercise. That makes me want to go back farther in time to see what was happening over the prior year or so.
–July 78 stands out above everything. It’s well-known in AT40 circles that both 7/1/78 and 7/8/78 each had 8 debuts, which makes it easy to overlook that 6/24/78 had 6. I’ve always thought that May-June 78 had quite a few pretty weak tunes, which perhaps made the charts ripe for a mass purge.
–Less notable spikes are in May-July 77 and Oct-Nov 79; there weren’t many down weeks through much of 82 except for August and the very end, a little surprising since this was part of the “clogged charts” period.

There’s a radically different story as 83 begins:

The first 10 months of 83 is by far the longest-lasting average to below-average period here; I totally did not see that coming. There are other dips around Nov 85-Feb 86 and the first four months of 88, too.

But there’s much less action on the high end. Between 12/10/83 and 3/14/87 there was not one four-week block during which more than 20 songs came on the show (for contrast, there were over 25 such blocks in the period covered by the first graph). Late 83 tried to make up for what had happened earlier, but after that things were overall more settled.  When I went back a few years ago to look at charts from 84 and 85 in preparation for assembling Top 40 playlists, it felt like chart runs had become much more even and predictable than the wild late 70s. The part of the chart from Jan 84 through about May 85 seems to confirm that impression—a lot of stability in turnover there.

I don’t know if there are any big lessons here—you’d expect a lot of jitter in a chart like this, along with occasional forays far away from the mean. “Slow action” doesn’t necessarily mean “slow fallers,” either, which was what I originally sought. Another thing to consider before trying to interpret too much is that Hot 100 chart methodology was not consistent throughout (for instance, there was a big change around May 83). It might be interesting to look more into musical history to see if there’s anything more than chance variation going on, e.g., I see a number of the slow periods happened around the first of the year—were fewer singles traditionally released right before that? Thoughts and speculation are more than welcome.

—

It’s time to circle back to this week’s chart for a song, one of the four I mentioned at the beginning. The Greg Kihn Band came out of the Bay Area and were on a local label, Beserkley Records. “The Breakup Song (They Don’t Write ‘Em)” had peaked at #15. I bought the single pretty early on; it was very much one of my favorites then. As much as I like “Jeopardy,” I’ll take this one every time.