I know, it’s maddening. People watch 10-hour series that take five hours to get good,
but your 101-minute comedy is too long. Michael Bay makes three-hour Transformers movies, but your 95-minute drama is too long. Critics love seven hours of Sátántangó, but your 18-minute short is too long.

Sorry, but it’s probably true. I learned this the hard way on my first feature film, Jake. After our initial 118-minute cut, we proudly got it down to 104 minutes. We couldn’t cut a frame more! We locked picture and sent out our perfectly formed film to festivals.

Final runtime? 88 minutes. After a series of festival rejections, we took another look. Suddenly, once-impossible cuts seemed obvious. We literally replaced 12 minutes of subplot with a sound effect. Other tightens were shorter — in some cases, a few frames here, a few frames there. But, it all added up.

At least we weren’t alone. Stanley Kubrick legendarily trimmed both The Shining and 2001: A Space Odyssey after they were already in theaters. The Brown Bunny and Southland Tales each went through massive re-edits after Cannes. Most recently, David McKenzie premiered Outlaw King at Toronto, only to trim 20-plus minutes before its Netflix release a mere two months later. And from Apocalypse Now to Annie Hall, there’s no shortage of films that went through famously protracted edit sessions to find their best shape.

But the Kubricks and McKenzies of the world get a chance to make a second impression. You won’t. If your film is flabby, it will struggle to get into festivals, it will struggle to get a good reception at festivals, and it will struggle to get distributors. So: Don’t rush it. Your priority isn’t to rush the edit. It’s to make the best film, to look at every scene, every moment, and make sure it’s in maximum service to the film.

Here’s how.

Start later, end sooner. You probably already know this: Get into a scene as late as possible, get out as early as possible. But how do you know when that is? There’s trial and error, of course, but a great way to see what other filmmakers have done is to read the scripts of films you know well. I remember being struck reading the script to Rushmore, noticing a few scenes that had extra exchanges at the end that weren’t in the finished film. It wasn’t that those lines added nothing— they just weren’t worth the extra screen time.

One increasingly common way to get into a scene earlier is to underlay dialogue over shots of people arriving. In effect, the conversation starts over the car door closing, the host rising to greet the guest, then suddenly snaps back into sync with a line. I don’t always love this — usually because it’s an obvious last-minute way to tighten a scene — but that doesn’t mean it isn’t better than the alternative of watching a car door close, watching somebody slowly rise from their seat and so on. (A recent example of this is House of Cards S6E7, where Claire’s speech to a rally runs under footage of a torch being brought in. The shift from underlay to real time is so awkward it’s hard to imagine it’s pre-planned, but the content doesn’t merit any extra screen time, and it’s a necessary sacrifice to keep pace.)

“Start later, end sooner” isn’t just true on a scene-by-scene basis; it works for the film as a whole. Are you starting with exposition or scene-setting that audiences can easily figure out on the way? Do your audiences really care what happens to every supporting character? Do you need the last line of dialogue, or do we know what the character’s going to say? Without resorting to spoilers, my favorite ending of the year is A Quiet Place. It’s unexpected and yet perfect. We know exactly what happens after the film ends, and we don’t need to see it because the film has said what it’s going to say. So, ask yourself: At what point have you said everything you’re going to say, and why haven’t you cut to black after that point?

Find the “scaffolding” and rip it out. When you start writing a script, there’s an initial area from which everything gets built. In Jake, it was the lead character’s place of work. I wrote several scenes there and developed the character around it. It was only well after we completed our initial edit that I realized it was adding nothing thematically. They were also the first scenes that we shot, and while there are some moments I’m proud of, there are others that really didn’t work that great. But I still thought I needed them. Eventually, I finally realized there was only one relevant plot point that removing those scenes would complicate, and a sound effect would fix that problem.

This is always a challenge in editing your film — working out the difference between what was important to you in envisioning the film and what’s important to the viewer now that it’s been shot. I use the term “scaffolding” because it’s what you’ve used to build your filmic universe that can be removed once it’s done. It’s not necessarily plot or setting, of course. It could be lines of dialogue that are redundant because the performance or production design tells the tale. It could be moments that drove the theme home in the script that are thuddingly redundant on screen.

ADR is your best friend. Bottle Rocket is the film I’ve seen more than any other, but it was only when I caught up with the Criterion Blu-Ray that I had a mind-blowing revelation. In the original cut, it was the police’s discovery of Bob’s pot plants that sent Anthony, Dignan and Bob on the run, a series of events that took ages of screen time. In the final cut, it’s all gone, replaced by one line at the end of the bookstore heist: “We’re heading back to Bob’s, then we’re going to hit the road and go on the lam.”

If you think about it too much, it doesn’t make a lot of sense — why wouldn’t you just go on the lam straight away? But you don’t have time to think, partly because it fits with Dignan’s character and partly because it’s a fast-paced scene. We cut right to the only scene at Bob’s that’s needed, and then we’re on the road.

Once you start noticing how ADR can help cheat things, you’ll see it everywhere. The new film Overlord has a particularly obvious example, where Wyatt Russell’s Corporal Ford sends Jovan Adepo’s Boyce to go find two other soldiers. It’s a weird little moment, but it’s quickly rolled past once you start seeing stacks of deformed bodies getting destroyed by flamethrowers. And it points to a truth about editing you have to embrace: Sometimes it’s not about the best solution, but the least worst. And while I haven’t seen the deleted scenes to Overlord, I’m betting there’s nothing important I’m missing.

Production isn’t over until you lock cut. The notion that you’re done shooting when you start editing is a doggedly 20th-century one, at best. But many indie filmmakers let themselves get locked into a corner with the footage they have, instead of breaking out with the footage they can get.

One of the most common self-created prisons is “we can’t get the camera/DP that we shot the film with.” I’m here to tell you: It doesn’t matter. My film was shot on a Red One, but when we were editing, we realized some establishing shots would both add clarity and let us tighten some scenes with underlay. We shot them one afternoon on a DSLR we had sitting around, with the thought to maybe get them later on a higher-resolution camera. We never did.
Fun fact: Literally no one has ever pointed that out or cared. (Having an awesome colorist helped, but still.) And if Peter Jackson can use a GoPro in a 3D 48fps production of The Hobbit, you have my full permission to use a DSLR insert or establishing shot in your Alexa-lensed film.

Also, royalty-free stock footage is plentiful and abundant! For $60 — sometimes less! — you can get that shot of rolling waves or establish a field that helps bridge an awkward gap. A good colorist (have I mentioned that they’re worth their weight in gold?) can integrate said footage with the look of your film. And if it allows you to jettison 45 seconds of awkward stage-setting, it’ll be the best money you spend on your film.

A few frames every few seconds adds up quickly. Let’s say that you shoot every dialogue scene in shot/reverse shot, cutting between for each one. Let’s say there’s 11 lines of dialogue exchanged every minute of your film. Let’s say you have a 90-minute film. That means, if you shave three frames off each cut between speakers — at ten cuts per minute, that’s 30 frames per minute. That sounds like nothing, but it’s 2,700 frames over the course of your 90-minute film. If you’re shooting at 24 fps, that’s 75 seconds less of what might be dead air.

A similar thing is true for documentary editing. I see many documentaries where people’s long, sometimes awkward and “um”-filled interview responses are overlaid with pictures. It would be the easiest thing in the world to trim those sentences, help the subjects sound more thoughtful and make their point more clearly, and trim two to three seconds out of every sentence. If you do that, say, 60 times in a doc, your film is three minutes shorter, and you’ve most likely lost nothing of value and gained clarity.

I’m not saying everything has to go by as quickly as possible. But by tightening things up as much as possible, you’re giving yourself not just more time to play with, but the ability to slow the pace at key moments to make those moments more meaningful. If there are pregnant silences everywhere, none of them are meaningful. But if one suddenly jumps in at a key time, it can carry a powerful emotional weight.

At 90 or 15? Keep going. Making a film is a marathon, and it’s impossible to go full-tilt every step of the way. So it’s understandable that when target numbers are reached, people relax their discipline. A “good number” for a feature is popularly considered 90 minutes; for a short, 15 minutes. I often hear filmmakers use those numbers as targets.

But just because that’s the running length you’ve gotten your film to doesn’t mean it’s the right length for the film. Can you do a lot more with a lot less? Brimstone and Glory, the Mexican fireworks festival documentary, is one of the best films I’ve seen this year, full of explosions, danger, character, beauty, and impossible spectacle. Runtime: 67 minutes. Purportedly, that’s commercial suicide, but I just saw it in a New Zealand cinema. You’ll be lucky if your film gets there. And if it’s a 75-minute film trapped in the body of a 90-minute film, chances are it won’t.

Get eyes on it. You are never as acutely aware of the failings of your own film as you are while watching it with other people. Now, most of us can’t afford to get 80 randomly recruited strangers and give them polling cards — but that, quite honestly, is a good thing. You don’t need that.

For our 118-minute first cut of Jake, we had a test screening on our living room TV with four people who represented the range of tastes of people we might thought enjoy our film, from mainstream-leaning to experimental. While talking with them was quite helpful and led to some good changes, I think 75% of the cuts we made after the screening were ones that were blatantly obvious during the screening — mostly indulgently extended scenes that we enjoyed but that played much longer than our narrative required and where the gag played itself out quickly for first-time viewers.

Now, we were fortunate in that our viewers worked in the film and TV industry. Yours may not. But that doesn’t mean their viewpoint isn’t equally valuable. The key to remember about any feedback is this: Most people are 100% correct in their ability to identify problems with your film. However, they don’t want to sound unhelpful. So they’ll also suggest a solution. And that solution, unfortunately, is usually a bad one.

But don’t throw the baby out with the bathwater. Take the note with a smile, no matter how painful it is — they’re doing you a favor, after all — and think about what might fix it. Often, it’s setting up something more clearly 10 minutes earlier so that stakes are clear, or something else that’s nowhere near the actual point where the viewer experiences the problem.

Elisions are exciting. There’s a fine line between exciting the viewer by withholding information and confusing the viewer, it’s true. But most films stay safely on the former side of that line. They underrate the intelligence of the audience, and often the audience repays that disrespect with disinterest.

Some of cinema’s most memorable moments are about what isn’t shown. There’s the off-screen heists in Reservoir Dogs and The Old Man & the Gun, the mysterious contents of containers in Belle du Jour and Pulp Fiction. Or think of the difference between the theatrical cut of Donnie Darko, with its intriguing unexplained edges, and the director’s cut, with its unnecessary explications.

More radical cuts can be even more exciting. I recently watched Alain Resnais’ Muriel, which often elided whole scenes; for instance, cutting from a person picking up their first bite of dinner to another angle putting the fork down on an empty plate. On the more mainstream side, the opening of Widows cuts thrillingly between scenes of domesticity and a heist gone wrong, quickly abridging the latter for maximum impact. Or, if you want to go more experimental, check out the films of Angela Schanelec, who pushes “what you need to know” to the extreme, often demanding viewers either do legwork to keep up or else live with uncertainty.

Now, this may not be for you. But the joy of nonlinear editing is the ability to prototype all sorts of different variations quite quickly. Take advantage of it. If a scene isn’t working for you, take it out. And maybe that encourages a style where you can take other things out. Or maybe not! But you don’t know if you don’t try.

No one cares how hard you worked for a shot. When I saw Be Here To Love Me, Margaret Brown’s documentary about Townes Van Zandt, I noticed Cowboy Junkies thanked in the credits, along with some other biggish names. I asked why they hadn’t made the cut. Brown said after watching the first cut her first editing decision was to take out every time somebody called Van Zandt a genius, no matter how big of a star they were. It’s a comment that made me wish she could have been in the editing room of dozens of other documentaries. Be Here To Love Me is, like all the best music documentaries, a story about a human, not a sales pitch for a musician.

So, get rid of your celebrity talking heads if they say nothing! The same goes for drama directors: Be ruthless with your unnecessary cameos, or your tricky shots, or that expensive-looking location. It’s an act of tremendous discipline to step back and systematically go through every single shot in your film and ask yourself, honestly, “Why do I need this?” This is one reason that, despite being a professional editor, I had someone else edit Jake. Peter Evans, the editor, wouldn’t know what was easy or hard on the day. He’d just make the best possible cut of the scene and discard what seemed irrelevant.

Another anecdote: Almost a decade ago, I was flying to Los Angeles and arriving at the same time as a screening of Troll 2 and the documentary about it, Best Worst Movie, with its main subject George Hardy in attendance. I happen to share a deep love of Troll 2 with Andrew Todd and Johnny Hall, the directors of Ghost Shark 2: Urban Jaws, which was filming at the time. I asked if they’d be interested in having me get a quick shot of Hardy railing against the menace of ghost sharks. Instead, Todd wrote a two-page scene for me to shoot with Hardy, Juliette Danielle (The Room) and Alan Bagh (Birdemic: Shock and Terror). And shoot it I did.

Bringing those three actors together in a scene would have guaranteed Ghost Shark 2 a certain notoriety, but the fact was it didn’t really fit the tone of the film and sent the story off in directions that didn’t pay off. So, the directors cut it. But it also led to a whole new set of scenes featuring Danielle that worked much more cohesively. I know it wasn’t an easy decision for Todd and Hall to give it up, but in the end, it’s the film that matters.

Just because the talent needs it, doesn’t mean you do. Every actor has their own needs, and some of those may involve pacing through their performance in certain ways to get to places. There’s one scene in Jake where an actor wanted to take a moment, turn away, turn a water tap on, turn it back off and then turn to the other actor before delivering a line. I was 99% sure it wouldn’t make the cut, and that most of the moment of processing would be happening on the other actor’s face. But I let him do it anyway because it was what he needed in the moment and because I could cut around it cleanly if it didn’t seem necessary. Besides, I could have been wrong and it might feel necessary in the edit.

The water tap didn’t make the edit. Lots of moments didn’t. I’m not trying to throw that actor — or any! — under a bus. It’s incredibly difficult to give a film performance and know precisely how much will be carried on your shoulders at any given time. And an actor has to be prepared to be 100% believable any time the director chooses to cut to them, and make whatever transitions that seemed to work on the page believable in real life. But cinema isn’t real life. And those two beats the actor takes before replying may be two beats you don’t really need.

The variant of this for documentary is that often directors want to include things that the subjects felt were really important. This is tricky, because of course you want to maintain a proper relationship with the subjects. But it can also lead to cumbersome, lengthy interview grabs that explain things in unnecessary detail or nuances of story that aren’t relevant for anybody involved. You can’t betray your talent if you’ve made promises, but you do need to do them justice. Sometimes that justice comes in doing what’s best for the film, not what they think is best.

Don’t forget: Fat adds flavor. Let me add a giant caveat to everything I’ve just said. The phrase “penny wise, pound foolish” usually applies to pinching pennies as false economy. But the same can apply to injudicious editing. Forever, it’s been a rallying cry of producers to “cut the fat.” But what does that mean? For too many, the mantra is “everything that doesn’t serve the story has to go”.

I kind of hate that mantra. Film is not a plot delivery device. If you want the story told efficiently, read Wikipedia. (Ditto for documentary: If all you’re doing is illustrating a Wikipedia article, don’t you have better things to do with your time?) Back in the day, I saw Attack of the Clones on an IMAX screen because that version was only 120 minutes and the film was notorious for various annoying subplots. I thought that version might cut to the chase more quickly and jettison unnecessary material. Instead, it lurched wildly from moment to moment with no sense of pace, and was quite probably a more painful sit than its 142-minute theatrical cut. (To be fair, I’ve never bothered to find out.)

The power of cinema isn’t in torrents of exposition or in swiftly dispatching one plot point to move on to the next. It’s in the held moment that shifts the viewer’s emotional understanding of a moment, of a character, of the tone of the film. Take the stand-out “Cheddar Goblin” scene from Mandy. It could be excised in full with zero impact on the storytelling of the film. But who the hell would think that’s a good idea?

Ultimately, a runtime is just a number. Unless it’s strongly outside the standard deviation, people won’t remember how long your film is when it’s done. What they’ll remember is how long it felt. So, ask yourself honestly: Does your film feel shorter than its runtime? Or longer?

Doug Dillaman is an American-born filmmaker and freelance video editor who lives and works in Auckland, New Zealand, where he is making his second feature film, Gut Instinct. Learn more about his debut feature, Jake, at www.jakethemovie.com.