Driving Precision:How Mitsubishi elevated their marketing with Meta CAPI
by Hazem Alsadi – Platforms Specialist @ Annalect
With an automotive story that began in 1917, Mitsubishi Motors is a renowned Japanese automobile manufacturer with a legacy of innovation spanning over a century. Mitsubishi Motors Europe is based in the Netherlands and it serves as the European Headquarter for Mitsubishi Motors Corporation. Known for its wide array of eco-friendly technologies, reliable and efficient cars, Mitsubishi is on a mission to build a better future by delivering greener, safer and more affordable modes of transport, all while supporting environmental sustainability
Contextual background
Driven by its commitment to continuously elevate its marketing operations, Mitsubishi sought to enhance data accuracy and achieve better ads personalization on Meta campaigns. To bring this vision to life, Mitsubishi partnered with Annalect to set up Meta Conversion API that could help to realize Mitsubishi’s objective of creating impactful and precise marketing efforts.
Striding towards first party data strategy with Meta Conversion API
Facebook Pixel is one of the primary marketing tools that Mitsubishi uses to track users’ key interactions on its website such as applying for a test drive, receiving an offer/quote, sign up forms and completing car configuration steps. In other words, Mitsubishi is keen on tracking and reporting on its full marketing funnel to cover each and every step of the customer journey. Depending solely on Meta Pixel to keep up with the latest development in the ad is bringing numerous limitations. These limitations are exemplified by data connectivity issues between browser and Meta, privacy regulation that is gradually leading to less performance insights due to browsers blocking third party cookies and hence missing data points. Consumer privacy is no longer a surging trend, and it becomes a must have in today’s marketing activities. To address this issue, Annalect recommends server to server integration, also known as Meta Conversion API. This innovative solution enables advertisers to share web and post-conversion data directly from Mitsubishi’s servers to Meta’s servers and give Mitsubishi the control to decide what data is shared at which point in the journey. The server-to-server integration bypass reliance on 3rd party cookie and thus preventing signal loss that otherwise will occur with 3rd party cookies. Together with PHD, Mitsubishi media agency, Annalect advised Mitsubishi to track the majority of its key user interactions using the Conversion API.
KPI’s & event tracking
After a thorough data analysis and a successful alignment with the PHD team, Annalect set up Meta Conversion API on a list of 7 key events as shown below. These events were already tracked via Google Analytics 4, but partially tagged via Facebook pixel.
| Event names | Description |
| aanvraag_proefrit | Test drive request |
| brochure_download | Download brochure |
| nieuwsbrief_inschrijving | Register for newsletter |
| aanvraag_offerte | Request an offer |
| start_configuratie | Start the car configurator form |
| complete_configuratie | Compete the car configurator form |
| aanvraag_contact | Request to contact for more info |
Redundant setup and deduplication
To achieve accurate tracking, Annalect always recommends using both Meta Pixel and Conversion API. However, simultaneously tracking the same conversion events with both tools, it’s essential to ensure that the same event is not recorded twice, which would falsify the data. To prevent this from happening, Annalect relied on generating and passing unique event identifiers for each conversion event tracked simultaneously via Meta Pixel and the Conversion API. Passing these identifiers will allow Meta to recognize, compare and keep only one instance of the matched event. The deduplication set up helps us to avoid double-counting conversions and ensure that Mitsubishi’s tracking remains reliable and accurate.
Event enrichment
To enrich data quality and make sure that Meta targeting capability reaches its utmost relevancy and efficiency, Annalect aligned with Mitsubishi’s IT Team to push customer information (see below) in the website’s data layer structure. Customer information here represents the first party data that is very important to utilize to improve Event Match Quality score (EMQ). Based on the quality of customer information that is passed with each server event, Meta provides a matching score. Meta recommends achieving a score of at least (5.8/10). This in turn will help to activate Meta’s matching technology. When matching is activated based on a good EMQ-score, Meta will connect an event fired on Mitsubishi’s website to an actual Meta user account. It is important to remember that sensitive customer information will be hashed out before sharing it with Meta. This means that information will be coded in random numbers and letters, and it will not lead to the identification of a natural person which will safeguard user privacy.
- Customer Email (Hashed)
- Customer Phone number (Hashed)
Technical setup
Mitsubishi Motors in the Netherlands relies on Google Tag Manager as its main tag management system. Based on the discussions with Mitsubishi’s IT Team, Annalect advised Mitsubishi to leverage its existing Google Tag Manager infrastructure to implement Meta Conversion API. Several preparatory steps were followed to kick off:
- Setting up Google Cloud Account
- Creating a custom subdomain on Google Cloud Console
- Setting up server-side GTM container and connect it to Google Cloud Project
Baseline setup & Initial results
Annalect finalized the baseline setup of CAPI configuration on August 12th. As shown below, the EMQ scores were very low in the first 2 days due to some missing events parameters. As mentioned previously, the Event Match Quality score should be at least a 5.8 /10 to help Meta match a user activity on a website to an actual Facebook account.
After troubleshooting the server events and configuring the missing customer parameters, EMQ scores started to surge. We can see below a performance overview of all the events as taken from the Meta Business Manager.
Compared to the initial set up results, Annalect saw excellent progress in the quality of events passed to Meta. Ensuring the server events were enriched with impactful parameters and securely integrated into the system was very rewarding, as the results exceeded our expectations.
Pre and post CAPI measurement
After the initial technical setup and optimizations, Annalect partnered with PHD to align on a pre and post CAPI media test. The focus in this test is on the most important KPI on Mitsubishi’s website, namely Aanvraag_proefrit (Test Drive Request). PHD started to use CAPI events on August 16th and reviewed pre implementation results as well as post implementation results – after 4 weeks.
| 16th July – 14th Aug 2024 (Pre Conversion API) | 18th Aug – 16th Sep 2024 (Post Conversion API) | ||
| Ad set | Results | Ad set | Results |
| ASX PHEV | 11 | ASX PHEV – CAPI | 9 |
| Eclipse Cross | 8 | Eclipse Cross – CAPI | 12 |
| COLT | 5 | COLT – CAPI | 8 |
| COLT – zero | – | COLT – zero – CAPI | 11 |
| Total results from 4 ad sets | 24 | Total results from 4 ad sets | 40 |
66% increase in the number of additional conversions reported
Based on the table above, we can see that compared to a similar campaign that was launched prior to CAPI, the number of submitted Test Drive Requests jumped to 40 leads post CAPI implementation. It’s worth noting that the same budget and nearly identical audience attributes have been used in building the campaign. However, we can observe the significant impact achieved through Meta CAPI. Similarly, we have monitored the average Cost Per Lead during the first two weeks of using Meta CAPI. We were surprised to find that the average Cost Per Lead in those two weeks outperformed the all-time average Cost Per Lead during the entire Q2 of 2024. In other words, CAPI reduced the cost per lead by approximately 13.60% in the first two weeks.
During the last two weeks of Meta CAPI, Cost Per Lead dropped by 63%, recording an unbeaten milestone compared to the full Q2 of 2024. With this unsurpassed success in terms of cost reduction, we can safely say that Meta CAPI saved 63% of the amount of money Mitsubishi pays for every submitted Test Drive lead.
Conclusion
Meta Conversion API brought a significant transformation to Mitsubishi’s social strategy for Meta. By leveraging the power of first party data and server-side integration, the brand overcame many limitations on the browser and achieved accurate tracking and targeting relevancy. The enhancement in its personalization has led to a dramatic reduction in cost per lead.
“We recently started Meta CAPI from Mitsubishi Netherlands. The first results are very promising. In particular, we see a clear decrease in costs per lead. And that is something that we as advertisers always find very pleasant”
Robbert van den Heuvel
Manager Marketing Communication – Mitsubishi Netherlands
“Implementing Meta CAPI with Annalect has significantly improved our social advertising performance. Including but not limited to more accurate tracking, larger retargeting audiences and a lower cost per action compared to the pre-CAPI period”
Yannick Roosendaal
Social Lead – PHD Netherlands.