How many AI generated experiences do consumers want?
January 11th, 2023by Alex Bainbridge
Is it thousands per city like today? Or is it tens of thousands perhaps? Or could leading retailers just have 10 experiences per city that are each highly configurable & personalised?
Today’s upper bounds example is Rome on Viator, where Viator has 7000 things to do just in Rome and the surrounding area. I am old enough to remember when Viator was sold to Tripadvisor, at that time they had a selection of about 10,000 heavily curated experiences globally, so 7000 just in one region shows the extent their catalog has scaled up.
Yesterday, speaking at the Skift Megatrends conference, Johannes Reck (GetYourGuide) suggested there is still upward growth to come:
“The catalog will just have to be a lot bigger. We estimate there are roughly 1 million experiences today. There will be many millions of experiences in the future”
Can tech handle this catalog growth?
If you increase to 20,000 experiences per city (as might happen with generative AI) – it breaks a lot of tech used in the sector just due to the data size required and the manual data management processes that still exist.
Additionally, discovery user interfaces will be overwhelmed, primarily due to too much choice between indistinguishable products.
Curation doesn’t help much as curator A may want subset 50 of 20,000 in a city, curator B may want subset 75 of 20,000 in a city, and there may be no overlap. So the OTA itself still needs the complete catalog (if trading with a multi-curator model).
Is the Uber ridehail model an alternative?
With Uber, they don’t have a selection of 10,000 A>B taxi routes in each city and then ask the consumer to select from a long list. Instead you as as consumer just say where you are and where you want to be, and the product is generated & priced at that point. i.e. it is effectively one route product and the main selection you are presented with is what kind of vehicle you want.
Generative AI can enable this for us, in local tours & experiences. Tour itineraries can be generated on demand – a consumer may select a “food tour” theme or a “history tour” and then adjust what is proposed by the AI.
However this breaks how OTAs currently enable discovery, distribution & operation.
The tour guide challenge
Ultimately, the challenge with both approaches is that operating (with only tour guides) is not really possible:
- Either tour guide prices must go up (as group sizes will go down as products become more long tail)
- Or you digitally tour operate part of the experience (more on this strategy in the coming weeks)
Any thoughts on what happens next?
Image: Dall-E AI
This content is protected by copyright. Link sharing is encouraged but duplication and redistribution is illegal
If you prefer to comment on LinkedIn, here is the thread