Mint Apartments · Platforma
Direct bookings on your own site: portal commission returns as margin
If you sell nights through Booking or Airbnb, commission is not “marketing cost”. It is a tax on every stay you could have closed yourself. At Mint Apartments the guest books on the operator’s domain, pays online and gets a lock code. Estimate: 10–15% cheaper for the guest than the portal*. That is the only public, clickable booking example I show — no invented occupancy rates.
What “on your own site” means, and what is still an OTA
A site with a “book on Booking” button is not direct booking. It is a banner. Direct booking: live calendar and price, payment on your domain, confirmation and onward handling (code, house rules, language) without a middleman in the transaction. The portal can stay a channel. It does not have to be the till.
At Mint, calendar and availability come from Previo. I did not write a channel manager. That is the SaaS versus custom panel rule: off-the-shelf where the market is standard, custom where the process is yours.
When the maths closes, and when it does not
It closes when you have repeat demand, a recognisable brand or paid traffic you already send to a property page, and commission eats a double-digit share. It does not close when 90% of nights still come from OTAs and the site has three visits a week. Then channel and offer first, not a payment engine.
- Count annual commission, not “2% here or there”.
- Add desk / WhatsApp hours on “where is the code”.
- Subtract the Platform build (from PLN 25,000) and gateway upkeep.
- If payback sits beyond 18–24 months on your numbers — I say so.
A PLN 2,000 Start brochure will not carry this. Calendar, payment, Tedee/Nuki locks and a multilingual assistant are a platform, not five pages with a form.
Check-in without a desk is not a gadget
A lock code at 23:00 is part of the funnel: “arrive when you want” fails if someone must be on duty. A guest assistant (WhatsApp, apartment context, human escalation) takes repetitive questions. Estimate 8–15 h/month* — with an asterisk, because 36 apartments in Gdańsk is not your 4-room guesthouse.
I do not ship an assistant that invents discounts. HITL: allowed actions, a log, a human on disputes. Same spine as human-in-the-loop automation.
What I will not promise from Mint as “your result”
- I do not transfer Mint occupancy onto your property.
- I do not guarantee the OTA disappears — it often stays as fill.
- I do not build a PMS from scratch if Previo (or another standard) already holds channels.
- I do not show I DRIVE CARS as live — it is not in production.
Public proof is mintapartments.pl. Everything else is your numbers in a 20-minute audit. If you lack volume, we stay on a company website and do not fake a booking engine.
A stack you can live with after handover
Astro/React on the front, Previo for stock, payment on-domain, locks, WhatsApp, language versions. After invoices, the code is yours. NDA is standard. I will not hold you on a theme you cannot export. Upkeep is the gateway and the PMS, not twenty WordPress plugins.
Commission maths a layout cannot cheat
Direct booking on your site cuts OTA commission. At Mint the guest pays 10–15% less than on Booking.com — a number you can check on a live path, not a slide. The portal stays as a channel. The site takes people who already know you or arrived from a campaign onto the domain. If 90% of occupancy must be bought from an OTA tomorrow morning, “your own site instead of Booking” is a slogan, not a strategy. Channel first, then close on the domain.
I do not build a calendar from scratch inside a website budget. A PMS (at Mint: Previo) holds stock, channels, settlement. The site captures intent and hands the guest to a known system. SaaS or a panel explains why sewing a PMS at PLN 2k is vanity. The Platform package (from PLN 25,000) starts when that integration and the payment path are real, not when someone wants a “Book” button on a brochure.
- Rate parity: what the guest sees with you versus the portal — no surprise at lock.
- Cancellations and terms that match in the PMS and on the site. Drift kills trust.
- Mobile: a booking that needs pinch-zoom returns to the OTA.
- A confirmation mail that arrives. A funnel without mail is leaky.
When to leave Booking as the till and not pretend “direct-first”
A new property with no brand, no returning guests, no campaign budget — occupancy first, then direct. The site is then a brochure with a clear “book” that may point at an OTA or a PMS widget, as long as the price does not lie. Direct-first makes sense when you already feel commission as a cost, and you have someone to handle mail and check-in. Otherwise you will build a pretty till nobody walks up to.
I do not promise the “site will replace Booking in 90 days”. I promise a path the guest can finish on your domain, and an integration I do not pretend I wrote from scratch. If someone sells a homemade booking engine at brochure price — ask who watches overbooking on Saturday at 23:00. I want that to be a system that already does it, not my weekend.
A checklist before you paste a Book button on a brochure
Parity with the OTA rate, the same cancellations, a PMS that watches overbooking, mail that arrives, a phone path without pinch-zoom, who answers the guest after lock. Without that the button is a poster. I do not build a calendar from scratch at website price — at Mint it is Previo. The Platform package (from PLN 25,000) starts when that list is real. A new unbranded property may honestly point at the portal. Direct-first without a channel is a slogan. The 10–15% less for the guest at Mint you can check on the live site, not on a slide of “industry” commissions.
Questions
Can I keep Booking and still take direct bookings?
Yes. I often do: OTA as a channel, your site as the cheaper till. Condition: the same stock, no double bookings.
What does it cost?
Platform: from PLN 25,000, not the Start brochure package. A precise figure after the audit, once we count commission and handling hours.
Will you write a calendar from scratch?
Rarely. If a PMS/channel manager exists, I attach it. A custom calendar makes sense when the process is unusual and the off-the-shelf tool lies.
Is an AI assistant mandatory?
No. Booking and check-in first. I add an assistant when question volume justifies it — HITL, not a decorative chat.
Will the site replace Booking in 90 days?
I do not promise that. The portal stays a channel. The site takes guests who know you or arrived from a campaign onto the domain. If tomorrow morning 90% of occupancy must be bought from an OTA, “direct-first” is a slogan. Channel and occupancy first, then close on your own domain. At Mint the guest pays 10–15% less than on Booking — a number from a path you can click, not a slide. A new unbranded property should often keep OTA as the till and not pretend a strategy it cannot carry.
Why not build your own calendar at website price?
Because overbooking on Saturday at 23:00 should be a PMS job, not my weekend. At Mint that is Previo: stock, channels, settlement. The site captures intent and hands over the guest. Sewing a booking engine at PLN 2k is vanity. The Platform package (from PLN 25,000) starts when integration and payment are real, not when someone wants a Book button on a brochure. Rate parity, the same cancellations, a confirmation mail, mobile without pinch-zoom — without those the guest returns to the OTA.
Is an OTA widget on your domain already “direct”?
Sometimes it is a more honest first step than a lying price. Direct means the guest finishes and pays with you, and commission does not eat margin the same way. A widget that shoves them onto Booking at another price teaches the guest the site lies. Better a clear button to the portal or a real PMS widget with parity. I do not mash Mint (a booking funnel) and Plumm (an operations product) into one universal “company site with booking” case.
Let us count commission before we code a calendar
Twenty minutes: whether direct booking has a right to close. Mint is clickable under work.
Book a 20-min audit