This is a transcript of SYS Podcast Episode 268: Judging Screenplays By Their Coverage, Data Analysis With Stephen Follows.
Ashley: Welcome to Episode #268 of the Selling Your Screenplay Podcast. I’m Ashley Scott Myers, screenwriter and blogger over at www.selingyourscreenplay.com. Today I’m interviewing producer Stephen Follows. He just did a big report in conjunction with The Script Lab where they analyzed around 12,000 screenplays that came through the various contests. And then they generated a bunch of stats based on the reader critiques that these various screenplays received. We talk through this mountain of data and try and find some things that are useful to screenwriters. So stay tuned for that interview. If you find this episode viable please help me out by giving me a review in iTunes or leaving me a comment on YouTube or retweeting the podcast on Twitter or liking or sharing it on Facebook. These social media shares really do help spread word about the podcast so they’re very much appreciated.
Any websites or links that I mention in the podcast can be found on my blog in the show notes. I also publish a transcript with every episode in case you’d rather read the show or look at something later on. You can find all the podcast show notes at www.sellingyourscreenplay.com/podcast, and then just look for Episode Number #268. If you want my free guide- How to Sell a Screenplay in Five Weeks you can pick that up by going to www.sellingyourscreenplay.com/guide. It’s completely free, you just put in your email address and I’ll send you a new lesson once per week for five weeks along with a bunch of bonus lessons. I teach the whole process of how to sell your screenplay in that guide. I’ll teach you how to write a professional log line and query letter and how to find agents, managers and producers who are looking for material. Really it’s everything you need to know to sell your screenplay. Just go to www.sellingyourscreenplay.com/guide.
So now let’s get into the main segment. Today I’m interviewing producer and statistician Stephen Follows. Here is the interview.
Ashley: Welcome Stephen to the Selling Your Screenplay podcast. I really appreciate you coming on the show with me today.
Stephen: That’s all right. Thanks for inviting me on.
Ashley: So to start out maybe you can give us a little overview of your background. Where did you come from and how did you get into what you’re into now?
Stephen: Yeah, sure. So I’m British and I’m based in London. My background is I’m a producer primarily and then also a writer and made a load of short films, sort of hundred on short films of different types and sizes and things and built that into a production company, which I still run with my business partner Ed from Somerset House in London, which is a beautiful building rebuilt in 1776. So one of those sort of very British iconic kind of things. Yeah, so that’s my main living. Then maybe say six, seven years ago I started to publish stuff I would just do for friends where we would be having a conversation about producing or writing and someone would assert some industry adage about this never works, that always works, never do this.
And I would always go away and have a look and wanna see if that was true or not, and sometimes you can find out sometimes you can’t. So I started to share this research I was doing and it’s kind of grown from there. I now publish every single week and I publish all sorts of research first hand. Most of it is first hand new primary research on different aspects of the film industry, all trying to help everyone, from aspiring and up and coming people to established professionals trying to understand how the industry works, what the trends are, and trying to democratize that information. Because if you work for a studio, if you have unlimited money, there are plenty of places you can go and people you can hire to tell you how things work.
But if the vast, vast majority of us, we have to figure out as we go and if you’re figuring out how to make a movie when you’re making your first movie, that’s a very bad way of learning. That’s like learning to swim by being in the middle of the ocean.
Ashley: Yes, you’re right.
Stephen: You can’t… it takes you two years and a hundred thousand dollars of someone else’s money just to go, “Oh, yeah, I shouldn’t do that.” And if everyone else does the same journey then that’s mad, we should be sharing that information in advance. So that’s my background. Then I always look for, kind of, new areas that I can research and things that other people haven’t looked at or new data sets and that kind of open up the possibility that we as a community can know what’s a fact, know what’s an opinion and know where creativity comes in. Yeah, so I try and do as much of the different kinda stuff as possible.
Ashley: So you mentioned the producing, the writing, do you also have a background in sort of statistics and math? Like where does this stuff come in into your background?
Stephen: No. The last qualification I had in maths was probably when I was 16, I don’t know. I just… I have a head for numbers and I can see systems where other people just see chaos.And I enjoy it, and so I feel like partly it’s a joy and partly it’s a responsibility to be the person that understands the film industry and understands how to analyze and quantify things. And so I very much do it as a hobby. When I was a kid, I knew I would either be… either do sort of something proper and then film on the sides or do film and then do some academic things on the side and I went for the latter and I’m glad I have. So film and creativity is my day job and rigor is my hobby.
Ashley: Yeah. So let’s dig into this new report that you did. I guess you guys analyzed about 12,000 screenplays and the covers that went along with those. Maybe you can just give us a quick, you know, 15, 20 second… I think you sort of covered some of that, but maybe we can dig in just some of the specifics of this report. What is this new report all about? And then we’ll start to go over some of the lessons that you’ve learned from all this data.
Stephen: Yeah, yeah. This was really fun actually. There was an infographic years ago of a script reader who’d read three or four hundred scripts and he or she had summarized the key problems they saw in the scripts and where the characters are from. And John Screencraft and I both remembered it and we’re chatting about a year or so ago, just over a year ago. We said, “Wouldn’t it be cool to do something on a bigger scale?” And they’ve got a huge amount of scripts that have been submitted and coverage that’s been created for competitions and reports. So we figured out a way to keep privacy as not an issue because, you know, I don’t want everyone’s email address, but at the same time, be able to analyze all the scripts and all the scores that the readers gave them and look for correlations.
Ashley: Okay, so let’s dig into this a little bit. So maybe, I guess the first question would be like, what are some of the findings? Are there certain genres that readers like more? Are there things like page lengths within this certain parameters get… tend to get better coverage? Maybe you can start to talk about some of the results of all of this analysis.
Stephen: Yeah, sure. I mean, I think the first thing to say is that what we’re measuring here is what script readers think. And so this is not measuring quality in an objective sense, because we can’t know if the script readers are biased, if they’re wrong. Really, we’re not measuring art, we’re measuring the gatekeeper nature of script readers. You know, if you’re trying to get your script read by a big producer, you’re trying to get ahead in the competition or provide get evidence to show that your script is good, you have to interact with readers and that’s what we’re measuring. So we did find that there are correlations between certain genres like Animated and Thriller did tend to get higher scores than Comedy and Fantasy. But that’s not the sort of focus on that we’re really going for, because that’s interesting, but it’s not useful.
What we would constantly trying to do is think, “Okay, how can this be useful to screenwriters?” We’re never gonna create a formula for, “Here’s how you write a good script,” because hard work and talent is what you need. But can we give writers clues of where to look in their second and third draft? Can we say to them, “Look, it seems to be within the genre you’re writing there are certain things that are more important than others. Go back and check whether you’ve put time into those things.”
Ashley: And so what’s an actual example? Can you actually give us like you were talking about take a specific genre, Thrillers or something.
Stephen: Sure.
Ashley: What are some of those lessons that you were actually able to pull out? What would you recommend to someone who has a Thriller script?
Stephen: Yeah, that’s a great example. So as well as the overall score that the readers gave the script, we also had scores for a load of different smaller things like tone, characterization, pacing, dialogue, format, things like that.
Ashley: And what are those scores? Are they like based on one to 10, is it just a good, pass, consider, recommend?
Stephen: Yes. Yeah, no. It’s a one to 10 and we normalize them so that if there was a particular reader who was always a bit more generous than the others, we would lower those scores and so, we did that so that they are consistent data set. Then we’re looking at how they relate to the main score. So if you spend your time rewriting and focusing on a particular one of these elements, is it the element that’s most likely to move the overall score? What we found for thrillers for example, the plot was the most important or most connected to the overall score, meaning that it’s really important that you impress them with how they think of the plot. Whereas if you look at the other end of the spectrum for Thrillers, things like theme, the concept and originality were quite low and it doesn’t mean they’re not important.
All of those things still do correlate. none of them are irrelevant. It’s just that they were less important when it comes to the overall score. And if you compare that with something like a family film where catharsis was the number one thing, and for family films plot is right in the middle. It’s quite far down the list. What’s really fun about this is, this is exactly what you kind of expect as a film fan. If you’re watching a thriller you do want it to be a twisty, turny plot with reveals and things like that. But if you’re watching a family film you want there to be a resolution. You can’t have the open ended kind of who knows what will happen that you have in perhaps a drama, where there is a story but fundamentally it’s about a question of humanity and it’s never answered.
That doesn’t work in family. Family films are about, here’s a problem, here’s a solution and everyone’s happy at the end because you know you can’t… I mean, I’ll give you an example actually. There’s a fun video that went viral years ago where for Christmas, I think Toy Story three had just come out, and a load of boys or brothers had cut to out the last bit of the film so that they made it so that when all the toys are about to go into the incinerator, and it looks absolutely certain that they’re gonna die, they just made it cut straight to the credits implying that they just died. Right? Then they showed it to their mom and they secretly filmed her watching this. So she’s watching this scene and then the credits roll and she’s like, “What! What! They’re all dead?” And like a whole world fell apart.
Then obviously they reveal it and it’s very funny. But the reason it’s stuck in my mind is because it’s not just a choice. That would be fundamentally wrong because there would be no catharsis. Whereas we’ve all seen dramas where even in say action films like… without wanting to spoil it for anyone who hasn’t seen various films, there have been very recent big budget temples that have ended with bad guys winning and half of all the good guys dying. And like, that’s okay, as in it’s not fun, but it’s great entertainment. Family films, no way Jose. And we can sort of show that with these correlations.
Ashley: So maybe we can run through some of these other genres, because I’d be curious just to kinda hear your results. And again, it seems like some of this is very intuitive, like as the example with drama, I’m assuming that the plot is not as important but the characters are very, very important.
Stephen: Exactly. You’re absolutely right. Characterization and the voice, the writer’s authorial voice is really important with drama and plot is down the list. But you hit the nail on the head there where actually this science, this sort of… this analysis is not to tell you that you’ve been doing it wrong. This is not to tell you that the world is upside down, and you didn’t realize it. We know this. Good writers spend time on their art. They listen to people, they have film fans, themselves. This is just providing evidence, and there maybe a couple of places where we’ve all succumbed to the industry folklore that actually is wrong in one or two places. So this isn’t about saying, “We’ve discovered that characters are important in dramas and plot’s important in thrillers.” It’s to say, “Look, this is how important and this is a reminder.”
So we can sort of check all of this with our own experiences. In answer to your question, a couple of other examples. You’ve got characterization being number one for dramas and for comedies, which makes sense. The actual events they go through are far less important than who they are and what we see in them. Historical films, it’s about tone. It’s about the tone of the piece. I don’t think there’s any other genres where tone was number one and so that’s kind of interesting. They are kind of more about the pattern and the painting. If you look at Mr. Turner for example, there is a film with a very loose plot, but it’s a wonderful snapshot of a time in someone’s life. And you wouldn’t come out saying, “That plot point was good. That plot point was good,” you’d say almost like a Turner painting, “The landscape that they painted was really important.” So that’s kind of interesting and relevant. And so a lot of them…
Ashley: Yeah. What about… I’ve heard something especially people trying to break in, they’re always writing low budget horror scripts. What finds did you find for horror?
Stephen: Yeah. Horror actually has an overlap to some degree with a bit of drama where characterization is top. The voice is quite important as well, which is kind of interesting. One of the things that I’ve… I’ve studied horror a lot as a genre in the past and one of the things I found is that almost every other genre has a very strong correlation between what critics and audiences think of the film and how likely it is to make a profit. So if critics think a drama is bad it’s far less likely to make money. It’s just there’s a correlation there. But with horror, it’s almost irrelevant. So what’s interesting with Horror is the quality of the art when it comes to the finished film is not very important, and there are other things that are. Whereas with drama and documentaries they’re very, very connected.
You can’t make a bad drama and expect it to make money. There may be one out of however many but with drama you’re focusing on the quality bit. And so that’s those correlations that kinda makes sense because with horror, a good one can break out but ultimately a really good concept, like The Purge. The Purge has not got very high scores from audiences or critics, but it’s done staggeringly well, because it’s such a good idea. It’s an idea that we can all engage. It’s like a Brewster’s millions idea where you go, “Yes, I get that. That makes perfect sense.” One thing that was interesting was that the least important factor of all the ones we looked at across all genres was formatting. No case was irrelevant. In all cases it did have an effect.
But the way I read that is that formatting is important. If you break the traditional formats thing that you’re supposed to do you can lose marks, but there’s no way of gaining marks if you see what I mean. Like get it perfectly right and then forget about it and work on what really matters. I think there is an obsession in the industry with formatting. It is important, but there’s also… it’s like wearing clothes or washing. Yeah do it, but that’s not what’s gonna get you a date, that’s not gonna get you that job interview, you know, the basic bar. But what you should be focusing on is what’s specific to your genre and what your audiences are expecting when they’re buying a ticket to your kind of movie.
Ashley: Yeah. What about action films? What findings did you find on that?
Stephen: Well action is very plot orientated as well; plot and characterization. Interestingly, conflict is not as important in action films as you’d think. In fact, conflict was more relevant in animated films than it was in action films, but more so than in comedies. I think what’s interesting about this is that it’s confirming what we already know about the genres. And there were other things that we found that confirm our expectations for genres. So for example, one of the things that we did was that we looked at the sentiment of the scripts. So building on a lot of brilliant research done by scientists and researchers in the past, we were able to take a sentence and run it through a sentiment engine, and it comes out with a number between minus one and one.
So if the sentence is “FU I hate you”, then maybe that’s a minus one and if it’s, “I love you”, then perhaps it’s a one. And if it was, “hello”, it might be a zero, right? And so if you do that for every line in every script, and then you look at the journey of the script, you can get a rough trace of the plot. Obviously, there are always some scripts that might break the mold and if there’s a script called Opposite Land, we wouldn’t know that. But when you’re looking at 12,000 scripts, you kind of get the sense of it. And if you then average the sentiment for an entire… for the whole script, you get a sense of whether this is broadly a happy or sad film. Then when we correlated it with scores, we found something really interesting which again makes sense. Which is, for all but one genre. The happier the script, the worse it performed. And so…
Ashley: Really?
Stephen: Yeah. It was just fascinating.
Ashley: What was the one genre that defied that?
Stephen: Well, what do you think it was? So the one genre where happy…
Ashley: Like family because you told us the families you need that [Inaudible 00:18:08].
Stephen: Yeah, well, yeah. Family, the difference was very, very minor. It’s comedy where it’s significant, like, statistically significant but it’s happier. This is interesting because when we think of it as perhaps as writers, but definitely as film fans, we always think about the ending defining the movie. And if you think about an action film, you think about the bad guy lost and got killed, and whatever and you think of that as the resolution. But when you look at the journey of the script there across the hundred pages, let’s say, it’s probably only the last five or 10 pages where the good guy wins, or heroine wins or whatever. The majority of the script is about conflict and about losing, obviously not in a straight line, that’s quite boring. This was most acute in Dramas and Thrillers.
So if your drama or your thriller for the majority of the script is fairly happy, then actually that’s probably that’s statistically not gonna help. It doesn’t mean make every situation incredibly sad but it is worth thinking about that. That if you have conflicts that are too easily solved by your hero or heroines, and if you have moments without conflict, then of course, it’s not gonna be nearly as successful.
Ashley: Yeah. And so that’s… and I guess that when you first started talking about this I was a little bit confused. So you’re not talking about the resolution of the movie at all. You’re talking about the journey of the whole thing. So when you say a script is on the happier side. It has nothing to do with the last five minutes where the hero… because almost every movie, the hero wins in the end, but that doesn’t necessarily mean happy.
Stephen: Also, one interesting thing about the end of a movie and catharsis is that actually it can be really hard to tell. So you’re right. The research we did here is about the average of the whole script. So across 100 pages, if 90 were sad and 10 we’re happy, that’s 90% sad which is very sad. But when you think about resolution and catharsis, it can get really complicated. I was involved with a BBC Radio Show years ago with Mark Kermode who’s a big film reviewer. We were talking about what we could study for the for the radio show. He said, “I’ve always wanted to know whether films with happy endings make more money than films with sad endings.” Now we weren’t able to do the research and the reason is, what is a happy ending?
And the example that kind of really proved it for us was Gladiator. So throughout Gladiator, he starts at the beginning a well-respected, powerful general with a wife and kids. Everything is stripped from him physically, metaphysically, his wife and kid are killed. But because of the very beginning we see glimpses of the Elysian Fields they have in his culture. And at the end of the movie, when he finally gets to kill the guy who caused all of this for him, and then he dies, our hero dies, he gets to go to heaven and see his wife. So is that a happy ending? He lost everything and then he died, but in the context of the movie because of the way that they shot and edited it, actually it feels like catharsis and he’s finally with his wife and kid again.
I couldn’t tell you… I’m sure people can write pieces on this, but is that a happy ending? It certainly feels… because you get a sense of catharsis, but no one would say, “Oh great. He died.” So it’s interesting, whether we’re looking about the journey or the destination.
Ashley: Is there a way for us to run these reports, this happier report? Is there a tool we can upload our script and you know, and then it will spit out a score for us?
Stephen: The sure answer is no. We did think about creating something like that. But my concern is that this is advice that works across all scripts on average, and every individual script is different. You have to be the thoughtful writer who takes all of this as advice but then chooses what to do individually. I think you could run any script through this but you’d find that it would bounce around and it might well be that there’s a bit that is quite sad that is actually quite important to this, or quite happy that is quite important to the script because it shows the characters’ dreams and hopes that are later crushed or vice versa. No program could, at the current state of AI, could understand that and so we don’t want people to go too far into thinking this is a formula or that they can cheat the system.
This is research to prove what we already know, to look for correlations and then sort of hand it over to the screenwriter and say, “There you go. Go do your craft and take this advice as an advisory, but you can choose what you want to do.” We found that in a few places that the advice was actually quite measured. So you talked earlier on about the number of pages which is always a concern that everybody has, writers and producers and certainly there’s an old producers joke when someone hands you a big script you pick it up and hold it for a second and say, “No, no, I’m sorry, I can’t do this, it feels expensive.” And what you mean is too many pages. We were really interested to see if there was this sweet spot, this absolutely perfect place that everyone should hit the way they would on a word count for an essay assignment.
And the real truth is actually, it’s quite simple. Don’t make it very long, don’t make it very short. Beyond that make it good, which is not very practical advice, but is it the truth? There is not a magic number, it’s that you wanna be between 90 and 120 pages roughly, and it falls precipitously on either end. If you make a film, write a script under about 85 pages it tends to get quite a poor score and over 130 tends to get quite poor score as well, so that’s kind of interesting, that you want to try and shoot for the middle in the technical sense and then shoot for excellence within the content and within your artistic sense.
Ashley: The one thing… and as you’re talking about the page length thing I think is, it’s just something that comes up, especially a lot of newer writers, they want this sort of template or formula on how to write a script. So they’re the ones that are concerned with the formatting and the page length and those kinds of things. So I get all that. The one thing that I would just say, and I wonder how you guys have sort of thought this through. Your data set of 12,000 scripts, it’s distinctly different in that those people, while they may be professional readers, they’re basically being paid to read from page one to page 120, or whatever. And I think that’s not necessarily what’s happening when you’re submitting.
I mean, if Steve, you submit a 200 page script to a production company, even if it’s the most brilliant thing ever written, it may not even get read. But I guess my bigger point is there are gonna be some differences. Your data set is different than what is more typical of a screenwriter submitting their things. So how do you account for those potential differences?
Stephen: You’re absolutely right. I think that’s why it was so interesting that it was so important that everyone understands that this is analysis of this data set within the context of these kinds of things. The majority of the scripts that we were analyzing were submitted to open screenwriting competitions. What that means is the majority of the scripts are by amateur writers and that’s not a slow on their abilities, that’s just to say, they’re not professional, they’re not being paid as their primary income. But it does also include some very experienced writers, some Hollywood stars and obviously some winners of the competitions because by definition someone has to win. Some of the scripts are very good, but the majority were amateur scripts.
So I would suggest that the work we found here is no most use if you’re really focusing on script competitions. It’s not exclusively that. There were also script reports, but you’re right, every studio and every producer will have a different focus on what they want from a script report. If they literally just want someone to sort the wheat from the chaff then reading 10 pages and dropping it is perfectly within their boundaries of their rules. If they’re looking for a script report or it’s in a competition, they will read all of it. And I think that this isn’t a panacea. This can’t solve every problem you’re gonna have in the industry but this is a quite deep detailed study of one aspect of it.
I think that, what you said earlier on is very true as well, that early writers often want the secret sauce, they want the formula and I know that there’s sort of a weariness of some very experienced writers that say, “Look, just work hard on your craft. Write. Write. Write.” and I back that up. I’d say that I know [laughs] I know that’s doesn’t feel like helpful advice. But there are things we can do like with this data set and you can use it to understand the challenges maybe facing you and there may be illustrations and proof of things you half knew but now you kinda know. But ultimately it’s gonna come down to you working very hard and getting feedback and still writing, finishing scripts. All the things that everybody else says.
There was a few little things in there that surprised me that… they’re not major findings but they certainly… I mean, there was one for me that always that when I started the whole project about a year ago, I really wanted to find the answer to. Not because it’s very consequential, but just because it was something that bugged me for a long time and I didn’t know the answer. So in the industry there are many people who will tell you that voiceover is bad, and very rarely do they go into more detail. But when they do, they say it’s a literary form, the internal monologue of a character is not what cinema is. Cinema is about dialogue and action and inaction, but it’s visual. It’s show, don’t tell.
If you’ve got voiceover, then you are writing a novel and you haven’t embraced the formula… no you haven’t embraced the format. The counter argument to that is usually equally short, which is usually just the words Goodfellas.
[laughter]
Stephen: And neither argument is particularly sound. Neither is wrong, but neither is particularly kind of conclusive and it just comes down to whether you buy it or not. So I was really keen to go in and have a look at the correlation between voiceover and the quality of scripts. And what I found was no correlation at all. So it doesn’t matter the amount of voiceover you have on average. Obviously how you use it is incredibly consequential, but actually the amount of voiceover isn’t a bad thing. And so that was pleasing in itself to know that, “Okay, we have we have an answer to that.”
But it also had this sort of little follow up of, “Well, why do people think it?” So I’ve had to think about that and I’ve evolved my thinking now. My belief now is that there is still a correlation between voiceover and bad movies. Because I think that it does exist, just I have no evidence for that beyond my own experience watching movies, but I now put the blame or at least the cause on editors who are either working under the instruction of producers or trying to save a film, or cover up plot holes, or you know, all that sort of ADR kind of stuff. So voiceover in a movie might be correlated with bad movies, but voiceover is in a script is not correlated with bad scripts, which is kind of pleasing, isn’t it?
Ashley: Yeah, that is interesting. And the voiceover thing is something that’s bugged me for a long time. I’m in a writers group and basically, there’s 15 other writers sitting in the audience. One writer has actors on stage reading their scripts, And I’ve put up scripts that have voiceover, other people in the group have put up scripts with voiceover, and every single time the voiceover just gets pounded. The writers go, “You don’t need to voiceover it. Show don’t tell.” and so I just… it’s, part of me doesn’t quite believe your findings, you know.
[laughter]
Ashley: Because I see it with my own eyes that just for some reason people have this sort of, it’s in their head that voiceover is hacky even in the script level. And I just see it every week in and week out. If someone puts up a script and there’s a lot of voiceover they’re gonna get hammered.
Stephen: Well you’re right… and that actually relates back to something you said before, which is that there are many gatekeepers. Let’s say that because this is an empirical fact that within our data set voiceover was not correlated with quality at all. They’re unconnected. However, if you’re sending you’re script to a producer who believes it is connected then maybe it does matter. Because he or she, this producer might throw your script out. Even though they’re wrong they might throw it out, which means it does matter. So it’s one of those weird things where I’m… we’re right empirically. But in the real world, it still comes down to what people believe. And if enough people believe it, it becomes a practical truth.
As in if every producer is gonna throw your script out because it’s got too much voiceover even if they’re all wrong, if you wanna get picked up, you’ve gotta take some voiceover out. So it doesn’t end the entire debate, but it definitely confirms that it’s not automatically a bad thing in any kind of logical sense.
Ashley: Yeah. And so was this analysis on the voiceover, was it sort of like a sliding scale where you tried to basically see if more voiceover lower overall score or was it just sort of a binary thing more if you have voiceover don’t have voiceover.
Stephen: No. Good question. What you’re looking for is correlations. So there’s a thing called the Pearson coefficient correlation correlation coefficient. And it basically is a way of saying, if you’ve got two sets of numbers, how are they correlated? If column A goes up, column B always goes up, you end up with a correlation of one which is entirely positively correlated. And if whenever one goes down, and the other goes up, then it’s minus one, and if it’s zero it means they’re unconnected and you need to get above not point two or below minus two to actually say, “Okay, there is something going on here.” Because with any large data set you’re always gonna have lots of noise and you’re gonna have… if you look at enough sets of data you’ll find some that correlate, even if it’s erroneously.
So using those methods we didn’t see any correlation like that. Now obviously we’re not using AI to understand how you’re using voiceover, so it could well be that half the people are using it brilliantly and half the people using a terribly and on average it looks like it doesn’t work. We don’t know. That’s too complex kind of question for this kind of data set. But I can tell you that it’s not automatically a bad thing, which is nice to know.
Ashley: Yeah, yeah. Were there any other specific things like voiceover that you guys really dug into, other just interesting little tidbits?
Stephen: Yeah. So we ended up doing kind of two different things that they were the same day set at the same time. One was this headline thing where we’re looking for correlations between certain types of things and success. And the other was looking at what the average screenplay looks like. So as for more correlations around success, you probably need to have a look at the report because there are some complicated we could use sentiment to work out plot arcs and look at whether they were successful or not. That actually is quite a complex topic that needs the graphs because you can see the whether the sentiment goes up towards the end or towards the middle.
And so there’s a few more things like that, but we are probably a bit too complicated to talk about just as audio. But in the second half of the things we were doing, where we were looking at the average screenplay, you know, what’s normal. So this isn’t even just saying whether something’s good or bad, but it might give you some just some advice about what’s average for your genre. For example, the shortest scripts were horror scripts and animated scripts, which you often find with the movies, and the longest were historical and faith based films. Then you also look at how many characters are in movies, and so thrillers and horrors tend to have a smaller number of speaking characters 25 to 30, whereas animated and historical films tend to have sort of 40, 45 characters.
So we did things like that where we would look to how much whether scenes were set, whether they…. westerns are mostly set outside and comedies are mostly set inside. We looked at how long dialogue was and we looked at whether… horror, again, horror and thrillers happened more at night, family films tend to happen during the day, which is kind of interesting. We also did a whole load of work on swearing, which I’m not gonna go into, just because this podcast will suddenly get explicit very quickly. But we… so there’s no need to have an explicit one again. We looked at the three worst swear words, and we looked at which ones they were used in, and how and how frequently. One of the things that was really fun to look at was the…
Ashley: Was there correlation on swear words and the reaction to the readers?
Stephen: Yeah. There actually was. This is one of those really interesting things that I am happy to talk about and it’s in the report. But I worry that one or two people will take the wrong end and will change their scripts because of it and let me explain. So there is actually a correlation between swearing and the quality of the script according to readers. So if you… we added that we create a swearing score where in each of the three words, words got different points depending on how bad the words are, more details in the report, but it allowed us to have a score for how sweary the script was overall. And for as this sweary score goes up, as the film has more of these words, the better the review score was right up until the various swearier scripts, the ones that use it a huge amount and then the quality drops again.
So generally more swearing equals higher scores and what worries me is that somebody will open up their screenplay and drop a load of F bombs and think they’ve improved their script. They’ve probably done the reverse and they’ve definitely not improved it. But when we were trying to work out, “Okay, what’s going on here? Why…” because clearly you can’t just sprinkle some swear words in and improve a script. That doesn’t make any sense. But then we found that there was a very strong correlation between swearing and the scores that readers gave for the category called Voice. You often hear about producers wanting to find writers with a really strong voice and they write a spec script, not because it’d be picked up but because people will be able to read your voice.
And it seems that readers have a belief there’s a correlation between swearing and voice. It might well be that they’re right and that they’re Tarantinoesque kind of swearing is poetry, and it might be that they are wrong and they just like swearing and put that down as, “Wow, this person is a really brave thinker, really this is the voice of my generation.” We can’t tell…
Ashley: Was there any division in terms of the swear words and swear words that are like in dialogue that the audience is ultimately gonna hear, or the swear words that are in the character or just the subscription the action lines that the audience will never actually know? But those to me are more about voice.
Stephen: You know what, that is an incredibly annoying question because it did not occur to me and I wish I’d done it [laughs]. That’s a great question. We were looking at swearing across the whole script. And you’re right. What would’ve been fun was to look at there’s any swearing in the screen descriptions. Although, I was about to say sometimes you have… like there’s a place in the UK called Scunthorpe and within that is a four letter word a show won’t say. There are other things like that you don’t want to… that one of my co-authors surname is Cockcroft and so within that you might get swear words. We were looking at standalone swear words but we weren’t distinguishing between whether they were in the action or the dialogue and you’re right, maybe there was a difference between that.
I would’ve thought my [Inaudible00:37:39] guess would just be that almost all the swearing is in the dialogue. I think it’s a pretty brave writer to…
Ashley: Well, I’m just… Yeah and I’m trying to think… the guys like Shane Black and Quinn Tarantino, I’m pretty sure they’ll use a swear word or two in their descriptions.
Stephen: Yeah, exactly. And maybe that’s it. I mean that’s really interesting, actually. As I said, I wish I’d looked at that. Great question.
Ashley: I’m curious. It was probably a year ago and I’m sure this is something you and John talked about in working on this report. I think it was The Blacklist, and they were actually selling a service, and they got a lot of blowback for it. But they were selling a service that could kind of run these sort of statistical reports on your script and they eventually distanced themselves from. Did you ever get a look at what they were doing and kind of maybe try and improve on that? Was there anything you thought through on that stuff?
Stephen: That’s pretty interesting. So I don’t have any inside knowledge on that one. I was a viewer of that as well. I had the same journey that a lot of people did, which is I was conflicted, because what they’re doing is very reasonable science and if it helps people then there’s no harm. They weren’t forcing anyone to use it. But on the other hand, there is a danger in trying to suggest that a rudimentary, even if it’s a bit complicated, but what it’s fundamentally doing is not understanding the script in any kind of artistic sense to then tell you what to do with your script. That is probably not gonna be very successful until AI gets incredibly smarter. And even then, maybe, maybe not.
So I don’t wanna talk about that individual case because I don’t know enough about it. But there are other services that are doing that. And quite often the template seems to be to me that there’s a PhD student who has learned how to do all this sort of stuff who then thinks, “Great. Well, this is a tool that can help writers.” And they mean to help writers, there’s no there’s no malice there. And they build a tool, and I’m not gonna name names, but there are a few services out there who are either launch or prelaunch who I’ve talked to and seen, and they offer all manner of graphs and charts from your script. But the problem is, I’m not convinced that that works on an individual script basis.
Because if you looked at… we didn’t do this, but if you look at a Tarantino script, it might fail by some average measures of how it compares as an average script because it’s quite different. What we’re looking at is the average script and actually you don’t want your script to be average. So this kind of research is very good on mass but individually is quite tricky to be able to tell. Also some of them also fail tests that humans would be able to see pretty quickly. So one of the services that I was looking at because I was chatting to them about it, that uses this kind of approach, they had used Deadpool as an example of a script. They’d put it into their system, analyzed it, created all manner of different graphs and things.
And one of the things that they did was that there was a graph that showed you when each of the characters showed up in the script. The idea was, you could see when your character was absent, or a secondary character was absent for the second act or the beginning of the second act and you could say, “Okay, I can go back and fill that in.” That’s great in theory, but the problem with Deadpool was that the same character was sometimes referred to as Wade and sometimes as Deadpool. And so this analysis said that Wade is a character that disappears and that’s weird and you shouldn’t do that. But anybody reading the script or watching the movie knows that it’s the same character. But how can a software possibly know that without unbelievable levels of understanding and logic and intelligence?
So the thing is that whenever you go too far into that much detail, you’re kind of missing the point, which is, write good work, get feedback, keep writing, rewrite, meet people, talk to people. That’s where you should be putting your energy. We’re not at a stage in 2019 where any computer program on an individual script basis can give you anywhere near the advice that a standard script reader or editor or producer could do in 10 minutes. So that’s my feeling.
Ashley: Yeah, and I’m going back to what you said about this measure of whether a script is a happy script or not. And as you were describing that, and I’m sure as a producer you’ve seen these scripts. So often you see amateur scripts where there’s virtually no conflict. My guess is those would fail, this happier score and I understand what you’re getting at. You don’t want people running it through a tool and having sort of a knee jerk reaction to this. But I can also see where a lot of amateur writers, they could run their script through something like this and they could see very quickly and inexpensively instead of spending $100, $200 on a script reader, there could be some of these rudimentary tools that just say, “Listen dude, this thing is way too happy.”
And really, what that means is you don’t have enough conflict. Because I’ve read those scripts, and they just don’t work, like drama is conflict. And I’m just gonna play devil’s advocate.
Stephen: No, you’re absolutely right. There might be things like that on a case by case basis. And certainly, I’ll give you an example we can both agree on, which is formatting. You should be able to run your script through an analyzer that tells you if it’s a script formatted correctly because that’s pretty obvious, right? And that would be very helpful, the same way that we use spell checks. Spell checks are invaluable and yet they’re doing exactly what you’re talking about and they are really useful. I think the tricky line is always where the art starts to take over. Spell checking is not fundamentally a creative endeavor. It’s logical. You either correct or you’re wrong. Whereas whether the character is nice or mean to their children in the beginning of the movie is subjective artistic discussion.
So that’s a [Inaudible00:43:20], there’s no hard line that we can all agree a spell check is good, we can all agree that you can’t just have a “yes this will make $10 million dollars” just from an AI reading it. But where in the middle we draw the line comes down to the writer, the problems they have, how smart the tools are. But remember also that the other factor is that the industry is hard to get into and sometimes it’s for the wrong reasons, but that actually can have a positive effect. So let’s say that we would take your example of a writer who’s an early stage writer, and you’re exactly right. They’ve written, what a lot of people do, a film without conflict, they run it through a program that accurately tells them that and it comes back with a sort of flag saying, “We think that your film has not enough conflict.”
Would they then know what to do with it? Arguably, would it not be better for them to have found the same result by going to a writers group, by getting some feedback from somebody else, by maybe paying for script report, although I obviously appreciate that a hundred bucks can be a lot of money when you’re trying to make a living as a writer. But there are other ways of doing it. And perhaps those other routes right now would provide more context, more understanding and more of a solution than perhaps a rudimentary machine. But it all depends on the writer, the script and the program, I guess. And I’m not against them. I don’t believe that they can never help. I just haven’t yet seen one that can work on an individual basis script by script and that is anywhere close to what you’d get if you asked a fellow writer or a writers group, you know?
Ashley: Yeah, no doubt. And that’s a fair point. I mean, maybe if you’re smart enough to understand what it means to not have conflict in your script, you probably don’t need to run that report.
Stephen: [laughs] And that’s exactly it. There aren’t many shortcuts in the film industry and the ones that do exist involve being pretty, rich or connected and maybe young. These are not things you can cheat really. And so any shortcut is gonna help you in the short term, but maybe hurt you in the long term. If you do get the job because your uncle gave you the job, maybe you’re not gonna have learned the lessons that other people would learn if they went in from the ground up. I’m not suggesting that everyone should have a hard time but if you found a way like that, that told you there wasn’t enough conflict, just as you said, maybe you wouldn’t know what to do with it. Because do you really know what conflict is and is this the way to find out what’s wrong with your script?
Ashley: Yeah. So what’s next for you Stephen? What are you working on now?
Stephen: Well, every week I try and do a new blog post, so I’m working on… and I have a few longer term projects on the go. One thing that will be probably coming out in a few months, I’m not sure exactly when, is that I’m working with a film financier who’s given me a lot of budgets and recruitment schedules for real films and I’m sort of working out with them what I can publish and what they wanna keep back, because quite often I verify the data but then anonymize what the films are so that we can all learn without people feeling like they’ve broken things. I’m doing some studies on cinemas, you know, the actual buildings and how many there are, what kind of movies they show.
Then every week I try and think of a topic that I haven’t looked at. Actually I would say that amongst my… most of the best articles I’ve ever done, all the most useful research have come from readers’ questions and from people who’ve said, “You know what, I’ve always wanted to know…” or, “Here’s this thing in the news, is it true?” Or, “Someone said that to me.” and so I would actively encourage your readers to throw any and all questions at me. If you go to www.stephenfollows.com S-T-E-P-H-E-N F-O-L-L-O-W-S, you can go to the contact page, you can drop me a line and it gets through to me. I don’t have an office, and I’m happy to, if I know the answer to your question, I can point you to existing research, mine or someone else’s I will.
If we don’t know, I’ll add it to my list of things to do. And as I said sometimes the best questions come from readers. Just before Christmas last year, I did a whole probably too large study into the question of whether Diehard is a Christmas movie or not. It was an absolutely fascinating topic because it made us discuss what a movie is. Is it the intent of the artist? Is it the commercial product or is it the cultural product? And each of them had different answers and I got carried away. I went probably went into too much detail but it was really interesting to look, to use such a silly question to look at the industry and to look at art and subjective decisions and also what movies mean to us and how they shift over time. There’s no question I won’t consider because you never know where they’re going to take you.
Ashley: Yeah, yeah, for sure. And I just always like to get any other channels that you use. Are you on Twitter, Facebook, Instagram, any of those, I’ll link to those in the show notes and I’ll put your website the www.stephenfollows.com and I’ll also put a link to this report for everybody as well.
Stephen: Brilliant. Yeah. And, you know, share it as much as you can. It’s a free report, there’s no sign up, there’s no email. Well, I have a mailing list but that’s a separate thing you can choose to sign up to or not. There are many places in the screenwriting world where it’s a lead generation for a $500 course. That’s not what this is. We wanna help writers, we wanna give you this information so that you can be free to focus on the art and to have validation on things that you suspect. So please have a read of it, share it with everybody and that’s where that journey ends.
Ashley: Perfect. Are you on Twitter, Facebook, Instagram, any of those?
Stephen: Yes. I’m not on Instagram, I’m on Twitter as @StephenFollows, I’m on Facebook which I think is StephenFollows.com spelled D-O-T-C-O-M. But you will put the links in the show notes. And like I said, if you want to sign up to my mailing list every week I’ll send out my new article, I send out some recent press from the previous week, things that are literally around film data and a link to historic pieces and that’s obviously free. You can sign up and jump off whenever you feel. And that’s an interesting way to connect as well if you want.
Ashley: Perfect, perfect Stephen. Well, I really appreciate you taking some time and coming on and talking with me today. This was a fascinating interview. I really didn’t know a lot about you before we started but this was really just fascinating, fascinating stuff. So thank you again for coming on and talking with me.
Stephen: Likewise. It’s been really fun. And as I said, you’ve asked me a question on something I wish I’d studied, that’s always fun for me. So thank you.
Ashley: So that’s perfect. Well take it easy Stephen. I really appreciate it.
Stephen: All the best. Goodbye.
Ashley: Perfect. Talk to you later. Bye.
A quick plug for the SYS screenwriting analysis service. It’s a really economical way to get a high quality professional evaluation on your screenplay. When you buy our three pack you get evaluations at just $67 per script for feature films and just $55 for teleplays. All the readers have professional experience reading for studios, production companies, contests and agencies. You can read a short bio on each reader on our website and you can pick the reader who you think is the best fit for your script. Turnaround time is usually just a few days, but rarely more than a week. The readers will evaluate your script on six key factors- concept, character, structure, marketability, tone and overall craft, which includes formatting, spelling and grammar.
Every script will get a grade of pass, consider or recommend, which should help you roughly understand where your script might rank if you were to submit it to a production company or agency. We can provide an analysis on features or television scripts. We also do proofreading without any analysis. We will also look at a treatment or outline and give you the same analysis on it. So if you’re looking to vet some of your project ideas, this is a great way to do it. We will also write your log line and synopsis for you. You can add this log line and synopsis writing service to an analysis, or you can simply purchase this service as standalone product. As a bonus if your screenplay gets a recommend or a consider from one of our readers, you get to list the screenplay in the SYS Select database which is a database for producers to find screenplays and a big part of our SYS Select program.
Producers are in the data base searching for material on a daily basis so it’s another great way to get your material in front of them. As a further bonus, if your script gets a recommend from one of our readers, your screenplay will get included in our monthly Best Of newsletter. Each month we send out a newsletter that highlights the best screenplays that have come through our script analysis service. This is a monthly newsletter goes out to our list of over 400 producers who are actively looking for material. So again this is another great way to get your material out there. So if you want a professional evaluation of your screenplay at a very reasonable price, check out www.sellingyourscreenplay.com/consultants.
On the next episode of the podcast I’m gonna be interviewing writer-director Chris von Hoffman. He was on the podcast before, which is on Episode Number #165. So check that out of you haven’t already listened to it already. I’ll link to it in the show notes. In this interview we talk about his latest film called Monster Party. It’s a new Horror Mystery feature film. So we talk through that project, how he wrote that and how he got that one produced. So keep an eye out for that episode next week. That’s the show, thank you for listening