I don’t have time these days (this year, really) to take the multiple days of compiling, researching, crunching, analyzing, and arranging into a coherent post. Life is too busy right now, and that means I’m going to instead just drop a few simple but (hopefully self-explanatory) stats. (Check the data cronch tag for an explanation of how I extrapolate viewership rates, given Netflix’s black box.)
Note: I didn’t include gen:LOCK, because it’s not on a platform anywhere near Netflix’s reach, and it was also released weekly, not as a single bingeable season. That changes viewership patterns significantly.
First, SPOP’s S1/S2 total viewership is just barely (and by ‘barely’ I mean ‘by about one-fifth of a percent’) more than VLD’s viewership for S7/S8.
3Below, the continuation of Del Toro’s award-winning Trollhunters, couldn’t manage more than 37% of SPOP’s S1/S2 viewership. 3Below peaked lower than VLD S8, and despite a healthy tail, its total viewership ended at 28% less than VLD’s S8.
Meanwhile, tDP’s S1/S2 total viewership clocks in 174% higher than SPOP’s, 175% higher than VLD’s final two seasons, and 470% higher than 3Below.
Which is to say: DW did its level best to bury VLD’s final season, but it sure looks like DW just ended up burying everything, instead.
No wonder people keep thinking tDP is a DW production. It’s the only American animated series on a global platform that’s performing at anywhere near the level you’d expect from a major studio.
the bar was already so damn low it was practically underground
Well, I guess I’ll start with this quick correction, with thanks to @jeannettegray, who pointed out I mixed up days (for S1-S6) and weeks (S7) which would explain why it felt wonky when I was looking at it. Here’s the corrected chart of time in-the-tail versus not-in-tail:
I included S8 in this version, but without a season coming after it, the only thing to track is the tail itself... and unless something truly wild happens between now and tomorrow, we’re already out of the tail. S4 and S5 had tails lasting 14 days each. I guess S8 isn’t the worst, since its tail looks likely to be 17 days. Yay, three more days of elevated viewership than the two worst seasons.
Unfamiliar with the expression ‘long tail’? Here’s a reference image showing the rise and fall of a season’s viewership stats. More info in this S7 post, or you can just check my data-cronch tag.
Or for a more pointed visual, here’s the past six months: S6, SDCC, S7, and S8... the last of which shoots up and comes right back down. Like a rock.
For S7, I included IMDB rankings per season, and the range was wide enough I had to take good/average/bad and split it into four: great, good, okay, poor -- and S7 had more episodes in ‘poor’ than any other. That’s what I mean when I say the bar was so low the damn thing was underground... and yet amazingly, S8 utterly failed to hop over it.
In fact, S8 did worse. Much worse.
viewer feedback via IMDB
Here’s all eight seasons, with their respective episodes averaged into each season’s value. It’s... well. Look for yourself.
For comparison, I also did a breakout, where the full seasons got split into 6 and 7, so we're comparing apples to apples with S3-S6 split seasons. The pattern is pretty consistent, between first-half and second-half, in that the season-half (or half-season) with the ‘finale’ is usually the one with the higher score. People like big finale set pieces, and those tend to get highest ratings.
Unless, of course, you’re season 8. Then people hate you.
S1, S2, S5/S6 (as a two-part season), even S7, all have higher ratings for the finale-containing half. And then there’s S8. I usually try to avoid a word as strong as hate, but... there’s really no other word that fits when the difference is that stark. People really, truly hated S8, in case the Rotten Tomatoes score wasn’t enough to tell you that.
Given the scores for S8, I had to create a new, even lower level to see how the seasons broke out in terms of ratings per episode. Now we have great (green), good (blue), okay (yellow), poor (red), and... terrible (maroon).
Yeah. So. There you go. Remember when I said last season that The Voltron Show had lost its title as worst episode ever, unseated by the Feud?
Well, the Feud has been unseated... by ten episodes in S8. Take your pick, there’s plenty to choose from. As of the time of this posting, the current three worst episodes are The Zenith (5.2), Uncharted Regions (5.3) and Clear Day (5.5). I’d list the ten worst but then I’d just be listing all but three episodes from S8, anyway.
However, it’s interesting to see the pattern form when you put the split-seasons back together. You can see how in some ways, there’s a certain level of audience approval gradually building, I guess as people got used to the story’s vagaries, err, style.
The number of ‘great’ episodes (8.6 or higher) slowly climbs; the number of ‘okay’ episodes gradually declines. S7 breaks that pattern, with only one ‘good’ episode, a few ‘okay’ and the rest at poor. S8, well. Yeah.
Okay, moving along, to the IMDB traffic. This is another viewership-style stat, since a site’s page tends to get hit the most when people want to know about the show they’re watching.
The color-dot indicates the ranking for the week of release. S6′s biggest interest was the week it released, which tells me people were going direct to IMDB on the day of release or the next day. When that happens, it seems to correlate with viewers coming to rate every episode, often with a strongly positive slant. They want to get right in there and make their happiness known.
S7, the leap upwards came a week later, and the utter wildness of the rating (plus the delay) seems to be related more to rubbernecking and controversy, rather than viewers reporting in. S8 not only debuted lower than the previous two seasons, it also had a delayed reaction (indicating more controversy)... and it dropped right back down, in one of the biggest post-peak drops I’ve seen yet.
viewer feedback via twitter
About six hours after S8 appeared on Netflix, I caught this sentiment analysis of the twitter stream for ‘voltron’ as a keyword. This is from midnight California time (point of release) to about 6am California time.
This is the tracker where I can’t review sentiment, and judging from the other tracker I use, the negative is higher... but more negative than positive is never good, no matter the totals.
Twelve hours after release, I checked the other tracker. Traffic overall was pretty low compared to previous seasons (and nowhere near what S7 had prompted, not even in the same county). Sentiment was running about 3:1 negative --- that is, for every for-VLD tweet, there were three against.
(If it makes anyone feel better, #Allura was the center of the maelstrom, with Shiro a not-quite-close second. No, despite the impression on tumblr, people weren’t sitting back and accepting Allura’s fate.)
But seriously, the response was otherwise pretty subdued. This is a general traffic pattern + predictive, for #voltron as of two days after release:
Which is pretty much... nothing at all. If I didn’t know better, I’d think no season had been released at all. It shouldn’t look that flat when a season drops, and it sure as hell shouldn’t be predicted to fall within the week.
And back to the sentiment analysis, showing the trends over the first week and a half after release. It’s continued to fall in a semi-regular pattern, so this is a fair-enough idea of what it looks like. (It hasn’t changed much; it just keeps gradually dropping.)
24 hours after release, negative-to-positive was nearly 1:1; since then, it’s stayed pretty consistent with two positive tweets for every one negative. Which, okay, you might say, two-thirds of people expressing an opinion seem to be somewhat okay with things, right?
Except I think there might only be like a few hundred of them. I mean, the peak up there, of total mentions? Maybe 1200, total. Sure, it’s great that 510 people had something nice to say while 498 were unhappy, but... that’s practically three people in an empty auditorium, if you go by twitter’s usual traffic levels.
And no one’s sticking around, either. Once again there was a post-release bump thanks to the simmering controversy... and then everyone hung up the phone, left the building, went on vacation, but they sure checked out.
That leap upwards? That’s not the week of release. That’s the week after. The week of release (second bar from the right) is flat for just a bit, before spiking upwards. And just like the wiki-extrapolated viewership stats, it falls promptly afterwards. Like a rock.
(I so want to crack a joke about rocks fall, everyone dies, but... #TooSoon?)
There’s one last chart to explore, but this one’s a bit of a doozy to explain. Once you see what it’s saying, though, hopefully you’ll find it as interesting (and illuminating) as I do. I’ll leave it here for everyone to ponder, and get into it in the final S8 installment.
What are the chances of this hypothetical sequel of Voltron having better representation when it comes to POC, LGBTQ+ Rep, and overall better story direction/management? Do you believe that DW would be willing to gamble on such an endeavour, especially now that LM and JDS no longer have any involvement with the show?
I think it’s time media companies realize the real gamble is continuing to think that a team of all-white, all-straight, all-abled, all-men can deliver stories that satisfy the breadth and depth of a diverse and intersectional audience.
The majority of VLD’s flaws can be traced to the combination of two inexperienced and limited-vision EPs, and their chosen staff of all-male (and nearly all-white) team of writers. Other than one woman of color (who was also under-assigned despite her proven track record) and a single freelancer episode by a non-white man, the rest of VLD’s writers were white men.
As the showrunner, staffing was JDS’ responsibility; failure to affirm diversity behind the camera lies on his shoulders — and that of his boss, who approved such regressive hiring choices. No one was in the room to speak up for disabled characters, or female characters, or queer characters, or characters of color. Without those voices present, VLD’s progressive (and subversive, even) themes were even more easily dismantled in S3-S6. By S7-S8, those themes were practically torn out by the roots.
Think bigger picture: it’s not just one show. It’s getting a company to recognize the days of the all-male and/or all-white creative team need to be left behind.
But since there’s always the chance that VLD was a one-off, a complete anomaly in DW’s usual hiring patterns, I figured I could take an hour and do a sampling of credit listings on IMDB for DW's television division.
Data cronching behind the cut.
First: I grabbed directors and writers, as the easiest comparison points. I snagged 289 listings from 18 shows, broadcast going back to 2017 or so.
(Too late, I remembered a staff writer gets credited for every episode while on staff, and their ‘written by’ credits are usually much less. For now I’m leaving it on the grounds that while it’ll inflate the numbers, it’ll also help illustrate who gets put on staff more often. I’ll leave fleshing out the exact ‘written by’ credits if I get time to do a deeper dive.)
Here’s the shows whose staff info I grabbed:
3Below: Tales of Arcadia
All Hail King Julien: Exiled
Dawn of the Croods
Dinotrux Supercharged
Dragons: Race to the Edge
Harvey Street Kids
Home: Adventures with Tip & Oh
Kung Fu Panda: The Paws of Destiny
Lego Jurassic World: The Secret Exhibit
She-Ra
Spirit Riding Free
The Adventures of Puss in Boots
The Adventures of Rocky and Bullwinkle
The Boss Baby: Back in Business
The Epic Tales of Captain Underpants
Trollhunters: Tales of Arcadia
Trolls: The Beat Goes On!
Voltron: Legendary Defender
Here’s the breakdown of women-to-men in writing and directing roles.
In hard numbers, that’s 17 women directing compared to 67 men, a 1-to-4 ratio. For writers, it’s 51 women compared to 110 men, a 1-to-2 ratio.
Here’s the difference in the number of episodes they’re assigned/credited...
(It may be quite interesting to see how those values change, once the ‘staff writer’ inflation is taken out of the picture.)
If you look at the average for each gender, the split is somewhat close. Women direct an average of 8.5 episodes, while men direct an average of 8.6; women write an average of 10.2 episodes, and men write an average of 12 episodes.
Not all writers or directors work on only one series. Of all the people of a gender in a category (directing or writing), here’s the percentages that have credits for multiple series.
Just shy of 12% of DW’s women directors ever work on a second series. 37% of male directors do. The percents for writers is distorted out of context, though. In hard numbers, 10 women writers have worked on more than one series, compared to 9 male writers, which means 20% of DW’s women writers work on two or more series, vs 9% of men.
I grabbed race where I could (or where it was obvious from photographs), but it’d take a lot more digging to really nail down that list. I would be very, very surprised if the percent non-white in either category is greater than 5%, though. Just a hunch, but if I can find a way to determine that, I’ll put it on the list.
At the absolute minimum, DreamWorks has a gender issue... but I’d be willing to bet we’d find an even larger disparity for race, sexuality, or ability. (Note: those last two are a privacy thing, so the best anyone can do is a general guess. HR may have some idea, but it won’t --- and shouldn’t --- share self-reported details, so don’t think I’m holding out on you.)
It’s time again for a data cronch, or at least one that can compare the first two weeks of the final season to all the previous seasons. Before I get into S8, there’s one image from S7 that should give a stark sense of how S7 did.
(If you’re confused about some of the terms --- the tail, the lull, the hype threshold, etc --- see this post for a complete explanation.)
This particular chart compares the length of time between seasons to the length of time a season was in the tail, meaning above a set viewership threshold.
[whoops did my math wrong, will fix in the followup]
And then there’s this, from Tim Hedrick:
Let’s just say it’s not necessarily as good a thing as you think it is, Tim.
I’ll explain what I mean along with viewership and engagement stats for S8, behind the cut. Spoiler: it’s a lot of whimper and not much bang.
viewership stats comparison
Here’s the first two weeks of all eight seasons. S8 is the dashed line.
It started even with S7, and dropped quickly to run above S4 and S5... and by the second week, it’d dropped to run at the bottom rung with those two seasons. It took a hard hit on the holiday (much harder than the previous two years), then bounced up again above the tail threshold.
Here’s the same numbers, but adjusted to put each season’s peak together. This gives a better sense of the season’s tail, compared to the rest.
It’s cutting across the seasons. It’s not as severe as S1/S2 (which began at tremendous heights and thus had farther to fall), but it’s not as flat as S6 or S7. It seems to be falling in line with S4 and S5, for the most part. In case you’re unaware, those are not good seasons to duplicate, numbers-wise.
A third way to compare the tails is to look at their heights and their lengths. For this one, I adjust to line up the ends of each season’s tail (that is, when its viewership stats drop below the 3K threshold for more than two days running). Right now I’ve got S8 a little out in front of S4 and S5, duration-wise, but unless some bombshell drops in the next few days, I expect the tail to be done by the end of this week, perhaps as early as Wednesday.
Since the above charts go day-by-day, it makes it a little hard to get a sense of just how each week stacks up, comparatively. Here’s a chart that gives you that, comparing each week across all the seasons.
(Note: we’re only one day into week 3, so S8′s value is artificially low, there.)
In the first week, it’s not too bad. S8 is lower than S7, showing some damage from the lack of hype, but overall it’s middle-of-the-pack, being 5th highest viewership for its first week. (Seriously, you can see the impact when DW’s marketing puts the hype machine in high gear, and this isn’t it.)
The second week is where the bottom drops out, and it falls to only a few thousand above S4 and S5. The real sign will be if it drops below both of those in the 3rd week; that’s when we could safely say S8 really is the worst season.
Now I’m going to flip the script and look at the rate of change between the weeks. It’s a bit premature for this since we need to be into the fourth week to really have enough data for rate of change, but what the hell. Why not.
In a rate-of-change graph, a positive number (bar going upwards) means an increase, and a negative number (which all of these are) indicates a decrease. The longer the bar, the bigger the rate of change.
I’ll use S1 as the example in plain english. These values are from day-of-release to the following thursday (one week of broadcast). Think of it as: if 100 people watched in the first week, 60 watched the 2nd week. That’s a -40% difference. There was a further drop moving into the 3rd, of 30%. Again with our simpler numbers: if 60 people watched the 2nd week, 40 watched in the 3rd week.
The 3rd-to-4th week is the smallest by far, and it’s a sign something must’ve happened. Marketing launched a post-release campaign, or there was a con that got people excited. Whatever it was (idk, ask @ptw30, who tracks these things), it halted the drop-off, keeping the loss down to 11%.
Now you can see why this chart shows the real meaning behind the one above: of all the seasons, only S3 had a bigger drop between its 1st and 2nd weeks. S8′s rate of change is -64%. That’s pretty bad for what should be the final season. Buzz should continue a little longer, as viewers process where and how the story concluded.
(The second bar looks horrendous for S8, because it’s comparing the full 2nd week to only one day’s value for the 3rd week, so ignore that for now.)
engagement stats comparison
I’m going to hop over to look at how much people are talking about VLD. I track searches on three terms -- “Voltron Legendary Defender” (series title), “voltron” (term), and “Voltron” (series title). I average the values, chart ‘em, and we get this lovely picture, with the release week marked in the season’s color.
There’s one oddity I don’t think I’ve pointed out before: the level of engagement dropped slightly, between two weeks before S7 and the week of S7′s release. That’s... not how these things are supposed to work.
There’s a particular detail I look at in this chart, and that’s the drop-off. Here’s a closeup of S3 through S8. The blocks in lighter green indicate where the 3rd week level of engagement remained above the week of release. The blocks in red indicate a drop in engagement.
Every season has some chatter in its release week, followed by a jump in the second week. In S3, S5, and S6, the chatter remained above the release week level. In S4, S7, and S8, chatter drops off precipitously.
In other words, inside of two weeks, the engagement levels have dropped to below that of the week of release. You could characterize this as the audience taking a week to react... and then dropping the topic. Something in S3 and S5 had audiences lingering a bit more. S6 is the ideal -- it didn’t just get people engaged, it kept them engaged, staying above the release-week levels all the way up to SDCC.
demand analysis comparisons
Now let’s take a quick look at Parrot’s “demand analysis,” which is not actually viewership statistics. It’s how much people are talking, overall, on a variety of platforms. While Parrot isn’t about to release its algorithms, reading between the lines tells me it’s taking traffic on twitter, facebook, instagram, news sites, forums like reddit, and other platforms, to arrive at a single value. It does this for however many series, and then ranks the results.
Here’s the historical track of VLD, with S8′s values added.
Note: white circles are placeholders, because I couldn’t find data for every single week. Given the way demand cycles tend to work, it’s likely that VLD stayed on the chart in the gaps, so I put in placeholders for those weeks.
Yes, Tim, VLD has leapt up to sit at #7, but it debuted at #9, which it hasn’t done since S3. (S4 never charted at all.) Worse, S8′s second week -- and judging from previous behavior, its highest peak -- only just brought it up to the level of the opening weeks for S5, S6, and S7.
What makes the demand analysis somewhat questionable as a point of pride is that this doesn’t track viewership but chatter. When you consider S7′s 2nd week was marred by significant controversy, and S8 prompted a rash of the same (if at a lower volume), I really can’t say that #7 ranking is necessarily a sign everyone is cuing up the show on Netflix.
It just means there’s a lot of talk. It doesn’t necessarily qualify this talk as good.
What drives home the dubiousness of this value (in terms of judging demand analysis as some sign of popularity) is the Business Insider article Tim linked to, which has this line:
Which I guess is somewhat more diplomatic than showing this:
That brings me to audience reactions, which are a fascinating (if you like trainwrecks) view of what people really thought. I’ll get into that in the next part.
I went ahead and compared the six weeks previous to release night (iow, stopping at the day before release) for VLD, TH, Castlevania, tDP, and She-Ra, in order of release. These are traffic rates for the first season of each show. (For explanations on how this equates to viewership, look at previous posts under the data cronch tag.)
There’s a bump for She-Ra in the 6th week out, probably due to the timing of NYCC. Same goes for tDP, whose bump was at like 7 weeks prior. I can’t recall the con where VLD got its last bump, but it was at 5 weeks before release.
What’s odd is TH is the sleeper hit of the five, with almost no run-up buildup in hype. Castlevania pretty much comes out of nowhere over a five-week stretch, but I can’t recall if I saw any promos for it before suddenly it was plastered all over Netflix’s banner. Comparatively, it looks like She-Ra’s had a decent number of people checking in on its wiki page, compared to the run-up for other shows.
And then there’s this.
That’s the wiki/viewership stats for tDP, in green, and that tiny blue spike at the far end? That’s the first night of She-Ra’s release. It’s pathetic. Even Trollhunters and its minimal hype did just shy of 20K on its release night. She-Ra barely managed 13K.
It probably doesn’t help that DW and/or Netflix decided to move it from its planned friday slot to a tuesday, four days early. If that confused the audience (or not enough got the memo), there should be a huge spike come friday night from people (finally) checking in. If there’s not, then either the release spike has ended up spread out over several days or more due to confusion, or... there’s no real release spike.
At this point? Barring a later spike, She-Ra’s buried. Hell, tDP’s long-tail is higher than She-Ra’s pre-release hype.
I don’t have enough data to say why this could be; what is clear is that She-Ra skews younger. But it could also be the usual sexism of boys (and often, also their parents) dismissing girl-centric stories as something boys can’t relate to (and we’ll leave for some other time the unpacking of that issue).
Then again, unlike DW’s other productions, She-Ra is absolutely toy-driven, thanks to having Mattel in the mix. So it could be that there’ll be a huge spike in She-ra related merchandise that’ll warrant continuing the series, even if the viewership and other stats appear to be low. Some signal of that might show up when I compare google/bing stats, but that’s not a certainty.
Been watching the numbers. Looks like so far, S8 is on track to either be somewhere between S4 and S5 in terms of the damage. It has potential to end up worse than S5, but will have to see how it goes in the next week.
Yeah, DW clearly intended to bury the season. So far, looks like that’s at least one thing DW didn’t half-ass.
Something I don't understand. According to those stats you posted, VLD has got more or less similar traffic rates as other shows like Castlevania and The Dragon Prince. So why is its Twitter and Tumblr presence so massively higher? For example the trailer for S8 on Twitter has 12k likes, but the trailer of S2 of Castlevania has only got 3.6k, almost just 1/4 of Voltron. I know you said Fandom can't be used to detect actual traffic, but I thought different Fandoms could be compared to each other.
You can compare fandoms, yes, but what you’re comparing in twitter stats vs the stats I listed is the difference between engagement and viewership.
We can determine engagement by looking at traffic on twitter, tumblr, facebook, instagram, and google searches. As google is the only one of those five that has a reach beyond 20% of the internet-using public, I find it’s the best for a blunt-force measurement of audience engagement, which really boils down to “how much people are talking about it.”
Wikipedia’s page views are an extrapolation of viewership. That’s a massive segment of the population who don’t get on twitter, use facebook for family stuff, aren’t into instagram, and have never heard of tumblr (yes, lots of those people exist).
Remember, a controversial show can get people talking but that doesn’t translate automatically to them watching. If you go by twitter stats, S7 should’ve been THE most watched season of VLD ever, ‘cause boy was it blowing up the charts in terms of engagement. Thing is, its viewership stats were just fractionally higher than S6, and still nowhere near S1/S2 levels. It didn’t bring in that many more eyeballs; it just got talked about. A lot.
The other thing to keep in mind is that you’re comparing one show’s seventh season to another show’s first season and a third show’s second season. VLD has had 2 more years to build up its audience than tDP, and a year more than Castlevania (which also doesn’t come with the massive marketing push that DW puts into even its half-assed-marketed shows).
On top of that, Castlevania’s audience skews older, and those of us with jobs don’t have the time to spend all day on various platforms, so our engagement may be as strong but it’s rarely as noisy. (in other words, we’ll miss stuff ‘cause we don’t have the time to go back and see everything in our feed that happened while we were busy elsewhere.) Meanwhile, tDP’s audience skews slightly younger, and twitter is much more a 18-30 kind of platform. So tDP’s audience is either slightly too young for twitter, while its other audience (parents) are just a little older than twitter’s core audience.
So, comparison graphs. Colors for each show in each graph are:
blue: Voltron (VLD)green: She-Ra (SPOP)red: The Dragon Prince (tDP)orange: Castlevaniapurple: Trollhunters (TH)
Viewership stats (via Wiki)
Stats run from 6/1/2016 to 11/17/2018.
VLD’s S1 viewership peaked at about 38K; TH’s S1 peaked at about 21K, same as SPOP over on the far right. Castlevania hit about 72K in its first season, so its S2 drop to around 48K is about standard. tDP managed almost 50K in its first season.
Viewership drops by 30-40% with each subsequent season for the average show; a drop of less than 20% means you have a major hit on your hands. The last season almost always shows a spike, as people come back to see how it ends. If a show ends without that spike, consider it dying a quiet death.
That reality of each subsequent season losing viewers means SPOP is in a really bad place right now, unless it can do a turn-around like TH and have a powerful ending. If SPOP has more than three seasons, by the time it gets to its final season (and that resurgent viewership for the finale), it might barely be a blip on the scale. The longer the show, the higher you want the S1 levels, to offset that expected decline over the seasons.
Engagement stats (via Google)
Here’s the chatter, of people searching for each show (or related topics like toys, merch, news, fic, art, etc). Same colors as above, for each show, for the same time period.
Unfortunately, Castlevania drowns everyone else out by such a degree that we can’t really see much until we get that out of the way. So this second one, I narrowed it down to ‘Castlevania netflix’ to temporarily quiet it, a little.
Compare this to the viewership, and you can see some interesting behaviors. People talked about TH a lot more than VLD, despite TH having such lower viewership stats. There was a major spike when SPOP released its character designs, and kerfluffles get people talking. For VLD, its S3 viewership dropped as in standard but the chatter went up; S3 and S7 came close to matching S1 levels of engagement (for not always good reasons, natch).
Notably, tDP may be somewhat quiet (comparatively) over on twitter, but it’s got people talking, too. If you’re wondering, the most common related search terms are for the cast overall, then ‘avatar the last airbender’ and then Ralya, Callum, Amaya, and Claudia. Going by frequency, Rayla is the most popular character by a large margin.
The dotted line in green, at the far right, is Google’s prediction based on the past few days’ traffic, as we haven’t quite completed the first week. (A query for a year+ gets compressed to a weekly Sun-Sat view.)
Speaking of which, let’s take it down to just the past 90 days, and remove the narrowing filter on Castlevania. This graph has a bumpier view because it’s back to a daily value for each.
tDP and SPOP look very close, and there’s really only a few points difference: tDP maxed out at 43%; SPOP maxed at 37%. (Basically, take Castlevania’s highest-rated day as x, and tDP’s highpoint is 43% of x.) If I narrow Castlevania down again, we can see tDP vs SPOP chatter.
People are talking about SPOP but still not quite to the degree of engagement as tDP. Then again, SPOP also hamstrung itself by releasing three days early without a lot of warning, so it’s possible we may see another spike similar to tDP’s once word gets out that the series is actually released. Or, once early adopters create buzz and lure more people into watching.
bottom line
What people talk about isn’t always what they’re watching. And people will watch and not have the place, time, or energy to chat about it online. All of these shows are on Netflix, too, so no advertising income as a comparison point. That makes engagement a more useful data point, because the income doesn’t change based on viewership.
By that measure, TH is probably DW’s most successful series of these, followed by VLD, and then SPOP. TH is the only one inside shouting distance of what tDP managed it its first season, and none of them come close to competing with Castlevania’s numbers.
Hell, if I were DW, right now I might be having serious regrets of making VLD a Y7 property, considering Castlevania’s M rating. By platform and viewership, Netflix might be the best home for adult-oriented series where stories can explore darker themes. Of all the franchises DW has right now, VLD was possibly the best candidate for taking advantage of the platform. That M rating hasn’t hurt Castlevania in the least, and in fact may’ve been part of what propelled it so high.
So, I’ve had a few questions on my S8 data cronch part 1, curious as to why I say S8 is so bad, when the numbers seem kinda mostly okay. A lot of those asks will have the context I’m referring to once I post the next part. I figured I’d give this particular comment a longer response of its own, though.
In the one with the coloured dots, S8 is doing about as well as S2, which is widely considered one of the best. And even if 7th place is not as high as it got historically, it's still very high.
I'm guessing you mean this particular graph.
First, other than S4 (which didn’t chart at all), the only other season that debuted at only #9 was S3, and it only charted for its release week... and disappeared. By that measure, well, sure. S8 is doing somewhat better, since it’s held on for two weeks. We can’t really say how well it holds up to other seasons, in terms of longevity in demand expressions, until we’ve got a few more weeks to compare.
Here’s what this chart isn’t showing you: the actual values. When S2 came out in January 2017, A Series of Unfortunate Events sat at the top, with a demand expression value of 17.4M; VLD had 9.4M. Main in the High Castle was 3rd, with 11.3M, and Stranger Things was 6th with 9.7M. VLD came in above Luke Cage (8.2M) and Orange is the New Black (7.3M).
For S8, VLD debuted at 9th with 19.6M. Some of the series above it: Star Trek Discovery (20M), Daredevil (28M), Sabrina (29.3M), Narcos (31M), Marvelous Mrs Maisel (33.6M), and Stranger Things (47.6M). I honestly think Parrot’s been expanding and tweaking its algorithm to include additional platforms and a wider global net. It means you can only really compare within maybe a few months; a breadth of two years has too much difference in the base math to be worth more than as a series of snapshots.
Because if I map out the actual values, you can see there has to be some inflation going on, as Parrot shifts things around.
Most of the major platforms -- twitter, tumblr, instagram, facebook, and so on -- existed two years ago, and none of them have ballooned sufficiently in traffic to make me believe that a #9 two years ago only needed to be 7M while getting #9 now requires a score of almost 20M. But since I don’t know precisely how the company is calculating its so-called ‘demand expressions,’ I can’t say for sure the cause of the value inflation.
Is it due to better data collection? Or a broader net (are they including weibo now? or some other massive non-english platform)? Did they change up how they weight an expression on one platform, vs another? Are they now counting every mention instead of bundling counts? I have no way to know, and that makes the numbers highly suspect in terms of any standalone comparison.
So the one thing I can do is say that given the context of other shows’ popularity at the time of each season of VLD, this is where it stood in contrast. The other thing I can determine is how long the season remained as a topic of popular conversation. In that, S1 and S2 were outstanding, with 8 weeks’ and 6 weeks’ longevity, respectively. S6 returned VLD to that level of engagement, staying on the charts all the way through to S7′s release, 8 weeks’ total. S7 managed six weeks and fell off again.
Here’s the other thing you need to remember: like I said in my previous post, this doesn’t measure viewership. It really only measures chatter. S1 and S2 stayed in the public’s awareness, but the chatter gradually faded. S6 got people talking above the initial release; honestly, Shiro’s return fired up the fandom to a huge degree.
And S7... well, there’s the sticky part. This ‘demand expression’ doesn’t say what people are discussing, only that there’s a continued audience chatter about it... and you have to take into consideration how the fandom completely blew up VLD’s corner of the intarweebs. I mean, it was off the charts, and it was hugely angry. Add in the various news zines catching onto the controversy, and the simplest explanation is this:
People weren’t talking about VLD as a show they liked. They were talking about VLD as the centerpiece of a controversy that raged high enough and hot enough that it took nearly five weeks to burn out. This wasn’t S1, S2, or even S6 levels of excitement. It was rubbernecking: there’s little the internets love more than converging on a fandom mid-implosion.
I’ve known marketing and upper-level execs who will swear that this is just fine, because at least people are talking. I can also tell you they’re only fooling themselves, because a brand will take a massive reputation-hit when the talk is that overwhelmingly negative. It’s one thing if people are talking about you approvingly and recommending you to your friends; it’s another if people are actively plotting your immediate and messy demise. If an exec wants to see the latter as positive compared to the alternative of no talk at all, go ahead, but don’t drag the brand down with those delusions.
But this is also what makes S8 an interesting comparison case, because it also left a lot of the fandom swirling in anger... but after the heights of S7′s fire, a lot of people are burned out, pun semi-intended. There’s still chatter, as befits the final season of a long series, but once you look at it in context of everything else... it’s as negative as S7′s response. It’s just not as strident.
And that context is what I’ll be covering in the next data cronch part. That’s coming up as soon as I can compile it into something resembling an organized post.
