Ultimately in 2024, we will enter a data-scarce environment & VRMs must be cautious not to rely solely on “AI-powered” solutions as a magic bullet. AI can be a powerful tool but is not a substitute for quality data cultivation and effective direct booking strategies.
This is the sole reason OTAs are coming after your direct booking guests: to increase their customer lifetime value through personalized booking journeys every single time.
VRMs must prioritize data collection by investing in the skills and technology necessary to leverage that data effectively to differentiate themselves from their competition.
Here are 9 questions to ask any AI-powered vendor before you engage with them.
1. Does your AI have a data threshold to ensure it’s effective?
The answer should always be YES. AI requires massive amounts of data, so any vendor promising to deliver without asking, “How much historical & incoming data do you have” is likely exaggerating their solutions’ capabilities.
2. Is there a certain amount of website traffic needed for AI to learn?
This is a feeder of the first question. If it’s a website facing solution, you’ll have to surpass a traffic threshold, usually 5-10k monthly visitors, for the ML models to be effective.
3. Can you provide case studies of clients that use your AI-powered solutions?
We are in a word-of-mouth industry. Trust your peers and proof of concepts.
4. How does your AI know about a specific guest’s intent when data is scarce & web visitors are more anonymous than ever before?
With the internet transforming in 2023 & guest intent data cultivation diminishing in 2024 (read more here), they should outline their use of lookalike models & audience cohorting. You should increase suspicion if you don’t understand their answer because “AI-magic” is a fantasy.
5. How do you integrate with our current property management system (PMS), and what level of support do you provide during the integration process?
Question them on their knowledge of specific PMS systems & their API relationships. If they beat around the bush, they’re hiding something.
6. How do you measure the success of your AI-powered solutions and what level of reporting transparency do you provide? Do you have an example report or dashboard?
If they aren’t A/B testing, they aren’t confident in their product. Revenue attribution is extraordinarily challenging, so if you’re not confident in the vendor’s reporting methods, don’t move forward.
7. Can you explain how the models learn and improve over time? Do you tweak them for specific clients?
There’s no such thing as “set it & forget it.” Models require tweaks because every VRM is unique, which leads to the question below.
8. Do you share or commingle data with other VRMs?
For the most part, the AI or ML engine should be siloed to your specific business. It must understand your core business & customers. Imagine if your social media feed used random people’s affinities… you’d likely stop using that app.
9. Who gets to keep the newly cultivated guest intent data?
100% of the time, it should be the VRM. This shouldn’t even be a conversation & you should request a mutual NDA be signed to protect this point.
In case you missed it:
