Data Deep Dive: Do Thursday Night Numbers help?

This is admittedly a controversial topic.  From what I’ve seen on the forums, people feel quite passionate about whether or not Thursday data should be used and what it means for the Friday 9a Pacific deadline.  As opposed to taking sides in that argument, what I’m going to look at more closely here is the degree to which those numbers help.

Why should you care?

The idea here is that the Thursday data can potentially be an indicator of how a particular movie will do the remainder of the weekend.  But what multipliers can you apply to Thursday data to figure out a range of what an entire weekend might be?  Is it different for different kinds of films?  This is what I set to find out.

Where’d I get my data?

This is the part that I’m a little squeamish about.  The data from this analysis comes from finding old articles off of ProBoxOffice.com going back to Week 1 of the Summer season.  I had trouble finding easy to locate sources of older information to the point that Google quit showing me ProBoxOffice.com results because they thought I was trying to hack them.

That said, I’m not 100% confident there isn’t a better data source out there, so don’t necessarily hate the analysis but feel free to hate the data and suggest a better source.  Expressions of my idiocy are also welcome, just be polite about it.

There are four potential problems with this data:

  1. It only includes new movies, not Thursday night revenue from existing films.
  2. Sometimes it doesn’t even include new movies as Thursday night data is often not available even for all new movies.
  3. Screen counts on Thursday sometimes do not match screen counts for the remainder of the weekend.
  4. Some movies come out on Wednesdays, so their Thursday data is second day data, not what amounts to late night opening day data.

Given all that, the analysis below throws out movies that opened on Wednesdays and omits movies for which there was no data (meaning new films with no reporting and old films that don’t appear to report at all).  If you’re interested, you can find the data I collected here:

https://s3-us-west-1.amazonaws.com/compare.fmlnerd.com/Historical+Thursday+Returns.xlsx

What were my findings?

Despite all those limitations (or maybe because of them), some patterns emerged.  There are 30 movies in this data set going back to Week 1 of the Summer FML season through last week and what I did was figure out what percentage of their overall box office came from its Thursday reporting.  If you rank the 30 movies in the sample based on that percentage, you get this graph:

Thursday

What you can see here is that 63% of the films in the sample brought in between 4.9% and 8.2% of their eventual weekend box office revenue on Thursday night, which gives us some good bounds to work with.

This past week, I rounded those bounds to 5% and 8% and used them against the ProBoxOffice.com Weekend Forecast for “The Visit” when I wrote:

At a $17.2M forecast, you would expect to see Thursday numbers in the $1M to $1.3M range.

Where did I get those numbers?  Simply multiply $17.2M by 5% and again by 8% to get your range.

The bounds work the other way too.  When you see the Thursday numbers, divide them by 5% and then again by 8% and you get a range of what it’s possible weekend would be.  For example, when “The Visit” brought in $1.02M last Thursday you could compute a range of between $12.75M and $20.4M for its entire weekend.

Part of the problem, though, is that for its Thursday estimate, “The Visit” was on 2,200 screens.  How many screens would it play on for the whole weekend?  In the ten minutes I had between meetings on Friday morning I couldn’t figure that out.  After the fact, we now know that it appeared on 3,069 screens all weekend, and that was among the reasons why it broke the bounds and actually brought in $25.4M.  So in this specific case, the Thursday numbers gave a thumbnail that this film was doing well, but given the potential absence of other data it didn’t give an exact enough picture to necessarily change your lineup.  In other instances, it might give a clearer sense but in this one it was a little deceiving.

Here’s another curveball in the data.  Notice anything in common about the movies below 5% in the graph above?  It’s a small sample size but some of them are movies that appeal to an older or younger audiences.  In other words, people not likely to go to the movies late Thursday night so their totals from that timeframe are a smaller percentage of its overall weekend box office.

Now look at the moves above 8%.  Those are mostly movies that have a built in fan base.  Consider “Paper Towns”.  Hard core fans of the book rushed to see that movie Thursday night, got disappointed at how different the movie was, and word got out so those returns were a much higher percentage of its overall weekend revenue.

So back to the question: Do Thursday night numbers help?

Under certain circumstances, yes they do.  Namely, if the movie in question is new, does not open on Wednesday, actually reports it’s Thursday night numbers with roughly the same number of screens it will see over the weekend and the data is available before 9a, Pacific.  That sounds like a lot of qualifiers to work with but it does happen often enough that it’s worth looking out for the announcements Friday morning and using the 5%/8% multipliers to provide a guideline for most films.

4 thoughts on “Data Deep Dive: Do Thursday Night Numbers help?

  1. Hey, Pete –

    I came looking for this article to get the formula for figuring out what tomorrow’s numbers might mean for Goosebumps and (hopefully) Woodlawn.

    And in doing so, I realized that you’ve been talking about “screens” showing a particular movie. If you’re using Box Office Mojo’s theater count info, that’s not actually an indicator of how many screens a movie will be shown on, just how many theaters are showing it. So whether a theater is showing a movie 20x a day on several screens, or 1x a day on one screen, it counts as one theater. (This weekend, The Walk is losing all or almost all) its IMAX screens, but the theater count is just dropping by 20 because most theaters will still be showing it on “normal” screens even though they won’t have it on their IMAXs.)

    I don’t know if that changes anything … or if you’re not using theater counts but have a source for actual screen counts (if so, please share!)

    Anyway, I’m going back to over-analyzing what I think Goosebumps, The Walk, and Woodlawn will do this weekend. I’m convinced one of them will be the bonus. And I’ve got a little over 14 hours to figure out which one! (And hopefully squeeze some sleep in)😀

    Thanks for everything you do for the FML!

    Like

    1. Thanks so much for the kind words, I really appreciate it. I’ve seen you comment on Chatter threads with similar ideas about screens vs what actually gets shown and while I’m fascinated by that nuance, I too am unsure about how it impacts any analysis. I just don’t know and I’m not sure there is a source of data that would help us know. If you find one, though, I’m game to give it a whirl.

      As for this weekend, I like Goosebumps personally but am scared of Woodlawn. It just seems like with the faith-based movies you never know. Faith-based football movies, more so 8).

      Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s