Reverse Dynamic Pricing
And why DoorDash should not outsource their PR to the free version of ChatGPT.
Happy New Year!
During the holidays, I started drafting what was supposed to be the below on dynamic pricing and loyalty. Then this weekend, the “DoorDash gate” happened, and because it was so on point with the topic, I had to modify my intro!
If you are not chronically online, you might have missed this Reddit post blowing up on both Twitter and Substack. In a nutshell, an alleged developer from a delivery app (no name was given in the post) revealed all the dirty secrets of their employer and how they make money by abusing the system and their users.
When I first read the post, I did not really pay attention to it because nothing appeared more disturbing than usual: dynamic pricing already made headlines with Wendy last year, and Instacart was also pinned down for changing prices based on who you are. As for the crude terms used to qualify the user base, any hospitality worker has heard way worse in the course of their first service.
What keeps me going back to it is the reaction it provoked when it blew up. Numerous Twitter accounts started defending DoorDash, saying the Reddit author was a scammer. Then both DoorDash co-founders themselves went on Twitter to say it was not true.
There is not much that you can learn from a French politician except one thing: if they take a public stand to say something they are accused of is not true, then it’s definitely true! No innocent man gives attention to something they are not guilty of.
Back to my original draft, it’s aligned with dynamic pricing through the lens of the end user and why, as a hotel, you should use loyalty as a growth lever instead of trying to squeeze a few bucks from a customer.
This entire post is longer than usual due to a now very long intro.
Hotels are pouring millions into AI-powered dynamic pricing to squeeze every last dollar from each guest. What they haven’t noticed: Their customers is going to have soon the same type of weapon in their hands.
2025 was the agentic year. Every hospitality tool embedded AI functionality with varying degrees of success. Booking and Expedia released their travel itineraries, TripAdvisor integrated with OpenAI, and DoorDash launched Zesty, its AI concierge. On the back of house, Mews PMS made strategic acquisitions while Duve and Canary raised significant funding to accelerate AI-powered guest apps.
Despite the noise, consumers aren’t seeing much results. The same way people didn’t flock to book stays through TikTok (remember those predictions?), they’re not booking through LLMs either yet. Most people find the tech remains clunky and see ChatGPT as that thing LinkedIn bros use to generate slop.
Everything might change in 2026, we are entering the “personal-agentic” era, then 2027 will be the Agent-to-agent era.
Until recently, AI agents lived in enterprise land: GTM AI, SDR AI, customer support bots, chief of staff AI, you can simply find everything as long as your company is willing to pay the price. Building and maintaining these systems required deep pockets, Clay’s AI functionality runs around $100,000 per year. At that price point, personally I can’t afford it.
Then Gemini Pro and Opus 4.5 dropped. The economics shifted overnight, development went from writing code to talking into your phone, ”Here’s my problem”, and watching the machine build it. Early adopters jumped head first and start building assistant for them and their family: Clawdis being a prime example.
Which brings us back to hotels.
Hotel operators have bought into the AI dream: One guest, one price. The algorithm calculates the maximum each customer is willing to pay and charges exactly that. It’s the apex of revenue optimization.
There’s one problem. Dynamic pricing only works when information is asymmetric. You win if you know the market and your customer doesn’t. You win if you have the algorithm and they’re guessing.
But when both sides have the same tools? the power shift hands.
Simple example; It’s raining in New York, within minutes, bootleg umbrella sellers rise like mushrooms after a storm. Unprepared tourists face a choice, pay whatever the closest seller demands or get soaked. Risk adverse vacationer will eat the surge price for one reason: You don’t know if you’ll find a better deal at the next corner.
Now imagine you have a personal AI in your pocket. It scans every seller within three blocks. Tells you the cheapest option and exactly how far away they are. Suddenly, you’re skipping the guy in front of you unless he matches your target price.
Hotels are worse off than umbrella sellers. They’re selling perishable inventory. An empty room tonight is revenue lost forever and independent properties live two bad nights away from bankruptcy. There will always be one desperate enough to drop prices to fill their property with last-minute bookings.
Your personal AI will sniff it out like pig with truffles. Cancel the existing reservation, rebook at the lower rate, and the best part is you won’t even notice it happening.
Large hotel groups can absorb this. They have armies of data modelers running forecasts and predictions. They can spread losses across their network. A bad week in Chicago gets offset by a good week in Miami.
Independent hotels don’t have that luxury.
If you’re running an independent property, you have one defense: loyalty. The kind built on a guest database that actually knows your customers. The kind that comes from understanding behavior patterns well enough to anticipate needs. The kind that diversifies your revenue mix so you’re not bleeding out when one segment evaporates.
Build relationships stronger than algorithms. Make your property something people choose even when the AI finds a cheaper option.
Price will always matter for travellers. But exceptional service and strong branding can outweigh a higher rate. Authenticity and transparency become the differentiators that algorithms can’t replicate.
In a world where personal AIs will authenticate every claim and question every price, trust becomes the most expensive thing to lose. More expensive than a few dollars on a reservation.
The hotels that survive won’t be the ones with the best pricing algorithms. They’ll be the ones customers choose even when their AI agent shows them a cheaper option two blocks away.
The DoorDash founders' reaction was textbook Streisand effect. The core insight about personal AI agents flipping information asymmetry is spot-on though, and its already happening in adjacent markets like travel. What makes DoorDash particularly vulnerable isnt just the perishable inventory problem but that they're stuck between restaurants who hate the take rates and consumers who are increasingly price-sensitive. The "desperation score" stuff aside, the bigger issue is how quickly these marketplaces commoditize when both sides have algorithmic pricing tools. I've noticed this with Uber vs Lyft comparisons - my phone now auto-checks both and the price differentials have compressed significantly over the past year, probably because more people are doing exactly that.
Brilliant insight!