Wait times at Walt Disney World have been significantly higher than we expected over the last few months — 20% higher, year over year, in 2015. Since our crowd calendar is based on wait times, we needed to determine the causes. And we will have updated crowd calendars within a few days.
Today we’ll look at why we think wait times are higher at Disney’s Hollywood Studios. I’ll cover each WDW park in a separate blog post starting tomorrow.
We think higher wait times at Disney’s Hollywood Studios are almost entirely due to these 3 things:
- Closed attractions
- Mechanical problems at Tower of Terror
- Better handling of FastPass+ return lines in the park
We explain each of these in detail below. We also looked at whether Disney has changed the number of FastPasses available, ride capacity, the ratio of FastPass guests to standby guests, or whether guests are using FastPass+ significantly more now. We don’t see any clear evidence of those happening; that’s also explained below. (There’s also increased attendance at the other parks, which we’ll talk about when we cover them.)
Our analysis needs a little bit of math, but it’s all pretty simple.
Take a ride like the Tomorrowland Transit Authority. Its track holds 32 ride vehicles. Each vehicle holds 10 people. One complete trip around the track takes about 10 minutes, so each ride vehicle goes around the track 6 times per hour.
The number of people that the TTA can serve in an hour is 1,920:
32 ride vehicles x 10 people per vehicle x 6 times per hour = 1,920
If the wait at the TTA is 10 minutes, that means there are 320 people ahead of you in line :
10 minutes of wait / 60 minutes in an hour * 1,920 people per hour = 320 people
You can use the same calculation at rides with FastPass+, like Toy Story Midway Mania. TSMM has an hourly capacity of around 900 people. If the posted wait is 50 minutes, the number of people who get to ride before you is:
50 minutes of wait / 60 minutes in an hour * 900 people per hour = 750 people
Most of those 750 people will come from the FastPass+ line – in fact, around 80% of the people who ride before you will come from the FastPass+ line. We’ll use that information a bit later. For now, the thing to know is how to figure out the number of people ahead of you, using the standby wait and the hourly capacity at each ride.
Standby Wait Times for 2014 and 2015
We collect thousands of posted wait times from our users, from our employees, and from Disney’s My Disney Experience app, at every ride, in every park, every few minutes, every day. We’ve done this for many years, and we have over 8 million wait times.
Here’s the average posted wait in line for Toy Story Mania in January 2014 and January 2015, between 10 AM and 5 PM :
So in January 2015 at Toy Story Mania, there were around 90 fewer people ahead of you to ride TSMM between 10 AM and 5 PM, versus January 2014.
Here’s all the DHS attractions we track, for January 2014 and January 2015:
The red lines are for the Studio Backlot Tour, which closed in September 2014, and the Legend of Captain Jack Sparrow, which closed in November 2014. There were no people in those lines in January 2015. We think most of them got in line at other attractions.
There were also slightly fewer people in line at Muppets and Tower of Terror in January 2015. For the month, there were around 792 fewer people in line at all DHS attractions between the hours of 10 AM and 5 PM, as compared to 2014.
Here’s April 2014 and 2015. Easter was in April in both years, which draws high crowds in the weeks around it. Wait times were slightly higher in 2015. The higher number of people in line in 2015 almost exactly matches the number of people who were in line at the closed attractions in 2014 :
That’s kind of amazing. You can explain 98% of the increase in Easter wait times (122 out of 5,865 people is the remaining 2%) by showing that most guests who used to be in line at closed attractions, got in line at open attractions.
We summarize all of 2015 below.
But Wait…There’s More (Closed Rides, That Is)
The charts don’t include lines for The American Idol Experience, which closed in August 2014, or the Magic of Disney Animation, which closed in July 2015. These attractions didn’t have posted wait times, and we didn’t include them as part of our DHS crowd calendar. And they were not very popular. But American Idol had anywhere from 2 to 4 performances between 10 AM and 5 PM, and Animation had drawing classes, character greetings, and a few things to do. Some people were in those attractions.
It’s reasonable to say that some of the people who used to go on those attractions now get in line somewhere else. Here’s the average number of extra people in line at DHS in 2015, before and after Animation closed:
Idol closed in August 2014. Here’s the year-over-year difference in extra people in line for September and October (there’s some overlap with Animation, too):
Saying that people move from closed rides to open rides is obvious. It’s the scale of the impact that’s interesting.
How Ride Closures Increased 2015 Wait Times
Wait times are up 20% at DHS for rides open in both 2014 and 2015 (from 30 minutes to 36 minutes) so far this year. That works out to around 319 more people in line at the open rides. But the closing of Backlot Tour, Legend of Captain Jack Sparrow, American Idol, and Magic of Disney Animation account for all of this change and more. In fact, attendance may be down slightly at DHS. Here’s why we think that.
Of all the extra people in line at DHS at any given time in 2015, almost 45% happened at Tower of Terror in February and March:
Tower’s wait times were up 85% in February and 25% in March. It was up 4% in every other month of 2015. That indicates mechanical issues at Tower of Terror, not a brief flash of popularity.
A Smoother FastPass+ Return Experience Increases Standby Waits
When Disney World transitioned to FastPass+ in early 2014, it did not start well. (We wrote a blog post about it in March 2014.) Many guests didn’t understand how it worked. Verifying a reservation took too long, and Castmembers had to deal with MagicBand hardware issues. What used to be a 5-minute wait with paper FastPasses, often turned in to a 10- to 15-minute wait to use FastPass+. Guests complained about long FastPass+ return lines, when the whole point was to make lines shorter.
The problems with FastPass+ in 2014 meant FastPass+ impacted standby wait times less in 2014: Because people had problems getting in the FastPass+ line, a bit more of each ride’s capacity went to standby guests, reducing the standby wait in line.
Most of those FastPass+ issues have been resolved in 2015. MagicBands scan faster, and when lines start to develop, Castmembers have been trained to scan only one person’s MagicBand, instead of everyone in a group. Since it’s faster to get into the FastPass+ line now than in 2014, standby lines are slightly longer.
To put this in perspective, in 2014 we estimated that your wait to use FastPass+ was going to be 15 to 25% of the posted wait time. That’s closer to 10% in 2015. It’s not perfect – long lines still happen sometimes – but it’s better. And we think it’s increasing standby wait times slightly versus 2014.
Why Our Crowd Calendar Didn’t Anticipate These Changes
We track hundreds of variables for our crowd calendar models, everything from park hours to weather. One of variables is “Total ride capacity lost to closures.” And we had the right numbers in there, but the models didn’t adjust properly. At this point, we think that model didn’t give enough emphasis to the high number of ride closures. The models seem to work well with one or two ride closures (we’ve done pretty well at Epcot for 2015, even with Maelstrom and Captain EO closed), but we don’t think the model has ever seen this many rides closed in one park at once. We know that the model sometimes has trouble with things it hasn’t seen before, and we’re working with the company that wrote our modeling software to make sure we make the right adjustments .
We should have an updated Crowd Calendar out within a few days. Once we think we’ve got mass closures handled properly, we’ll do another update after that. We know both Soarin’ and Maelstrom will be closed at Epcot in early 2016, and we want to make sure we handle that.
Other Ideas We Considered
Lots of people have a many ideas about what’s causing the higher wait times. We tested many of them and found them less convincing. Here’s a short list of some things we tested.
Is Disney Intentionally Lowering Ride Capacity? No. I said this on a WDWToday podcast and in blog comments here. I don’t think it’s true now. You can explain virtually all of the wait time increases across all four parks by looking at attraction closures, more efficient FastPass+ lines, and mechanical issues. Disney management may be a lot of things, but they’re not Enron. I regret saying it, and I apologize.
My spokesman, Arthur J. Fonzarelli, will now read a prepared statement on my behalf :
Are Posted Standby Wait Times Artificially High? Not as far as we can tell. It’s well known that the posted wait time outside a ride is usually wrong. It’s wrong at every theme park, including Disney. And it’s often wrong on purpose: sometimes it’s meant to be a pleasant surprise, such as when the posted wait time is 60 minutes and you only wait 45. Other times it’s a way for park management to tell you to go somewhere else, such as when Toy Story Mania has a posted wait of 80 minutes at park closing, when the actual wait is 15; they want to close the park on schedule and avoid paying overtime.
Our Lines app lets our users send in both Disney’s posted wait, and the actual amount of time they spent in line. We compared the ratio of actual waits to posted waits for 2014 and 2015 over 11,000 samples. That’s not a huge number. At many minor attractions, we don’t have enough actual wait/posted wait pairs to say anything. But for the big attractions, as far as we can determine, the ratio of actual wait to the posted wait time has not increased significantly across WDW. (If anything, the posted waits are getting slightly more accurate at some attractions.)
Is Disney Prioritizing FastPass+ Guests Over Standby Guests More in 2015? Not according to Castmembers or our counting.
When Disney adds FastPass+ to an attraction, it has to set aside a certain amount of the attraction’s capacity for FastPass+ riders. The default is around 80%, meaning 4 FastPass+ users will ride for every 1 standby guest (a 4:1 ratio). If a bunch of FastPass+ users show up all at once, the Castmembers at the ride can bump up the ratio to 10:1, meaning 91% of the ride’s capacity goes to FastPass+.
If that doesn’t reduce the FastPass line, Castmembers can take 20 FastPass+ guests for every 1 standby guests – 95% FastPass+ allocation – until the FastPass line disappears.
How do we know this? We asked Castmembers what the rules were, and they told us. Then we counted the number of people in line to check their answers.
Last Thursday, we sent Seth, one of our researchers, to stand in line at Peter Pan. It had a 60-minute standby wait. When he got to the front of the line, just as he was about to board, he turned around and walked back through the line, counting the number of people in the standby line as he passed them. He counted 196 people, and the posted wait was still 60 minutes. That indicates around 82%% of the ride’s capacity was being held for FastPass+ guests, almost exactly the 4:1 ratio the Castmember said it would be.
Seth found the same 4:1 ratio at The Many Adventures of Winnie the Pooh, Princess Fairytale Hall, Jungle Cruise, Buzz Lightyear, Space Mountain, and later at Toy Story Mania. The Castmembers at Seven Dwarfs Mine Train said they used a 3:1 ratio (i.e., 75% FastPass+ and 25% Standby); we counted the number of people in line yesterday and that was accurate. It’s a Small World was slightly higher on Thursday (around 90% / 10%), but back to normal yesterday. So with only a few exceptions, 4:1 seems to be the standard ratio, and has been for a long time.
Is it FastPass+ Supply or Demand? Fred, Steve, and I spent a long time looking at this. As far as we can tell, there’s no clear trend on FastPass+ supply, demand, or the relationship to standby wait times.
First, a couple of things: We track the FastPass+ availability every few minutes, at every attraction, every day, up to 60 days in advance. As I write this, we have more than 16 million FastPass+ data points. Knowing when FastPasses run out helps us tell you which FastPasses you should get in advance, and which attractions should have FastPasses on the day of your visit.
Here’s a chart showing Great Movie Ride’s average FastPass+ availability and standby wait times, by week, between 10 AM and 5 PM, for 2014 and 2015. The Fastpass+ times represent the amount of time before a Fastpass+ reservation became unavailable. So for example, if at 10 AM the last available reservation was for 10:45, that would be 45 minutes. If on a Tuesday we saw the last reservation get nabbed on Sunday that would be two days.
The lines in the top half of the page show when FastPasses run out. Red lines are for 2014, blue for 2015. Higher lines mean FastPasses ran out faster.
There’s a small gap in 2014 FastPass+ data, where our code broke and we didn’t collect anything. But even then, it’s clear that for much of 2014 and 2015, you could get a FastPass+ for Great Movie Ride within 15 minutes of “now” – an “instant” FastPass. If demand for FastPass+ was up in 2015, you’d expect FastPasses to run out sooner; the blue line for 2015’s FastPass availability would be consistently higher than 2014’s red line. But it doesn’t look like that.
The bottom half of the graph shows the average posted wait time at Great Movie Ride. Again, red lines are 2014 and blue lines are 2015. Wait times are significantly higher since early April 2015, even when you could get an “instant” FastPass on the same times both years. In fact, it was harder to get a Great Movie Ride FastPass in late July 2014 than in late July 2015, and waits were higher in 2015. So no matter whether FastPasses ran out faster or slower in 2015, wait times seem to be higher.
Here’s MuppetVision 3D. We wouldn’t recommend using FastPass+ at Muppets. The thing to note is that there doesn’t seem to be a clear relationship between FastPass+ availability and standby waits. The two lines don’t move in tandem.
Here’s Rock ‘n’ Roller Coaster:
FastPasses ran out as fast or faster at Rock ‘n’ Roller coaster for 8 weeks in the second half of 2014 versus 2015, but wait times are higher (or the same) in 2015.
Here’s Toy Story Mania. Again, it was harder to get FastPasses in some weeks in 2014, yet wait times are higher for the same weeks in 2015:
“But wait!” you say. “What if Disney increased the supply of FastPasses AND guest demand also increased? That would explain the higher standby waits.” We thought of that, too. We don’t think it’s happening. Here’s how we checked.
If Disney increases the supply of FastPasses at a ride, they increase the amount of the ride’s capacity that’s dedicated to FastPass+. Remember that the default is a 4 to 1 ratio, or 80% – the Castmembers told us that. So that hasn’t changed.
“But maybe management increased the supply and didn’t tell the Castmembers!” We checked that too. It doesn’t seem to be happening. Let’s say management decided to allocate 90% of Toy Story’s capacity to FastPass+. If that happened, just 81 people in line would result in a 60-minute standby wait. We counted the number of people in line, though, and compared it to the standby time; it’s the same 4:1 ratio as it’s been.
To finish off the Studios, here’s Star Tours. There were 6 weeks in 2014 where it was harder to get a FastPass+, yet wait times were lower:
Here’s Tower of Terror:
There are actually a few weeks here where you get something that kind of looks like a pattern. Late June’s FastPass+ availability and standby waits look the same for 2014 and 2015, and the high FastPass+ availability for the week of 8/30/2015 matches the lowest standby waits of the year. So that’s promising. But overall, FastPasses have been available consistently about 1.5 hours in advance since the end of June 2015, while wait times have fluctuated between 30 and 70 minutes, so that’s hard to explain.
We also don’t think there’s a clear relationship between FastPass+ availability and standby waits in any other Disney park. We’ll show that over the next few days.
FastPass Supply and Demand In General We also looked at the possibility of any other combination of change in FastPass+ supply and demand. On the supply side, there are 3 possible options:
- Disney has increased the number of FastPasses available
- The number of FastPasses available has stayed the same
- Disney has decreased the number of FastPasses available
The same things could have happened on the demand side. That gives us 3 x 3 = 9 possibilities.
To see whether any of them are true, it’s helpful to consider what would happen to the FastPass+ availability lines in the graphs above in any scenario. Here’s a chart showing what you’d expect to see with those lines, for all 9 possible changes in FastPass+ supply and demand:
We don’t see a consistent pattern in the year-over-year availability of FastPasses across DHS attractions. That means we can probably rule out all of the options except for the ones in the upper left (“higher supply/higher demand”) and lower right (“lower supply/lower demand”). But as we’ve shown above, counting the number of people in the standby line doesn’t seem to support either of those two scenarios. That’s why we don’t think there have been major changes to FastPass+ supply or demand.
Of course, Disney could be changing the FastPass+ supply dynamically every day (without telling Castmembers to change the FastPass line ratios), differently at each attraction, based on things like projected hotel occupancy or weather, in a way that produces the trend lines above. That would be really hard for us to see (and harder to prove). It would also be moderately complicated to implement, and I’m not sure what the goal would be there. But it’s possible.
So that’s where we’re at for now. Closed attractions and better FastPass+ systems seem to be the best explanation for Hollywood Studios wait time increases. We’re making adjustments to the models and calendars, and we’ll have updates out within a few days. Thanks for reading. See you tomorrow.
 In the real world, the actual number of people who ride per hour is almost always lower. Rides sometimes stop temporarily to let people on or off, or for temporary mechanical issues.
As long as we use the same attraction capacity for 2014 and 2015, the calculations work out the same. We used the same attraction capacity for 2014 and 2015 which means we are assuming the “real-world” capacity was the same in both years.
 Our crowd calendar is based on the standby wait times between 10 AM and 5 PM, at certain attractions in each park. We’re using the 10AM-to-5PM time period here for consistency.
 Backlot Tour was closed in February 2014 and February 2015.
 Our wait time modeling software is Treenet by Salford Systems. It’s a stochastic gradient boosting modeling engine. It’s more accurate than regression models, which we used in previous versions of the crowd calendar. Our lead statistician, Fred, came to us from Statistics Canada. Our other statistician, Steve, was a statistician at the Centers for Disease Control. That doesn’t guarantee they’re right all the time. But they have some idea of what they’re doing. 🙂
 What does the “J” in “Arthur J. Fonzarelli” stand for?