1) What is the current price accuracy threshold?
Our current acceptable threshold for measuring price accuracy differences is from -25 EUR to +7 EUR. Skyscanner accommodates price accuracy variations through the categories of good, acceptable, and bad accuracy.
-
Good accuracy: This is defined as any price displayed on the partner’s site that is within the range of -1 EUR to +1 EUR of the price shown on Skyscanner’s booking panel.
-
Acceptable accuracy: A range that includes a deviation of -25 EUR to +7 EUR from the initial price shown on Skyscanner's booking panel versus the price travellers see on the partner’s site once redirected.
-
Bad accuracy: Any price change that exceeds the acceptable range, encompassing price deviations beyond -25 EUR to +7 EUR from the price initially shown on Skyscanner's booking panel versus the price travellers see on the partner’s site once redirected.
2) Why are we changing the price accuracy threshold?
Price Accuracy is fundamental to maintaining customer’s trust in Skyscanner and driving conversion rates. Skyscanner is always looking for ways to improve both the user experience and partner performance. The current acceptable threshold of -25 EUR to +7 EUR was established during the Covid period. Over the last year we have started to observe more how our travellers respond to price changes between Skyscanner’s booking panel and our partner’s site once they have redirected. It is apparent from our data that even slight price changes affect travellers' trust and likelihood to make bookings.
By taking the median effect on partner booking rates for different price change ranges we can see below that the percentage of lost bookings correlates directly to price percentage changes.
The graph below further demonstrates that even a slight price change can lead booking losses.
3) What exactly is changing?
To improve the impact price inaccuracies are having on traveller bookings, we are changing the acceptable price accuracy range. This metric will now be defined by two criteria: a threshold of -4% to +1% and the existing threshold of -25 EUR to +7 EUR. The acceptable range will be determined by whichever of these criteria is met first.
As such, the new metrics will be as follows:
-
Good: Price displayed on partner’s site when redirected is within the range of -1 EUR to +1 EUR from the initial price shown on Skyscanner's booking panel.
-
Acceptable: A range that includes a deviation of -4% to +1% or -25 EUR to +7 EUR (whichever comes first) from the initial price shown on Skyscanner's booking panel versus the price travellers see on the partner’s site once redirected.
-
+1%
-
The negative impact on bookings remains noticeable even for low value tickets as travellers' sensitivity to price changes are heightened when dealing with tickets that are of lower value.
-
Adding a +1% threshold in addition to the existing +7 EUR threshold will account for nearly 95% of loss booking associated with price changes.
-
-
-4%
-
The data analysed suggests the current -25 EUR threshold remains acceptable and practical and found that -25 EUR corresponds to approximately -4% of the average ticket cost observed in all price accuracy events.
-
This decision is based on the observation that negative price changes have a lesser impact on bookings.
-
However, our primary concern is still the impact on travellers when making bookings and we are not ignoring the potential harm caused by unrestricted negative price changes. Also, such price changes would also suggest that partners may not have been as competitive by displaying higher prices on Skyscanner's booking panel, potentially resulting in redirect loss.
-
Therefore, incorporating a -4% threshold alongside the existing -25 EUR threshold will effectively manage this situation.
-
-
-
Bad: Any price change that exceeds the acceptable range, encompassing price deviations beyond -4% to +1% or -25 EUR to +7 EUR from the price initially shown on Skyscanner's booking panel versus the price travellers see on the partner’s site once redirected.
4) How will the new threshold work?
By introducing an acceptable threshold range of -4% to +1% in addition to the existing range of -25 EUR to +7 EUR, we will assess whether the price is within the acceptable bounds using the first applicable threshold.
For example:
5) What is the expected impact of these changes?
Our global target of maintaining below 5% price inaccuracies remains unchanged. With this adjustment, we expect a subset of redirects which presently fall within the acceptable threshold of -25 EUR to +7 EUR to be reclassified as bad price accuracy as the threshold becomes more stringent.
We will temporarily disable machine learning automatic quote suppression for a month as we expect bad price accuracy to increase. During this time, quotes will not be suppressed even if a partner's bad price accuracy is above 10%. This change will allow partners time to work towards reducing their price inaccuracies to achieve a score of below 5%, aligning with our 5% target.
Comments
0 comments
Please sign in to leave a comment.