What is the logic behind the dynamic caching system?
The new caching system uses machine learning and performance metrics to adjust cache durations dynamically:
Based on historical data, an initial cache duration is determined using machine learning.
This duration is regularly updated (hourly) according to performance metrics.
The key metric compares sampled prices from cached data against prices returned by your API to ensure accuracy.
The overall goal is to reduce traffic to downstream suppliers while maintaining data quality and price accuracy.
Are there any changes to the Booking page caching?
No. There are no changes to the Booking page. The current caching time remains 10 minutes.
How does dynamic caching impact the caching of empty responses?
There is no current impact on caching for empty responses. However, we are working to make this cache setting more dynamic as part of ongoing optimisation efforts.
Will the change affect traffic to my API?
You may see some variation in API call volume between Skyscanner and your systems, depending on how frequently you refresh your quotes. Overall, the update is expected to reduce live API calls while maintaining price accuracy and improving performance.
Will this impact my results on Skyscanner?
There will be no change to your impressions or exposure on Skyscanner. The rollout only affects backend caching behaviour.
Do I need to make any changes to my API integration?
No change is required to your API integration.