"Tracking" Doesn't Measure Ticket Sales. It Measures Memory, Interest and First Choice — Weeks Before Anyone Buys a Ticket

This weekend's two big openings arrived carrying three different forecast ranges apiece — this library printed all of them, unmerged, in the Godzilla piece below. That spread isn't a scandal. It's what happens when an entire industry prices a Friday using surveys of people who, at the moment they were surveyed, had bought nothing, seen nothing, and in many cases heard of the film for the first time from the survey itself. That's what "tracking" is. Here's the machinery, from the services' own definitions and the trades' reporting on them.

The four numbers, defined by the people who sell them

The vocabulary comes from National Research Group (NRG), for years the field's defining firm. In NRG's own tracking primer — an internal methodology document, cited here for its definitions — data is collected by survey every week, Monday through Sunday, with comparison tables delivered Mondays, and a film's position is broken out by weeks before opening. The four core metrics:

• Unaided awareness — the percentage of respondents who name the film unprompted. Top-of-mind. NRG states it correlates well with opening-week box office.
• Total awareness — unaided plus aided awareness: respondents who recognise the film when shown a list of titles and stars.
• Definite interest — the percentage of those already aware of the film who say they're definitely interested in seeing it in a theater. NRG itself flags the trap in the denominator: because it's measured only among the aware, a barely-known film can post a spectacular definite-interest score on the strength of its few early fans.
• First choice — respondents are shown every film opening over the next nine weeks and may pick only one. A major title entering tracking therefore "eats up" first choice from everything else on the calendar — one film's rise is mechanically another's fall, with no ticket sold in either direction.

What the surveys are, and what they're for

The trades' reporting on method is older than the current cycle — tracking is a secretive business, and the named sources here are dated, which we flag rather than hides. Screen Daily's 2008 account: a film enters the surveys about six weeks before release, the surveys lean on heavy cinemagoers, and the pre-opening forecast carries a stated 15% margin of error. TheWrap's 2010 anatomy of the business names the era's major firms — NRG, OTX and MarketCast, with Screen Engine entering under former OTX research head Kevin Goetz — and describes the polling: NRG multi-mode (phone, internet, in-person), MarketCast online with a weekly sample of 2,000 frequent moviegoers.

And then the quote that reframes the whole product, from Vinny Bruzzese, then president of the worldwide motion picture group at OTX, in TheWrap: "The real use of tracking is to measure the effectiveness of the marketing weeks out from release, so you can make adjustments while there's still time." The opening-weekend prediction, in his words, is a "parlor trick." The forecast the headlines trade in is the exhaust of a system built to tell a studio, while it can still act, whether its advertising is registering — awareness — and whether it's converting — definite interest and first choice.

The famous forecast misses are part of the record

The same sources keep the receipts. TheWrap (2010): Sex and the City 2 forecast around $60 million, opened $36.8 million; Shrek Forever After tracked near $90 million, opened $70.8 million; The Karate Kid beat the top of its range by roughly $20 million. Screen Daily (2008) holds the widest miss on the books: The Passion of the Christ opened to $83.8 million against tracking estimates of $15–30 million — a figure worth noticing this weekend, since that same film is one of the three non-English No. 1 openings our Godzilla coverage keeps citing. Family titles break the models in both directions: Alvin and the Chipmunks doubled its estimates at $44.3 million; The Golden Compass undershot a $30–40 million forecast at $25.7 million. Two structural reasons recur: at near-100% awareness, a tentpole's tracking stops discriminating — everyone has heard of it, so the survey can't price it — and definite interest measures stated intent, which OTX's Bruzzese was already warning, in 2008, doesn't guarantee attendance.

PostTrak is exit polling, not pre-release tracking

PostTrak isn't tracking. It's exit polling, run by Comscore with Screen Engine/ASI, surveying audiences during a film's first two weeks of wide release — demographics, what triggered attendance, and sentiment scores (Comscore's page cites a 74% "total positive" norm). Tracking asks people who haven't gone; PostTrak asks people who just did. When a headline cites "audience scores" beside "tracking," it's stapling two different instruments together, and the staple is where the error gets in.

This weekend's forecasts will resolve the way they always do — into an audited gross that flatters one model and embarrasses another. Our practice, set out in the pieces below, follows from everything above: print the range, name the source, and grade the forecast after Sunday — never average three surveys into a number none of them published.