You may notice some differences between the performance numbers shown in your new Campaign reporting dashboard and those previously shown in your Messages v1 (classic) dashboard or other channel-specific legacy reporting dashboards.
This is expected and reflects an important improvement in how we attribute value to your campaigns.
Why Are the Numbers Different?
In short: we’re now using new, more advanced attribution models.
Previously different Triptease products (such as Messages or Metasearch) used different methodologies for tracking bookings. This meant:
Cross-product interactions (e.g. Messages and Metasearch influencing the same guest) were only taken into account for the acquisition channels, not the Messages channel.
If multiple interactions occurred in one channel, attribution for that channel followed a last-click model. Previous interactions were ignored.
For example, in the Messages v1 (Classic) dashboard, if a guest interacted with several messages before booking the last message clicked is credited with the booking.
As another example, if a guest interacted with a message and with metasearch, in the channel-specific performance dashboards both channels would independently claim the booking.
What’s new in the Campaign Reporting Dashboard
The new dashboard uses a new guest journey-based attribution model. This now take into account every Triptease interaction that influenced a booking - across all products, channels, and campaigns.
It captures the full value of your marketing activity across every Triptease channel, giving you the insight you need to optimize campaigns and understand your guests’ complete journey to booking.
So, if a guest saw an on-site message, clicked a Metasearch ad, and later booked after clicking a Paid Search ad, all three interactions will be recognized. This allows us to provide a more complete picture of how your marketing efforts work together to drive direct bookings.
As a result:
Depending on the attribution model you are using, attribution for some bookings will now be split across multiple campaigns or channels - helping you understand how channels interact to influence guests at multiple points along their booking journey.
You may see slightly different total booking numbers for individual campaigns or messages compared to your old channel-specific dashboards, because we’re now representing outcomes in a more accurate way.
Overall, this gives you a more realistic and holistic view of your marketing performance.
Attribution models available in the new dashboard
To help you make sense of your marketing performance with this new guest-journey based attribution methodology, we're providing three attribution models for you to choose between to suit your needs and preferences. These different attribution models will help you explore different perspectives on your data.
Here’s what each one means:
1. "Weighted" attribution
In this model, each influencing campaign or channel shares credit for a booking. The amount of value attributed to each channel is proportionate to the number of interactions each channel has had with the guests.
For Retargeting, 'views' are weighted at 1/10th of the value of a click in the algorithm to reflect the fact that retargeting views are generally considered to be less influential overall compared to when a guest directly clicks on an advert.
Here is a worked example to explain how this attribution model works:
If a guest clicks a meta link twice and and a message once, meta would claim 67% of the value and messages would claim 33% of the value
If a guest clicked two different messages from two different campaigns, each message campaign claims 50% of the value.
If a guest clicks paid search once, meta once and a message three times, then paid search and meta each claim 20% of the value, with messages claiming the remaining 60% of the value.
Best for: Seeing a balanced, holistic view of performance across all your marketing channels. This is the recommended model if you want to understand the full guest journey and how all your marketing activities contribute to conversion.
2. "Last click" attribution
This model attributes the booking to the final interaction before the guest booked, considering all Triptease channels. I.e. the final Triptease interaction before the booking receives 100% of the credit.
For example, if a guest interacted with Paid Search, Meta, and then Messages before booking, the booking will be attributed to Messages.
Best for: Identifying which campaign or channel typically provides the final push to conversion.
3. "Single channel" attribution
This model shows the total influence of each channel, regardless of whether other channels also influenced the guest before booking. This model is most similar to how the legacy platform works, and how most other vendors and agencies attribute bookings to marketing campaigns.
For example, if a guest interacted with Paid Search, Metasearch, and then Messages before booking, the booking will be attributed to each of Paid Search, Meta, and Messages, i.e the booking has been counted more than once.
Best for: Comparing results with your legacy dashboards or other external analytics tools, and for judging the overall impact of a channel in isolation.
Summary
Model | What It Shows | Best For |
Weighted attribution | Distributes booking credit equally across all influencing campaigns and channels in proportion to the number of interactions per campaign and channel. | Understanding the full picture of your marketing performance across channels. |
Last click attribution | Attributes bookings to the final influencing campaign. | Finding out which was the last message guests saw before booking. |
Single channel attribution | Attributes bookings to every channel that has influenced them. | Understanding the total influence of each channel on direct bookings.
Comparing with old dashboards or external benchmarks. |
Note: Attributed booking volumes for Metasearch, Paid Search and Retargeting may appear lower in the Campaign reporting dashboard than in legacy channel-specific dashboards as a result of these new attribution models. Actual performance has not changed.