The year allocated to a song comes from the source charts. If a chart covers a particular year (for example the CashBox end of year chart) then that obviously provides the year. For weekly charts the normal convention is to list the month during which the song entered the chart, so a song that was in a chart for 16 weeks starting "Dec 1949" will be allocated to 1949 despite the fact that most of its run was during 1950.
The years are extracted from all the charts that the song is in, the median (i.e. middle) year is then assigned to the song. This normally gives a reasonable result but can be fooled by strange cases (like "Stand By Me").
As with the final position there is no room for adjustment. If we don't agree with the assigned year we just have to live with it.
Previous Comments (newest first)
28 Nov 2013
An addendum to my last note on Song Year
Actually, you seem to have already implemented a better idea than I suggested. For songs that chart twice, you have the song twice (not the year!). This you did for another Clash song, where you have "Rock The Casbah" with Year "1982" and another "Rock The Casbah (1991)" with year "1991".
What I also wanted to mention in my last mail is that another benefit of separating songs (or having two years specified) is that when a song clearly charted multiple times (e.g. "Should I Stay Or Should I Go" which charted in 1982 and 1991), you could rank the two chart occurences independently. That is, it would seem "not correct" to rank a song based on combined 1982 and 1991 charting (I understand that you would combine charting for songs that charted in country A in year x, and country B in year x+1).
I believe that "independent ranking for songs that chart muliple times" was your intent for including some songs twice. However, based on the charts referenced both for the ranking of "Should I Stay Or Should I Go" (in 1991) and "Rock The Casbah" (in 1982), I wonder if this occurred in these two instances, as there are references both to 1982 and 1991.
Tracking which year should be credited is much harder than you think, if a song entered the charts in November and was in a chart for a total of 20 weeks how do you divide the credit? It could be that it spent 9 weeks in the chart the first year or just 1. Also, since the position is important spending 9 weeks below 20th position, then climbing to the top as opposed to spending 4 weeks at the top and slowly dropping down should deliver different scores.
Unfortunately we have very few charts where we know the weekly positions, most of the data just has an entry date, peak position and period. We think the current approach is the best way to combine input from as many sources as we can.
Where we have firm evidence for completely seperate runs (by the same artist) we do end up with multiple songs (indicated by the year, as you say). This is not perfect but we feel in combination with the merging of years, it is the "least bad" option.
In the case of the Clash's "Should I Stay Or Should I Go" this is more complex. In the UK the song was released as a double A side in 1982 so is credited as "Straight To Hell/Should I Stay Or Should I Go", while in the US there were two distinct releases "Straight to Hell" and "Should I Stay Or Should I Go". In 1991 in Europe "Should I Stay Or Should I Go" became a hit. Now we could have decided to change the name of the 1991 release (to add the year) and the 1982 release (to split the double A side).
Our view was that we didn't want to do that, remember that we do list all the chart details, so anyone interested enough in the details is able to unpick this complete history just looking at the Clash's page.
As we say, not a perfect solution, but good enough for you, the reader, to be able to get the details.
28 Nov 2013
More Thoughts On Song Year
I believe I understand your algorithm for assigning a year to a song, but I am wondering if you considered (or would consider!) another option. If a song charts (in the top 100) in more than one year, then my understanding is it will still only show up in the top 100 in one of those years. Though it might be difficult to implement, have you considered assigning multiple years to a song to accomodate this. Without doing so, it seems a song may be "unfairly" excluded from the top 100 of a particular year.
Though I'm not sure it would qualify as an example of "top 100 in multiple years", my interest in "song year" was in finding that Should I Stay Or Should I Go (by the Clash) has a year of 1991. The song was released (and was at least a minor hit) in 1982, and will always be associated with 1982 to me. I wasn't even aware it was released in 1991 (and became an even bigger hit) until looking at your references. Again, though I understand your reasoning, it just seems weird to me that this song is 1991 (but of course there could be other people out there that it makes perfect sense for, as they first heard the song in 1991).
I also just want to commend you on the information provided on this website. I just discovered it, and it is a wealth of information. I greatly appreciate the effort you must have put into it.
28 Apr 2013
Re: Song Year
Your method is giving the same weighting to countries with vastly different population sizes. In the case of 'Keep Searching' (We'll Follow The Sun)' this means that Norway is given the same significance as US in your 'median' methodology for assigning it to 1965. Even with the combination of the other '1965' countries, your readership is going to be greater in the US and Canada combined, thereby demonstrating a flaw in your 'more readers' logic.
Also, you are not including all countries where a song charted. For example with 'Keep Searching', this song charted in Australia at the end of 1964. Therefore, you do not have all observations to determine the 'median' internationally.
Therefore, once again, I suggest that the most appropriate way to assign ayear to a song is according to the chart year of its country of origin - 'I'm Telling You Now' in 1963 (UK) and 'Keep Searching' in 1964 (US) etc. This will better show the 'sounds' of a given year while also meaning more to your readership in terms of when their own country's songs charted.
Population of countries has nothing to do with it, if it did India and China would dominate. We base our calculations on market size. Norway is much smaller than the US, that is exactly why there is one song chart from Norway and (for this song) 7 from the US. So Norway is not given the same significance as the US, CashBox is given the same significance as WABC NY, Norway gets the same impact as Record World.
Of course we can only include dates from charts that we have, the fact that this was a hit in Australia in a chart we have no data for cannot influence the year we assign.
Apart from those minor issue there are two different problems with your argument, first of all the song "Keep Searchin' (We'll Follow the Sun)" is not "clearly" from 1964. Not only did some US charts assign it to 1965 (such as the CashBox one), not only did the song spend significant parts of its US chart run in 1965, but also the North American market for music is about the same size as the combined European one and every country in Europe has it in 1965.
But, even if we agreed that this song "should" be assigned to 1964, how could we implement that? We have to process more than 30,000 song artists, are you suggesting we should assign each one a country of origin? How long do you think it would take to work that out (where does "Fleetwood Mac" come from, for example)? Even if every artist was assigned a country how would you set the years for hits that were not hits there? What if an artist came from a country without a chart? Your suggestion just leads to a whole load of questions.
The median year approach is not perfect, we can agree with you there, but it is "good enough" in most cases, it relies only on the data within the charts (no additional unreliable source of data is needed) and it is implementable within the constraints of available effort.
Until someone can suggest an alternate approach that is feasible we'll stay with this one.
23 Feb 2013
I suggest that each song be listed under the year of its first chart entry. For example, I'm Telling You Now by Freddie & The Dreamers first charted in the band's home country, the UK, in 1963. For many, it is clearly a 1963 record and does not belong in 1965 just because in got to No 1 in the US in that year (two years after its recording and release).
As we have explained on the "Why is the year for XXX listed as 19XX?" FAQ our decision is that we use the median year, not the earliest.
We are based in the UK, so we would also see "I'm Telling You Now" as a 1963 song. However, the simple fact is that far more of our readers live in countries where the song was a hit in 1965.
So, no, we'll carry on using the median date.
26 Feb 2012
I Can Help
Shouldn't this one be in 1974?
The year indicated is the year that the song was a hit (not its release). This is worked out by looking at the dates in the contributing charts. The song is in 12 charts as 1974 and 14 as 1975, so the year assigned is 1975.
Even in the US (which was the earliest chart date) it spent some weeks of 1975 in the charts, in most countries it didn't enter the charts until that year.
The selection of the chart year rather than the release date is deliberate, not only does that better match what most people expect but it also makes it easier to calculate the date.
22 Aug 2010
Celine dion's song Power of Love was out at least by 1988 I remember it being one of those "our" songs at a club that a former boyfriend and I used to go to. We'd dance to it everytime the DJ played it.
Wikipedia has a detailed article which you might like to look at.
The song was written by Jennifer Rush and was a major hit all round the world in 1985, except in the US where her record label, CBS, considered it "too European" and when they eventually released it it only made number 57 in the Billboard chart.
Celine Dion's version of "The Power of Love" was released on her 1993 album "The Colour of My Love".
So I'm afraid that you are mistaken, the song you danced to with that boyfriend was probably Jennifer Rush's version, or maybe even Laura Branigan's version (released in 1987) but it certainly wasn't Celine Dion's version.
In general the years assigned to songs are those in which the song was a hit (not the release date), so its quite possible that a song's year will be listed here many years after it was available. We selected this method because it best matches most people's memory of the era that a song comes from, where a strict adherence to release date would run into trouble with different territories and with cases like "Reet Petite", which was released in 1957 but became a hit in most Eurpoean countries in 1987.