Airbnb co-host commission math: what 20% really costs you

In March 2025 I hired a co-host to handle guest messaging and check-ins on my Tashkent unit while I was traveling for six weeks. We agreed on 20% of gross booking revenue. By the end of the trip I had paid her $612 for what turned out to be 14 stays of mostly autopilot work — calendar already synced, smart lock auto-rotating codes, cleaner on a fixed contract. My actual time saved was about 9 hours total. I had paid her $68 an hour to forward two messages a day. The 20% number had felt fair when I signed it; the math told a different story when I sat down with the spreadsheet two months later.

This is the post about how Airbnb co-host commissions actually break down once you draw the cash flow. The three pay models, the per-stay-volume math, the trap of percent-of-gross, and the clause that turns a $700 mistake into a $0 one.

What an Airbnb co-host actually does (versus a property manager)

Airbnb's official Co-Host feature, expanded across more markets in late 2024, lets a second person have keyed access to your listing's calendar, inbox, payouts split, and reservation actions. It is not the same as hiring a property manager.

The difference matters because the price tag attached to each role is different by an order of magnitude:

• Co-host: handles guest messages, check-in coordination, occasional issue triage. Does not usually price the listing, run pricing tools, or write the listing copy. Often a friend, a neighbour, or a part-time gig worker. Cost: typically 15–25% of gross, or a per-task flat.
• Property manager: runs the whole listing — pricing, listing copy, marketing, OTA channel management, cleaning vendor relationships, maintenance, tax filings in some jurisdictions. Acts as the legal-ish hand on the unit. Cost: typically 20–35% of gross plus setup fees, with management contracts.

A short-term rental property manager at 25–30% of gross is buying you a turnkey listing. A co-host at 20% is buying you message replies and a key handoff. If you pay property-manager rates for co-host work, the math collapses fast.

This post is about the co-host pay decision, not the property-manager one. They are different products at different price points; treating them as the same is the first place hosts overpay.

The three pay models

There are three live options in 2026 short-term rental forums, and a fourth that mostly exists in older agreements.

1. Percent-of-gross

The co-host takes a flat percentage of every booking's pre-fee gross. The Airbnb Co-Host tool defaults to displaying this as a "service fee" split on the payout screen.

The math is brutal because it scales with revenue, not with effort. A two-night $300 booking and a seven-night $1,050 booking both take roughly the same number of messages, the same single check-in, the same one welcome note. The co-host at 20% earns $60 on the first and $210 on the second for the same work. Hosts who run upmarket properties at $200+ a night get hit hardest.

It is by far the most common model and by far the worst for the host with a stabilised portfolio.

2. Percent-of-net

The co-host takes a percentage of net revenue: gross minus cleaning, platform service fees, supplies, and any pre-agreed pass-throughs. Net for a $1,050 booking with $80 cleaning, $35 platform fee, and $15 in supplies is $920. At 25%, the co-host earns $230. At 15%, $138.

Percent-of-net aligns incentives. The co-host wants the cleaning to be tight, the supplies to be right, the platform fees to be minimised. They have skin in the cost side of your P&L, not just the revenue side.

This is the model most experienced hosts move to after their first year on percent-of-gross. The shift requires writing the contract explicitly — Airbnb's default Co-Host split tool does not natively support a netted percentage, so net-based co-hosts get paid by a side agreement, with the Co-Host payout split set to zero or a small ceremonial fixed amount.

3. Per-task flat fee

The co-host invoices a fixed dollar amount per discrete task:

• Check-in coordination (key handoff or smart-lock confirmation): $8–$15
• Pre-stay message thread (the cluster of "what time", "is parking free", "wi-fi password"): $5–$10
• Post-stay handoff (cleaner kick-off, damage walk-through if needed): $5–$8
• Issue triage event (broken thing, noise complaint, lockout): $25–$50 per event

A typical 3-night stay in a low-friction unit costs $20–$30 in per-task fees. The same stay on percent-of-gross at 20% on a $270 booking costs $54. Per-task is cheaper for low-friction stays and protects the host from paying for autopilot work.

The catch: the co-host hates this model. It caps their upside and makes them invoice. Most won't accept it unless you're a repeat client or the relationship is friend-of-friend.

4. Salaried (rare)

For hosts above 8 listings, some co-hosts move to a fixed monthly retainer ($800–$2,000/month covering up to N stays per month, with overage at per-task rates). This is a property-manager-lite arrangement and shows up in larger portfolios. Not relevant to a 1–3 listing operator and I'll skip it for the rest of this post.

The worked math: 20% of gross at three booking volumes

One $90/night listing. Average 3-night stay. Pre-fee gross per booking: $270. Cleaning $40, platform fee $11, supplies $8. Net per booking: $211.

We're comparing the three pay models head-to-head at three booking volumes. Numbers are USD.

A few things jump out.

Percent-of-gross and percent-of-net land within $25 of each other in this scenario because cleaning, fees, and supplies are a relatively small slice of a $270 gross. Where percent-of-net pulls ahead is on listings with premium nightly rates and proportionally smaller fixed costs, or on listings where the host runs higher cleaning fees on purpose.

Per-task flat costs less than half the percentage models at every volume. That gap is the math reason the model exists: it strips the co-host's pay from the listing's revenue and ties it to actual work performed. Add in occasional issue events and the per-task host pays roughly $140 at 5 stays a month, where the percentage host pays $270.

At 20 stays a month, the per-task host pays $558 and the percentage host pays $1,055 to $1,080. The percentage host is paying $497 a month — about $6,000 a year — for the right not to write a contract that itemises tasks. That is the single most common money mistake in solo hosting.

When the co-host actually earns their keep

The math is not "co-hosts are never worth it." It is "co-hosts are worth it in four specific situations, and percent-of-gross is the wrong model in three of them."

1. Time-zone offset

If your unit is in Lisbon and you're in Toronto, every late-evening guest message at the unit lands in your morning. The co-host's value is that they reply within 15 minutes when the platform's algorithm watches response rate as a ranking input. The 1-hour Airbnb response window matters: missing it 3+ times in 30 days nudges your listing down in search.

For this case, the per-task model often misses the shape of the work — you're paying for availability, not for discrete tasks. A small retainer ($150–$250 a month) plus per-event triage tends to fit better.

2. Scaling past three listings

At one listing, you handle every message yourself in 20–30 minutes a day. At two, you start missing one a week. At three, the missed-message rate climbs to one a day and your average response time creeps past the platform's threshold.

This is the volume where a co-host pays back even at percent-of-gross. The marginal hour saved is no longer a leisure hour — it's an hour you would have spent on listing #4 or #5. Co-host fees become the platform tax of operating at scale.

3. The premium guest tier

A guest paying $400+ a night expects messages answered like a hotel concierge. A 90-minute reply window is a missed message. If you can't physically maintain a 15-minute response cadence (and you can't, indefinitely), you pay a co-host whose entire job is to be on call.

For premium tiers the math also flips on percent-of-gross: a co-host earning 20% of $400/night feels less wasteful per hour because the listing's per-stay margin is thicker. The work is also harder — guests at this tier expect concierge-grade attention, not message templates.

4. Compliance-driven in-person check-in

A handful of jurisdictions — parts of Italy, Spain, Austria, France — require in-person ID verification at check-in for short-term rentals. If you don't live within 15 minutes of the unit, you cannot do this yourself. The co-host is no longer a convenience; it is the only path to remaining legal.

In that case, the cost is the cost of running the listing. Negotiate per-check-in flat fee plus a small monthly retainer for compliance paperwork, and don't pay percent-of-gross — the co-host's effort is mostly invariant to the booking's revenue.

For the legal mechanics in each market, see guest registration laws for short-term rentals.

How to write a co-host agreement that doesn't drift

Most informal co-host agreements drift within 90 days. The host realises they're overpaying, the co-host realises they're answering more messages than they thought, and the relationship sours. A one-page agreement, written up front, prevents the slow renegotiation.

Five clauses to put on paper:

1. The pay model and the rate. Spell out percent-of-net (with the netting formula), percent-of-gross, or per-task. No "we'll figure it out as we go." Specify currency, when the invoice is sent, and the payment due date.
2. Defined scope. What the co-host does (messaging, check-in coordination, cleaner kick-off) and what they do not (pricing decisions, listing edits, OTA dispute handling). Both sides need to know where the work ends.
3. Response-time SLA. A target time for first response to a guest message (commonly 1 hour during 08:00–22:00 local time). A short SLA is the thing the host is actually buying.
4. Exit clause. Notice period for either side to end the arrangement, usually 14 days. No early termination fee. The relationship needs an exit that doesn't require lawyers.
5. Data and access. What credentials the co-host gets, what they don't, and a written rule about not contacting past guests outside the platform. Doubly important under GDPR for vacation rental hosts if you operate in the EU — your co-host becomes a data processor and you are still the controller.

Five clauses. A few hours to write the first time, then a template you reuse. The agreement is what turns the working relationship from a drift into a contract.

One opinionated take

The default co-host contract — 20% of gross, no scope written down, no exit clause, no response-time SLA — is a tax on hosts who don't sit down with a spreadsheet for an afternoon. The four to five hours it takes to write a real one-page agreement saves the average host between $1,200 and $4,000 a year, and the savings climb with portfolio size.

If you have one listing and stable bookings under 8 stays a month, do the work yourself. The maths in section three is clear: at that volume the co-host is paying themselves $50+/hour to forward messages your phone already buzzes you about. Hire the co-host the day you start a second listing, not the day you start the first. And when you do, write the contract on net or per-task, not on gross. The host who pays percent-of-gross at scale is the host whose listings end up funding someone else's vacation, not their own.

For the broader operational picture, including how channel managers and tools change the make-vs-buy decision, read the channel manager break-even math — the same logic applies to every paid intermediary in a host's stack.