Reading · Australia In Cinemas

No Australian films released during the week so we get to see what happens when a behemoth of a film opens that all the exhibitors support (to the extent that there are virtually no screens available to show anything else).

The total Australian box office for the week was $31,489,669.81 across 163 films. THE ODYSSEY opened at $18.3 million, which is more than half the week’s takings on its own and makes every percentage below look worse than it otherwise might.

6 current-year Australian fiction releases in cinemas, earning $59,646.31 across 63 screens

3 current-year Australian documentary releases in cinemas, earning $5,186.00 across 5 screens
4 Australian films in rerelease, earning $2,946.00 across 4 screens

That’s $67,778.31 of Australian box office in a $31.5 million week.

The nine current-year Australian releases took $173,429.74 between them last week. This week they took $64,832.31. Most of that fall is LEVITICUS losing half its screens in one go, but the pattern underneath it is that every Australian title currently playing is in its second week or later, and nothing is arriving behind them.

Screens halved from 54 to 27 and the week total fell 64%. That’s a standard week five contraction rather than anything alarming, and the per-screen average of $1,665 is the second-highest of any Australian title this week, so the screens it kept are the ones that were working. It’s now the 3rd Australian non studio release of the year sitting above a million dollars. Hooray! I forgot to celebrate that the other week when it happened.

Held its 8 screens flat into week two, which distributors don’t always get. The week total halved, which is slightly lower than normal but $787 per screen on a small footprint is respectable.

Went up. Screens rose from 4 to 5 and the week total rose from $1,673.28 to $2,195.50, a 31% increase in week six. Small numbers, but growth at this stage of a run is uncommon enough to note.

Still on screens at week twelve. The odd detail here is that the weekend total was zero and the entire $597 came from midweek sessions.

Back again after a stretch off-screen entirely, on three screens.

The highest per-screen average of any Australian film in cinemas this week, ahead of LEVITICUS, on three screens in week four. Documentaries do this regularly and it still doesn’t get talked about enough.

Week twenty, one screen, and the week total nearly doubled on last week’s $55.

Twenty dollars. On a cumulative of $4.2 million. Somebody bought a ticket, and I hope they had a lovely time in the cinema by themselves or maybe it was 2 seniors tickets and they had a lovely time together.

The interesting thing that I’ve just realised is that Perth’s Revival House cinema isn’t included in these numbers. Last week was Aussie week at the cinema - Perth Revival House which only shows films in 35mm … None of these movies are in this list. Curious.

Hoyts | 1 screen | $1,057.00 week total

Roadshow | 1 screen | $871.00 week total

Walt Disney | 1 screen | $560.00 week total

Palace | 1 screen | $458.00 week total

No new Australian films this week, so no Availability or Awareness check. Instead I’m going to have a little existential crisis about AI and think out loud for a bit, so bear with me.

Last weekend I went to see Prompt: Make a Documentary, Zeke Morgan-Hind’s film about AI and filmmaking, which premiered at Revelation. I’m one of the talking heads, alongside Ben Young, Alex Proyas and a handful of others. If you’ve ever been interviewed for a documentary or created a doco, you know the shape of it. You sit down for an hour or two, you chat about a whole bunch of things that you’ve already prepared some answers for and about a minute survives to the final cut. It does mean the version of you on screen is a sliver of the conversation you actually had.

Zeke and I spent a long time on the money question (if AI replaces the jobs, who’s left earning enough to pay for the AI?), and the ownership mess underneath all of it which both are included in the doco. I’ll come back to that later in this piece, because I think it matters more than the parts we usually argue about. But we didn’t get included the environmental cost and how the conversation seems to be different at an independent level and studio level and what audiences might think.

The part I’ve been chewing on since, though, is smaller and more human and Zeke and I talked about it a bit in the interview but isnt’ included in the doco. Running through Ben Young’s sections, Basil Psanoudakis’s, and the Q&A at the screening afterwards, is one idea: a tool is only as good as the person holding it, and you can only hold it well if you understand what sits underneath it.

I agree with all of it, and I’ll add the selfish version. I like making things. I like writing this newsletter, cutting an EPK, the slow assembly of a film, doing the work to work out how a script can be improved. It’s hard and it’s satisfying, and the satisfaction is the whole point. So why would I want a machine to do the part I love? For me the joy is in the process, not the result, and I think that holds for filmmaking as much as it does for a Friday-night spreadsheet.

Here’s the moment it got uneasy for me, and it’s a small one. When I was cutting the Proclivitas EPK interviews in DaVinci Resolve, there’s a feature that stitches two clips together so cleanly you can shave out a pause and never see the join. Lovely tool. very useful when you only shot with one camera and don’t have anything else to cut to. And about thirty seconds in, I realised that with enough clips an editor could make an interview subject say more or less anything I wanted. (For now, a subject who moves around a lot makes that harder to pull off. That protection won’t last as the tech improves.) I was using it for the most boring, honest reason, tightening pauses, and the same feature, is a machine for putting words in someone’s mouth.

The tool that helps me tell the truth more cleanly is the same tool that helps someone else lie more cleanly, and the only thing standing between the two is the person holding it and whether they know what they’re doing and why.

Which is where I stop thinking about the tool and start thinking about the people. The next lot of people, in particular.

We learn by doing. That sounds like a poster on a classroom wall, but it’s the actual mechanism of this industry. Nobody arrives knowing how to cut a scene, run a set, or hold a boom for fourteen hours. You learn it on the job, badly at first, on the small films and the below-the-line roles and the unglamorous tasks that more experienced people are glad to hand down. Those tasks are the ladder. They’re also, almost exactly, the tasks AI is being sold to us as a way to skip.

The people who make it, even in the below-the-line roles, are disproportionately the ones who could afford to. Family support, a partner with a steady wage, savings, a cushion of some kind that lets you take the runner job, the deferred-payment gig, or the year you make your first feature for nothing. It has never been a clean meritocracy.

And it isn’t only at the entry level. The same filter runs all the way up the ladder, which I learned early, before I properly understood it. Back in 2011, at an emerging producer workshop with Screenwest, Sue Taylor spoke to us. She produced The Tree and Three Summers, and Minty, (which was one of my favourite shows as a kid). What stuck with me was her telling the room, completely matter-of-factly, that she’d often had to refinance her house to manage the cashflow on a production. At the time I filed it under war story. I didn’t really understand it.

I understood it the moment I went to make my own film. That’s precisely where you land, needing to bridge cashflow against money that’s coming but hasn’t arrived yet. You have to borrow it from somewhere and I could see then the obvious lever if you have it is finfiancing your home. Except I couldn’t pull it, because I didn’t have a “real job” the bank would recognise to refinance, which feels lije its own quiet joke about how this industry looks from the outside. And it makes me wonder how the producers coming through now will manage, when plenty of them won’t have the asset to begin with given how rediculous house prices are.

A feature film is a small business. The producer is a small business owner. So the most useful portrait of the film industry I’ve read lately isn’t about film at all. It’s the Australian Small Business and Family Enterprise Ombudsman’s Small Business Matters report from 2023.

Read it sideways and it describes us with uncomfortable accuracy. Around 43% of small businesses don’t turn a profit. Three-quarters of small business owners take home less than the average wage. They do it anyway, for meaning and purpose, and they’re more likely than the general population to volunteer and to do unpaid caring work on top. If you’ve ever met a producer, you’ve met that person.

The figure that stops me, though, is the age one. Just 8% of small business owners are under 30, half the peak of 17% reached in the mid-1970s. The most common age of a small business owner is now 50. The young aren’t coming in, and the report is plain about why: the economics don’t reward it, and the finance is hard to access, especially, it notes, for women. (That’s Sue’s refinanced house and my knocked-back loan, in aggregate.) That’s the canary in our mine. If the same wealth filter and the same economics apply to filmmaking, and I’d argue they apply harder, then the pipeline of new Australian filmmakers is thinning for exactly the reasons the Ombudsman lays out, and AI is arriving to thin it further.

The counter-argument, and it’s a fair one, is that there has never been a better time to make a film. That’s true. The camera in your pocket is extraordinary, the software is cheap, the barrier to shooting something has collapsed. But making a film and finishing a film are different sports. The deliverables, the chain of title, the QC, the legals, the distribution, and the sustaining of yourself financially across the years it takes: that infrastructure hasn’t got cheaper, and it’s the part that quietly requires either an institution or a cushion behind you.

Which brings me to a question I keep turning over and can’t land. Whilst doing the National Cultural Policy Review I assumed the answer was volume. Make more, give more films a shot at an audience, let people learn on the way through (there’s that we-learn-by-doing thread again), and once a film is finished the asset exists in perpetuity whatever it does on opening weekend. The trouble is the other case is just as persuasive. When it’s public money, every film funded is a film not funded, and spreading the same pool thinly can leave a lot of them not in the best position to recoup. Make fewer, resource them properly, get them seen. I can build either argument to a standard I’d believe, which usually tells me the real disagreement is upstream, about what the funding is even for, and I’m not sure we’ve ever settled that. So I’m going to leave this one open, because that’s where I am.

Which is why a small, boring decision about where I send this newsletter from has stopped feeling small.

This week I read Stephen Follows on what’s been happening to The Numbers, the box office data site a lot of us in this world lean on for US numbers.

His piece lays out something I have been aware of but the reasons why it happens realky shook me. Sites like The Numbers are now scraped by AI at industrial scale. By his account only around a tenth of their traffic is actual humans reading the site. The rest is bots, and the newest and worst wave is agentic, AI acting on prompts plus people writing their own scrapers, hammering the pages over and over for the data underneath - for gambling. The Numbers has fought back cleverly, even leaving instructions written for the machines telling them how to licence the data rather than take it, and that’s helped. But the line Follows lands on is the one that got me: mitigation isn’t escape.

This newsletter lives on Substack, in the open, which is the whole reason you can find it. That openness is also an all-you-can-eat buffet for a scraper. The section of his piece on how bad it really gets made me sit up and ask something I didn’t expect to be asking this year: should this move to a closed email list instead, a Ghost or a Beehiiv, where it comes to you directly and isn’t sitting out in the open to be harvested?

I don’t know, and part of what’s stopping me is money. Those platforms cost more than I can justify, and the only way to justify them would be to charge for this, which I’ve never wanted to do. So I’m sitting in the tension: the open platform is both why you’re reading this and why it’s exposed.

It’s pushed me somewhere stranger, too. I’ve started wondering whether the same logic applies to a feature film’s website as I’ve been thinking about how Australian films need to have the infomation listed for the SEO and AEO. Now I half-wonder whether the smarter move is a lean site carrying exactly the information you want an AI search engine to find and repeat, and not one word more, because everything you put up is read by the machines long before a person gets to it.

When I follow that thread all the way down, it lands back where this whole piece started. The bots want the result and skip the doing. They want the data, the words, the finished thing, extracted at scale, with no interest in the process that made it or the person at the other end.

So, in that spirit, a note on how this one was made.

Substack has added Pangram, an AI-detection tool, and a “how this was written” label. I’m glad they did, and I’ve left it on.

I use AI on this newsletter for two things. About twelve weeks in, I worked out I could speed up the weekly data by building formulas into the spreadsheet I assess each week. I know enough Excel to write them myself, but I’d just started working full-time, and doing it by hand would have eaten a whole day, so I had the spreadsheet turned into a template I paste each week’s numbers straight into. It returns the results (and yes, I glance-check them every week, and if you’re the sort of data nerd who’d enjoy the raw spreadsheet, say the word - I’ll send you a week - it’s a good time). The second thing is the shaping of pieces like this analysis: I vomit draft bring the thinking, the anecdotes and the argument, and I work them into shape with an AI before editing every line until it reads like me and says what I mean. I could write this all by hand and some weeks the thoughts have come out so clear I don’t need to move beyond my vomit draft. What I won’t hand over is the thinking itself. Every time I’ve tried to get AI to write this newsletter , it swings the whole thing toward the US industry, or invents Australian details that aren’t true, and this newsletter only works if it’s mine and it’s right. The assessing, the judgment, the calls about what matters, those come from my own noggin.

For the record, I’ve got Substack’s “don’t train on my content” setting switched on, for whatever that’s worth against a determined scraper.

And there’s plenty I wish AI would do and won’t: scraping Flicks each week to tell me how many session times a film has, not just how many cinemas it’s in, would change my week.

So would something that does my laundry and the dishes becuase most of the time they do not bring me joy - the dishes sometimes but never the laundry.

Till I see you at the cinema next,

Kate

No posts

BOX OFFICE RESULTS - 2026 - Week 29

Comments