Decision to Leave
Ben Johnson
Updated on 07/12/2023

In August 2022, we ran a campaign in The Netherlands for a top-rated movie that had a strong presence at the Cannes Film Festival. This was both a traffic-focused and an awareness-raising campaign, and the advertising platforms we chose were: Facebook, Twitter, Instagram, and YouTube.

This campaign was boosted by findings from our advanced audience research team at Gruvi Analytics. Gruvi Analytics strives to help filmmakers, producers and distributors make informed decisions for their movies by relying on social media listening and advanced audience analysis.

Through Gruvi Analytics, we were able to retrieve key data about the target audience interested in this particular film title, such as:
- Their age and gender
- Their location (by city)
- Their preferred social media and content platforms (e.g. Snapchat, Discord, FB, IG, Twitter, etc.)
- The language they speak/use the most on social media
- Which days of the week and times of the day they’re most active online
- The keywords, hashtags and emojis they use the most (for marketing purposes)
- Their preferred TV Shows, TV channels, Radio Stations, Newspapers, Magazines, Blogs/Websites, and top-followed Events (such as Festivals).

Based on the information above, we came up with high-precision targeting criteria that allowed us to not only narrow down and locate our target audience, but also make a few strategic decisions at the early stage of campaign planning:

  1. Since findings from Gruvi Analytics showed that 40% of our target audience uses English as their main language on social media, we decided to run two versions of the creative materials: one in English and one in the local language (Dutch). To be noted, this is not a commonly used approach as most distributors tend to focus on the local language only, which may exclude expat communities and non-speakers, so our goal was to test whether a bilingual approach would yield better results.
  2. To allocate a portion of the advertising budget to Twitter, where we saw a considerable community of people interested in the film title. Twitter is frequently ignored by advertising agencies as there is over-reliance on Facebook and Instagram, so we wanted to test how the campaign’s performance on Twitter would compare to these other platforms.
  3. We created two sets of audiences on Twitter:
    1. Custom Audience - this one was created by first identifying users on Twitter who have spoken about or interacted with the title before. We then exported the list of these usernames and created a custom audience list on Twitter. Next, we used Twitter’s algorithm-powered recommendations to arrive at a wider audience of people who Twitter considers to be similar to the users in the original list.
    2. Interest-based Audience - by looking at the demographics, location and interests of our target audience, we created another set of targeting criteria based off our in-house proprietary research to reach people who shared the same characteristics and behaviour on social media.


  1. Language

Our unconventional decision to run a bilingual campaign yielded positive results. The English version performed better from a click-action perspective, while the Dutch-version creative had better video retention. 

Although the difference between the two versions was not a considerable one, the findings suggest that it is important to invest in audience research to identify the important nuances that can otherwise get missed. For example, if the target audience is bilingual, consider running creatives in both the local language and in English so as to not exclude important parts of the population from your marketing.

pasted image 0
  1. Ad groups

The two sets of audiences performed similarly in terms of impressions, but showed different behaviour on other metrics. Interest-based audiences overperformed in CTR, while the Custom Audience had a stronger Video retention rate.

The Custom Audience suggested by Twitter’s algorithm was more interested in viewing the video creative during a longer portion of time and was less active in clicking on the landing page link to learn more about the title and buy the tickets.

On the other hand, our Interest-based Audience was much more likely to interrupt their video viewing in order to click on the landing page link and buy the ticket. Therefore, they showed higher % of clicks and CTR compared to the Custom Audience, which was more focused on engaging with the creative.

As a result, we could see significant differences in campaign performance for these two sets of audiences, particularly when looking at results per 1000 £ spent, where - all else being equal and the two audiences showing roughly the same number of impressions - the Custom Audience suggested by Twitter overperformed by 47% in video views, while our research-powered Interest-based Audience was much more action-oriented and overperformed by 47% in clicks with a 22% higher CTR. 

Twitter’s suggested Custom Audience was also 50% more expensive in Cost per Click compared to our research-powered Interest-based Audience.

L5NLYdWZXy5jtoAmhn4nMBysuY2NEubNzEv23Dlc7uUbe7vVp4gRYSFrR9CqV2xGHXT7Nga6WrmlTPjxskFWsh xwHypU1mRCwTY1LEmp V1O5nbdY

Our conclusion is that advertisers who aim at targeting an action-oriented audience with the lowest Cost per Click should invest in audience research and creating Interest-based audiences on Twitter. Interest-based audiences are likely to significantly overperform Twitter’s suggested audiences in terms of clicking on the ad and engaging with the landing page, which increases the CTR and helps reduce the Cost per Click.

  1. Results & showtime exit rates

In this campaign, Twitter has proven to be a reliable marketing channel, and also a relatively inexpensive one. Twitter helps to expand the target audience and to generate good campaign results at a reasonable cost. As shown below, this platform delivered the second-best result on clicks per 1000 £ spent, and also in showtimes exit per 1000 £ spent.

pasted image 0 1

  1. Historical Benchmarking

Twitter ads used in the course of this campaign outperformed all of our internal benchmarks across previous campaigns. 

In terms of ad interactions when looking at all campaigns we previously ran in The Netherlands across all genres, this campaign showed a 73% higher CTR. We could see that this campaign successfully targeted more click-oriented people than the benchmark on previous campaigns.

A WpqDBKYji28hZx9D2yX2evR9TugWD0NzkIbp3Q39E8qvRJV2Zh QPds6UaSn0l b4YO7wfCNf Mm8SBuwjduZgh1wTfa1cb4TEHHc6JIvX 1

Next, we decided to break down the performance of our research-powered Interest-based Audience versus the Twitter-suggested Custom Audience. When considering all campaigns that targeted the same genre worldwide, we could see that the Interest-based group overperformed by 49% in terms of CTR and achieved an impressive 3x the benchmark (310%) leg up on the Link click / Actions ratio.

Looking at all campaigns that ran in The Netherlands across all genres, the Interest-based group overperformed by 91% in terms of CTR and by 48% on the Link click / Actions ratio.

DzaupgooCsF76iFWpcOQ3hZ DXqKvO8 QI 0ps8tVQevf ZaK0zR6cglrqmnTGJ8Vv9tBxp1l4btLIp1N0RVe4a7Mw01Kzk5siektIabw2zeq0j5GuUpV7nqiW9rZpVnBFC1r1v07lWmaNHX 7ktMBw67LJENnGJiRWDeKFs7MYJeLwT 3gtXzB0 Q

Regarding landing page interactions, and even though the campaign ran during the slowest period of the year (August / early September), it outperformed all the internal benchmarks we have on record until this day. When looking at all campaigns we previously ran in The Netherlands across all genres, this campaign showed an 8% higher showtimes exit rate and interactions rate. Considering landing page interactions for campaigns we previously ran in all countries across the same genre, this campaign showed a 14% higher showtimes exit rate and a 23% higher interactions rate than our benchmarks.

jLEMMi7KyNHDQCSVff2u 8F0VC9KqQwOfZ2jDsSBAKzoyU4godx62BiF8HBrx2 xesqhRMsibmRbyVRY 33rFBdHfHDKr3lbLU9MAQag8MFj6xNrcT2tsS Py2uEIdgCqRLlyD4p3p2bipbrD0ojcEIniVjx4FffQVY lli8x I95ASKIFzj3p4lcg


  • Twitter audiences outperformed all our benchmarks across all channels on record when it comes to interactions with the landing page, proving that Twitter is a reliable advertising channel to use even in markets where other platforms might have a stronger presence (e.g. Facebook, Instagram) and in slower periods of the year (August/September).
  • Cost-wise, Twitter performed well compared to Facebook, signaling that distributors should consider allocating a portion of their advertising budgets to this platform. Furthermore, Twitter was the second cheapest channel when looking at results per 1000 £ spent, proving that it is an effective and inexpensive advertising channel.
  • In this campaign, Twitter’s algorithm has proven to be less reliable when suggesting target audiences than our audience research-powered recommendations. Interest-based Audiences built off our prior research scored a massive performance over the audiences suggested by Twitter’s algorithm in terms of both the CTR and the Click links / Actions ratio. This underlines the importance of investing in audience research prior to campaign planning to ensure the most high-yielding and cost-efficient performance.

The English-speaking segment of the target audience also performed exceptionally well, suggesting that distributors need to research the languages spoken by their target audiences and consider running campaigns with bilingual materials, especially in markets with strong foreign communities.

Share this article:
Decision to Leave