Understanding: why Ed Sheeran did better than Adele in the UK

  • 14

Autumn Leaves

Streaming smoothness

Ok, after sorting the Oasis 1997 album sales controversy let’s go back to Adele vs Ed Sheeran case.

I can already see many of you screaming – how the hell is 671,542 units with streaming better than 800,307 without streaming? Quite simple – 25 sold that much thanks to the streaming unavailability.

Nobody can deny today the overwhelming place streaming is taking in the industry and how profitable it is for everyone – whatever what some artists who always want more money try to claim. By the end of its chart run, no doubt Divide will have Worldwide various millions of equivalent album sales thanks to streaming alone. While speaking about first week results though, one can never forget that streaming music is a smooth consumption over time unlike purchase which is fully registered from day one.

I’m aware this concept may seem a bit obscure so we will highlight it with figures. Below are UK sales per format of Divide on its first week:

  • 405,351 CDs
  • 13,546 LPs
  • 78,944 streaming units

As we speak, streaming is more than 2,5 times bigger than download sales in the UK, so it is rather surprising to notice that download sales top streaming ones with a 2.2:1 ratio, especially considering Ed Sheeran unreal domination on streaming, isn’t it? How is that possible?

As mentioned, streaming units, unlike download ones, are built over time. In fact, current formula converting streams into albums in the UK suggests you need to play as many as 1,000 times a song from an album top 12 tracks before accounting for 1 album sale. In other words, if you stream Divide in full one time every day, you will have contributed for one sale after 84 days only. Yeah, it takes you 12 full weeks before equaling to one immediate purchase. If you stream, say, four times per day the album, you will still need 21 days / 3 weeks before recording one full streaming album sale.

The question now becomes – in terms of consumers, how does 78,944 streaming units translate?

4 thoughts on “Understanding: why Ed Sheeran did better than Adele in the UK”

  1. MJD…

    This is a little bit off-topic but I need to know your opinion on something.

    What do you think of Ed Sheeran placing all songs from the same album in the UK official singles Top 20 on the week of release, while the streams of those songs were also counted within the albums chart? Do you think that’s normal/acceptable?

    Of course, Ed Sheeran is just the best example illustrating that problem… This is happening again with Drake this week and will happen again every time a big acts releases their album from now on, unless the OCC establishes new rules to prevent this from happening. (France has the same problem…)

    I do find it quite annoying that the singles charts are becoming completely irrelevant due to this problem. I know album-only songs started charting within the singles chart quite a few years ago now when digital sales started, but with the streaming era it is becoming a real problem!! Especially since streams are double counted, once in the singles chart, once in the albums chart, which in my opinion artificially inflates “sales” and therefore gives an inaccurate picture of how successful an act/an album is.

    What do you think?? I do think the OCC definitely does have to implement some new rules to regulate the charts… Before they become totally boring and irrelevant (especially the singles chart).

    1. Hi Mirai!

      I pointed out this double counting issue way back in 2014 when the OCC first introduced streams in their ranking. The very first formulas I presented to factor in streaming happened in 2013 and involved no such double counting. It took Ed Sheeran smash to get people aware of it.

      While I think the OCC rankings are wrong there I consider chart watchers to be even more wrong than them though. Most of them are looking at it the wrong way. Rather than trying to set superficial rules like excluding songs or limiting ratios of streaming, it is needed to understand fundamentals of what we are looking for.

      Everyone expects one of the following two things from charts, which are fundamentally different:
      1) reflecting popularity
      2) accounting for sales

      The case 1) is very well represented by Billboard Hot 100 and BB200 rankings. They do use Streaming both sides, but that is a correct move since streaming songs do reflect the popularity of both songs and an album – Ed Sheeran is the perfect example of that. There is no “double counting” as Hot 100 points aren’t used for sales totals, they are only points. Thus, they cleverly separate sales accounting (Soundscan figures) and popularity indicators (Billboard Rankings).

      For people watching for case 2) they can’t accept that the same stream is being accounted for in both sides Singles and Albums. They are wrong in their expectations though. One looking for accounting sales in 2017 World just can’t expect to do so within’ “singles” and “albums” folders. It is needed to go further and compile an Artist ranking only – like CSPC – because it is impossible to assign one stream into a “Single”, an “album” or even a compilation. As streams replace the consumption of all those old products, as they really are “playlists” (hi Drake), the only solution would be separating all different sales lists (streaming, download singles, album sales) and creating a Top Artist ranking that merges them. Trying to account for sales while sticking in ‘Single’ and ‘Album’ folders is just impossible.

      Ironically, as much criticized as it may be, the Billboard provides a singles popularity list (Hot 100), an album popularity list (BB200), an artist ranking (Top Artist), and sales charts of al raw products (Album Sales, Physical Singles Sales, Hot Digital Songs, Top Streaming), creating the perfect body of charts one can hope for.

      1. Sorry for taking so long to reply. It is a tricky issue to address I thought I’d take my time before replying.

        First I would like to know… What were/are those formulas that you created to factor in streaming? Are these very different from the ones currently used by (for instance) the OCC? I’m very curious to know!

        You’re right. I didn’t really realize double counting existed, and I didn’t really realize how big of an issue that was until the Ed Sheeran case happened. But now it’s just so obvious that something is wrong… I mean, 3 weeks ago it was Stormzy, then Ed, and now Drake. Now we have a clear picture of what’s going to happen more and more in the future if nothing changes regarding the way charts work at the moment.

        Of course, I have to admit that I’m not very pleased by what you say… I like having dinstinct albums and singles charts in the “old-fashioned” way but I also know that you’re generally right in what you suggest and slightly ahead of your time… So it looks like the charts – as they’ve existed back to the 60’s for the OCC – will have to change sooner or later in order to adapt to the new era, before they become a complete joke.

        What I don’t understand though is… Why don’t you think we could just make albums song ineligible to chart in the singles chart? I mean, why couldn’t we just ask record labels to “nominate” singles-chart eligible songs (based on some additional criterion maybe?) and exclude the rest? To this day, it seems to me that it is still possible to distinguish a “single” from an album song. Like, the distinction is simply established by determining whether a song is promoted or not : is it getting airplay or is the record label pushing radios to play the song? Does it have a video? Has it been performed on TV or at events unlike the other songs from the same album?

        In Ed Sheeran’s case, we can clearly see that out of the 16 songs on the album, 3 of them are “proper” singles, all the others are just album songs (at least, for now). The 3rd and current single hasn’t been officially designated as such, but it is #5 on the UK radio airplay chart so we can clearly see that it is being promoted – whether it was done voluntarily by the record company or not… And usually – at least in my opinion – it is possible to determine which songs are “singles” and which aren’t quite easily. Therefore, why couldn’t we just keep the singles chart the singles chart and count the streams of the non-single songs for the album only? Why do you think this isn’t possible?

        I totally understand what you say about the Billboard charts. But do you really think there’s absolutely no way to keep just one singles chart and one albums chart? I mean, I don’t really like the way the Billboard charts are organized. Plus with the RIAA having different criterion for certifications, it makes the whole thing quite uneasy to understand… The OCC charts have always seemed much easier to understand… I’d like it if they could stay this way.

        An artist-based chart sounds quite weird… Do you know how the Billboard calculates their Artist 100 chart? Because to me it seems that newcomers would have a very tough time charting high on such a chart due to have a much smaller catalog/discography than well established artists. But maybe I’m wrong. Obviously though, an artist-based chart doesn’t seem as exciting as a singles or albums chart…

        Or maybe the solution could come from some technological evolution which would enable us to know if one song is streamed “as part of an artist’s album” or if it is streamed “individually” so that we could seperate those streams and count them in the different charts accordingly?

        Sorry for asking so many questions, but I’m getting kinda passionate about this issue… As you’ve probably guessed I’m quite “conservative” in the way I consider this issue, but I know that your analysis is more “fact based” and as I said earlier it’s probably closer to what will eventually happen than mine…

        One last thing: Do you think including streaming on both charts was a good idea? I mean, what if the OCC went backwards and decided to delete streams from the albums chart and count all streams on the singles chart instead? Wouldn’t it be more realistic and more accurate? Of course, the singles chart would still be the exact same mess that they are becoming… But at least there wouldn’t be double counting anymore!

        PS: if it’s easier for you, you can reply in French, because I think we are getting quite “technical” now haha

  2. Hi Mirai!

    Indeed we are getting into the details but this is indeed a very fascinating subject! Kind of deserve its own article – there will be more about it in the near future!

    Answering your questions still, let’s try to structure it to get understandable!

    1) Distinction Single or Album track
    You mention 3 proper singles for Ed. To me, defining which track is a real single or not isn’t the real technical difficulty but I’ll still answer on that. On most cases, which song is a single is kinda obvious. It gets less and less obvious in this days yet. Ed is a good example actually – How Would You Feel (Paean) spent 3 weeks Top 40 of the UK Airplay Chart. On its side, Galway Girl was the 3rd top selling track before reaching the Top 50. While now with the former dropping out and the latter climbing the situation seems clear, it wasn’t when the album first came out. Also, the Hot 100 always aimed to reveal to the industry which songs deserved the most investments for promotion. With that spirit in mind, excluding album tracks wouldn’t make sense – although of course streaming figures are easily available for them.

    2) Issues of exclusions
    I mentioned bigger technical flaws that this debate to not exclude tracks. The whole problem is on how do you tally your figures – do you set a weekly figure per song and every week your release to date is increased by that figure, or do you maintain real release to date totals. This brings major flaws to exclusion rules as a song isn’t by nature a single – at best it will become one.
    Just think about a track ‘selling’ 10k per week during 20 weeks before becoming a single. When it is announced as a single, what is its release to date tally? 200k, on which case its chart run is massively flawed, debuting after already moving significant amounts, or 0k on which case its sales total to date is massively flawed as it doesn’t reflect all persons which went after the song.
    In other words, excluding songs means creating a severe inconsistency between charts and sales, while to me the former should reflect the later.

    The issue may even become bigger if you exclude “album tracks” from the single ranking and thus count their streams into the album chart. It means the “to date” total of streams will be absolutely impossible to understand – some went into the album chart, others in the singles chart, both with different ratios (1/1000 and 1/150), it will be a huge mess and when OCC will tell you an album did 100k and a single 150k, you will have simply no idea of what it really means. So at the end of the day, figures will lose their meaning and charts due to exclusions won’t reflect sales anymore, thus that will basically break both popularity indicators.

    3) Streaming accounting for both sides
    Rather than wondering if streaming should account for albums or singles rankings, you should go back to fundamentals and ask – which purchase this new consumption way replaced? The answer is both. Thus, the answer is that streaming should really count on both sides since streamers use the service as a replacement of both a single or an album purchase. If you assign all streams into the singles chart, your “album” and “single” rankings will become “over 40s old music” and “under 30s music”, with only the public ranging from 30 to 40 years old accounting on both sides. I don’t think an album chart full of Elvis, Leonard Cohen etc and a single chart full of Chainsmokers is really good.

    4) Avoiding double-count of streams
    Obviously, while streaming has to be part of both charts since it replaced both products, we still need to properly separate streams in one or an other without duplicating everything. The very first formula I suggested in 2013 was an easy one to simplify the case – barely assume the weakest total of streams per song equal to “album streams”. If songs of an album have 100m, 90m, 85m, 12m, 11m, 11m, 10m, 10m, 9m, 9m, 9m, 8m respectively, assume 8m album streams. That’s 8m * 12 (songs) = 96m, divided by 1500, that’s 64k album sales. Then you remove those album streams from all songs to get “singles streams” which thus end up being 92m, 82m, 77m, 4m, 3m, 3m, 2m, 2m, 1m, 1m, 1m, 0. That way, proper singles continue to perform well, “album tracks” won’t disturb the highest region of the single chart anymore but still ‘viral’ songs to come are still visible since they aren’t fully excluded, and we avoid double-count. It isn’t progressive enough to me on this day and age, but if the goal is to get charts continue to look similar as old singles / albums lists, that’s a good solution. Also technically it is easy to track in the long run – even if the “weakest” song can change from one week to an other, to get the release to date total of a song it will always simply be its total streams minus the album streams.

    5) Artist-based chart
    In all fairness, the Billboard 200 is already pretty much an artist list since we mix together different records. It is album-themed, which is fine to me. I suggested in an other forum doing just that, an artist ranking featuring only ‘new’ material (less than 18/24 months old), with a second list for catalogs. It would be fascinating to see a “Top 100 Artists Catalog” list to see the Beatles, Michael Jackson, Queen etc complete catalogs compete week after week, and new artists will be compared fairly avoiding issues of some of them not being available for stream or whatever. Rather than a fixed criteria (xx months) to define what is new or old we can barely check, as you mentioned, what is being promoted or not to avoid issues with artists like Drake releasing albums very often.

    6) Certifications
    Believe me, while we think OCC figures are easier to understand, they are truly a mess. The album total for example, as we use only the Top 12 songs and artificially downgrade streams of the Top 2 tracks, it is absolutely impossible to re-calculate the number of “album streams” from the available data, so if the OCC doesn’t provide it, we have no clue on how many copies albums like Purpose or Views “sold” thanks to streams and no clue on which singles got double counted or not into those “album sales”. The singles side is just as messy. Since January they changed the ratio from 100/1 to 150/1 so now we have no idea of how many copies each song sold. For example ‘One Dance’ got its streams counted as 100/1, since 3 months its weekly sales are tallied using 150/1. Thus if you add its sales and its streams, you do NOT get its current OCC sales to date. Also, while they changed the ratio for charts, the old one remained for the BPI. For example, Shape of You is officially on 1336k, but I can tell you the single will be 3xPlatinum this week as with BPI ratio it is on 1711k. It really is a massive mess. All OCC historical totals are now flawed and since they hate contradicting themselves, they will remain wrong. That’s why no matter which formula one decides to follow, raw data charts and totals must remain independent, to be able to switch of method if the new market environment requires it.

Leave a Reply

Your email address will not be published. Required fields are marked *