The Scale I Use

Sometime in the late 80’s, Billboard published an insert in one of their issues called “How We Track the Hits.” At the time, they gathered their data for the Hot 100 by asking for ranked best seller lists from record stores and playlists from radio stations. They would take all of their lists and assign points to each listed song as follows:

Rank Points

1 25

2 24

3 23

4 22

5 21

6 20

7 19

8 18

9 17

10 16

11 15

12 14

13 13

14 12

15 11

16-20 10

21-25 9

26-30 8

31-35 7

36-40 6

Below 40 5

Then they’d total them up, giving greater significance to bigger stores and radio stations, and voila, the Hot 100. When I compile my Top 150 Hits of… lists, each song is assigned a point value for its position on the Hot 100 every week. I use the above scale, except I continue it as follows.

41-50 5

51-60 4

61-70 3

71-80 2

81-90 1

91-100 .5

If Billboard used their own scale for their own year-end chart, giving every song from 41-100 5 points, that could certainly account for many of the discrepancies between theirs and mine. But personally, I don’t think floating between 91 and 100 for five weeks is the same as floating between 41 and 50 for the same period. So I like my scale better.

I thought about adding 5 points to the entire scale so that #91-100 would be 5 points and #1 would be 30. That would be the same thing as awarding 5 bonus points for each week on the chart. But when I looked at a few sample results, it seemed to create as many discrepancies as it resolved, so I decided against it. Such point scales are fairly arbitrary anyway. The only truly accurate way to know what songs sold more singles and were played more on the radio is by having actual airplay and sales data, which Billboard started using when the technology became available in late 1991. So after ’91 my way of making my own lists gets a bit different too. More on that some other time.

One other tweak I make is that for each additional week at #1, the points for that week go up by one. So the first week at #1 is 25 points, the 2^nd week is 26, 3^rd is 27, and so on.

In the event of a tie, I list songs with the same total number of points by the following criteria.

1) The songs’ relative positions on Billboard’s year-end chart.

2) Peak position on the Hot 100.

3) Their relative ranks in Joel Whitburn’s Pop Annual books. If you don’t know what I’m talking about, check out recordresearch.com.

The Billboard chart year runs from Dec.-Nov, as do mine. So some songs that hit their peak in Dec. of a certain year will be listed on the chart for the following year. I know that for a few years in the 80’s, Billboard’s chart year began and ended in mid-Nov. I don’t know why and I don’t approve, so I’m sticking with starting in Dec. Furthermore, if an issue of Billboard is dated Dec. 1 that means it is for the week ending Dec. 1. Which means that for issues dated Dec. 1, 2, and 3, the majority of that week actually falls in November. According to Billboard, a chart year starting in Dec. starts on the on the first issue with a December date. For me, it starts on the first issue dated Dec. 4 or later. This affects under which year some songs end up being listed.

If the chart run of a song straddles the cutoff between chart years, as many do, the point total for it’s entire chart run is included, and it is listed in the year in which it accrued the majority of its points. For example, Smooth by Santana feat. Rob Thomas was Billboard’s #19 song for 1999 and #2 for 2000. Although it debuted in ‘99 and first reached #1 in late Oct. of that year, the clear majority of its points fall in the chart year 2000. So for me it’s #1 in 2000 and not mentioned in ’99. Get it?

If a song has an almost equal number of points in both of its chart years, I’ll favor the earlier, calendar year, even if its point total is a little greater in the year following, just because I like it better that way. How’s that for scientific?

Lastly, Billboard takes a one-week break from publication at the end of every year. Before Dec. 1991, it was their policy to freeze the chart from the previous week, with all songs maintaining their positions for those two consecutive weeks. What I do is look at the charts immediately before and after the skipped week and make up a mock of what the Hot 100 would likely have looked like if it hadn’t been frozen. Usually that only amounts to a difference of a few points in any song’s total, and often it makes no difference. But for one, it was the difference between being the #1 song of the year and being #2 (teaser for '88).