• Ben Johnson

The Audience Project: using audience behaviour to improve admissions and predict success.

Updated: Aug 28


TAP: using audience behaviour to improve admissions and predict success.

THE AUDIENCE PROJECT: A EUROPEAN MOVIE GOING DATABASE TO HELP THE INDUSTRY SELL MORE TICKETS


In 2018 Gruvi received investment from Creative Europe to start building The Audience Project (TAP), a database to collect behavioral data on cinema-goers taken from the various ad campaigns we run. This ever-growing data pool has enabled Gruvi to better target our ads on Facebook, Youtube, Google, etc in each country, exploiting relevant audience behaviours such as their tendency to purchase tickets for certain types of films.

KEY FINDING: targeting behaviour is 2 to 3X more effective than targeting purely by demographics or interest-based targeting available in Social Media

Gruvi has been in the film marketing game for 9 years. Over that time we have built a comprehensive marketing tech platform for the entertainment industry. Our tools are able to reach, track and record a global audience with intelligent, interactive advertising that encourages audiences to engage, while anonymously mapping their relevant behaviours. This data can then be used to retarget those users with relevant ads for other releases, and to prospect for 'new' audiences by modelling behaviour, such as the likelihood to buy tickets, etc.

TAP database collects and segments the audiences into clusters, identifying common audience behaviours and interactions that can subsequently be used to determine particular audience interests. With this ability, we can ensure we target contextually: delivering the right message to the right consumer at the right time, with ads that include the relevant point of sale.


In this post we will answer the following key questions, in order to understand the success of TAP:


1. From the millions of impressions generated during a campaign, how do we identify the potential buyer?


2. How do we leverage this data in existing and future campaigns?


3. How can we predict the success or failure of a film's opening weekend based on online behaviour?


For further information about the study CONTACT US


BEHAVIOURAL ANALYSIS: THE IMPORTANCE OF MEASURING ACTIONS TO MEANINGFULLY GAUGE AUDIENCE INTEREST


In these days of content abundance, there are more online ads promoting movies, TV shows, and video games than ever before. Most ad agencies still believe the most effective way to reach an individual user is simply by driving higher & higher ad impressions: the more spots they buy the more chances there are for that targeted individual to see and recall the ad.

The television industry has experienced such a surge in original content that annual cancellation rates have quintupled over the past 15 years (twice as many original scripted series were canceled last year than even aired in 2000). Since 1985, the indie film industry has seen a nearly twenty-fold increase in the number of theatrical releases even though ticket sales have remained flat (in 2014, the Head of SXSW’s film festival decried that “the impulse to make a film had far outrun the impulse to go out and watch one”).

Matthew Ball and Tal Shachar, AGE OF ABUNDANCE: HOW THE CONTENT EXPLOSION WILL INVERT THE MEDIA INDUSTRY

However, today’s online audiences are simply overwhelmed by the huge amount of content choice: a impressions based strategy no longer works when fifteen or more films are released in cinemas in a single week. This is why Gruvi focuses more on the significance of ‘Actions’ based reporting: we measure 'meaningful' engagement with our ads as an indication of audience interest. The Actions we record and highlight in our reporting include:


1. High retention or completed video views - does the viewer remain interested throughout watching the video, where does he/she drop off?


2. Post shares, comments, reactions - does the viewer share the experience with their friends?


3. Link clicks, location, cinema & showtimes interactions, and landing page exits to ticket purchases - does the audience member interact with the point of sale experience?


What we measure goes far beyond impressions based reporting. Reach is important, Frequency is also important but the clincher for a successful campaign orientates around what the audiences do, which is fundamentally linked to their audience behaviour. This understanding also allows us to run a campaign tactically, to make adjustments on the fly, while guiding the creative strategy. You can read more about our philosophy in a post and video dedicated to this subject, but the following quote sums up our thoughts on the industry currently:

TODAY'S REALITY: most agency media planning is rooted in a First World War-esque Strategy. Careful integrated planning conducted over many months, aiming at blowing the biggest budget available, in the shortest amount of time, with the maximum impact. Very little organic change happens during execution.

Basing your advertising around behavior is an important key to unlocking box office success. In the next section, we will discuss the results of a case study in the Polish market that comprised of a year's worth of advertising over many films.


CASE STUDY: USING TAP TO DEVELOP POLISH INDEPENDENT FILM AUDIENCES (WHO BUY TICKETS!!!)


The following case study shows the effectiveness of using TAP to exploit audience data collected from multiple campaigns across a range of different ad channels including Google, Youtube, Facebook, etc.


We started 2019 with the film “Everybody Knows”, directed by Asghar Farhadi and starring Penelope Cruz & Javier Bardem. During the first phase of the ad campaign to promote the release on Facebook we reached 578,278 people, with 5,070 (0.9%) clicking through to our showtimes & advance ticketing landing page.


We used TAP to collect and build a Custom Audience around first-party and second-party data to optimize the campaign. In the second stage of advertising, we looked to drive audiences from 'awareness' into 'engagement'. These custom audiences were used to identify 'look-a-like' audience groups (see below) that had similar buying behaviors, so we could also broaden the reach of who we were targeting effectively. During this second phase, we saw a dramatic improvement in user engagement: we reached 174,537 people, with 12.5% clicking through to our showtimes & advance ticketing landing page – an increase of more than 11% from the first attempt. The total number of visits to the showtimes and advance ticketing landing page totalled 51,000, and the film generated over 60,000 admissions by its second weekend on release.

A lookalike audience comes from modeling your custom audience to help find new people who are likely to be interested in your film or tv show because they're similar to your best existing audiences.

Over the course of the year, we continued to focus on building our database, which included audiences that:


  • interact with our ads to click through to our sites and ticket widgets - we refer to these as our Tier 1 Audience

  • clicked through to buy tickets - we refer to these as our Tier 2 Audience


The audiences were recorded against relevant metadata such as film title, genre, etc. We then used our Tier 1 and Tier 2 custom audience groups to continue to build lookalike audiences. Over time we were able to identify the audiences that paid to go see different genres of independent movies and use them in our campaigns to generate markedly higher levels of performance across various channels.


The success of using TAP can be seen in the campaign results delivered for subsequent releases from the same distributor in Poland. One notable example was "Midsommar" where we used the 2nd party data from the “Everybody Knows” campaign stored since March 2019 we were able to retarget audiences in the 40+ demographic. We reached 527,069 people via targeted ads, with 59,706 of them subsequently visiting the film’s showtimes website which represents an impressive 12% Click Through Rate Against Actions (the average CTR on a film campaign is between 0.5 - 1.0%).


We use several Key Performance Indicators (KPI) to monitor a campaign to assess its effectiveness. These are:

  • Action Performance Estimated vs Actual - the proportion actions delivered relative to what we anticipated on the media plan, described as a %

  • Action Rate Against Impressions - the proportion of actions (completed video views, clicks, etc) relative to impressions, described as a %

  • Click-Through Rate Against Actions - proportion clicks to the website measured against overall actions, described as a %

  • Click-Through Rate Against Impression - proportion clicks to the website measured against overall ad impressions, described as a %

  • Showtime Interaction Rate - the proportion of users interact with the showtimes unit, described as a %

The table below includes a summary of the KPIs described above to gauge how a campaign will actually impact sales at the box office. We condense box office into the measure called ‘Sales per Cinema’ (total number of tickets sold divided by the number of cinemas that screened the film) as this offers a better interpretation on the actual sales success of a film.



In the next section, we will examine how we use these KPIs to measure how the campaign is performing relative to A Sales per Cinema metric and to predict performance before the movie has released.

For further information about the study CONTACT US

PREDICTING BOX OFFICE SUCCESS BY OBSERVING BEHAVIOUR

We use our Action based reporting to look for the correlation between the campaign and its subsequent box office results. As early warning indicators into campaign health, we look at Action Performance Expected vs Actual and the CTR Against Actions at a later stage when the film is about premier we look at the Showtimes Interaction Rate. Overall it's important to look at the KPI group collectively as there are variables that impact interpretation (more on that later).


General indicators of film success in terms of cinema admissions are seen when:


  1. The Actions rate against impressions is higher than 10 to 15%

  2. The Click Through Rate against Actions is higher than 5 to 10%

  3. The Click Through Rate against Impressions is higher than 0.50 to 1%

  4. The Showtime Interaction Rate is higher than 10 to 15% on mobile


Warning indicators that the film may fail to reach admission target are when:

  1. The Actions rate against impressions is less than 9%

  2. The Click Through Rate against Actions is less than 4%

  3. The Click Through Rate against Impressions is less than 0.4%

  4. The Showtime Interaction Rate is less than 9% on mobile

To illustrate our findings the following table comprises the films we worked on in Poland in 2019 where budgets were sufficient to reach the target audiences. Ultimately we measured campaign KPIs and the films’ against Sales per Cinema. We also added colour coding to help highlight KPI indicators that exceed or fell below the performance tolerances mentioned above, green (when KPIs positively exceeded the tolerances) and red (when KPIs fell below the tolerances).


Source: https://docs.google.com/spreadsheets/d/1UYGBTP-cinE7CKeUOCyRuocFn_THLEe_P_YoUCTwzcg/edit#gid=140141


It’s interesting to note that films that performed well on 3 or more KPIs also performed well at the box office, with the exception of “Midsommer”. In that case, the campaign delivered a very impressive CTR Against Actions score, but this did not translate into impressive Showtimes Interactions Rates or Sales per Cinema. Initially, it seems that this audience’s interest in the campaign did not lead to a willingness to pay in advance to see the film.


A NOTE ON VARIABLES THAT IMPACT KPI INTERPRETATION

Individual KPIs and their interplay vary depending on the channels selected for advertising and how the advertising is run. The high performance of Actions (Action Performance Estimated vs Actual greater than 150%) relative to their estimates not only indicates a good performance on the campaign, it also dilutes the results for other KPIs like CTR Action Rate. A good example of this was for Extremely Wicked, Shockingly Evil and Vile where we achieved a near 800% higher action rate than was predicted. In this case, a 1% CTR Action Rate from 800% is comparable to a CTR Actions Rate of 10% from 100% estimate. Budget is another important variable, a low budget limits the reach of the campaign meaning the number of people we can reach within the target population will be smaller than required. In a future iteration of TAP we hope to compensate for this by designing dashboards that break down the campaigns by channel and by the KPI metrics above, in addition to other measures such as competing titles, demographics, location, time of day, creative used, weather, etc further segmentation, and analysis.


A strong measure of success is the Showtimes Interaction Rate. Once that crosses 20% we can see a strong correlation with positive Box Office returns, but there are some issues with this KPI to bear in mind. One is that older women do not tend to interact with the showtimes section of our tools. A low interaction rate is to be expected for campaigns that have a target audience of female cinema-goers age 50+ as was the case with "Midsommer", “Pavarotti” and “Red Joan” (even though Box office exceeded the distributor's expectations). The second issue is that most films only have cinema showtimes available a few days before their release, due to the way cinemas publish their program times. This is why during the pre showtimes phase of a campaign, we use the other KPIs and concentrate on the CTR Against Action Rate as the main indicator. When four KPI's align it's a very strong message for the distributor to push for more screens over the opening weekend and increase the budget. Conversely, if the results look poor across the board we will make the recommendation to cut back on the budget to save money for a future release.



For further information about the study CONTACT US


KEY FINDINGS


1. TAP, our Creative Europe funded Audience database, allows us to sort and filter potential audiences that engage with Gruvi campaigns into custom audience groups. These groups are stored against relevant metadata such as the movie titles, genres, etc in TAP's database for use in future campaigns.


2. Custom audience groups are audiences that have previously expressed a real interest in going to see a film. This data set offers real efficiencies in targeting proven cinemagoers with future campaigns. Custom Audiences also allow us to create ‘look-a-like’ Audience groups modelled on the most effective ticket-buying audiences, enabling us to broaden the effective reach of the campaign.


3. TAP also allows us to track KPIs to help predict campaign and box office success. The key metrics we analyze in the run-up to a film’s theatrical release centre on meaningful audience Actions such as. length of the video view, etc, and allow us to calculate the potential success of the campaign prior to the release of the film, no other effective measure exists in the market that is so audience-driven. For example the Showtime Interaction Rate has the highest correlation with Box office success at 0.8, so if rates exceed 20% this is an extremely good measure that the audience wants to see the film and are actively trying to buy tickets.


4. In order to give us the best view possible into better understanding and delivering on each campaign’ strategy, we look at the aggregate KPIs for that campaign before advising our clients on what next steps to take.

For further information about the study CONTACT US



HOW WE WORK
PRODUCTS
COMPANY
  • Facebook
  • Twitter