Response-Time Thresholds: Where Slow Replies Cost You Ranking
Response rate is a share, not a speed: every new inquiry must be answered inside 24 hours or it counts against you, and the window looks backwards so a bad month lingers. The published eligibility line is 90%, but the operators who rank well work to a much tighter number. This guide covers how to protect the figure with three kinds of reply, and how to read your own numbers before they drop.
The line is 24 hours. Nobody who ranks well is standing anywhere near it.
Last updated: September 22, 2026
Response rate is not a speed. It is a share: of every new inquiry and reservation request that arrives, how many got a reply before the 24-hour mark passed. That framing is the whole problem, because a metric shaped like that is hurt by neglect rather than slowness, and it is reviewed over a window that looks backwards — a bad month stays visible long after it ends. This guide covers what actually gets counted, how far the published eligibility line sits from the number operators who rank well work to, the three kinds of reply that protect the figure without living on your phone, and a ten-minute weekly read of your own numbers before somebody else reads them for you.
Key Takeaways
- 24 hours is the floor. Reply inside it and nothing is gained; miss it and nothing is forgiven.
- The window looks backwards. A weak month keeps counting until it ages out of the window on its own.
- A placeholder buys the time. One sentence, sent immediately, moves the clock from tonight to tomorrow morning.
- Templates cover the common half. Six answers cover most of what guests ask; only variables need writing.
- Escalation has a name. One person, one permission, one way to reach them when you cannot answer.
What the metric actually counts
Strip the dashboards away and there are two numbers, and only one of them is the one people worry about.
The first is the one that matters here: the proportion of new inquiries and reservation requests answered within 24 hours, measured across the recent past. Anything that arrives and is never answered sits in the denominator, unpaid. Check the exact wording and the window shown in your own host dashboard — that display is the authoritative version, and both its definition and its look-back have changed before.
Two mechanics follow from that shape, and both catch hosts out:
| Mechanic | Why it bites | What it means for you |
|---|---|---|
| The window is retrospective | Recent and older messages count until each ages out | A bad month lingers; recovery is measured in weeks, not in one good day |
| The clock is 24 real hours | It does not pause for your working day | A message at 23:00 expires at 23:00 tomorrow, wherever you were at midnight |
| Silence counts the same as "no" | Unanswered requests stay in the denominator | Declining promptly beats letting a request you already disliked rot |
That last row is the expensive one. Most hosts do not lose points because they were slow to say yes; they lose them because they never said anything.
One honest caveat: the definition, the 24-hour line and the eligibility threshold in your dashboard are disclosed by the platform, and those are the parts you can rely on. The working targets in the next section are a different category of claim entirely, and are labelled as such.
How far the floor sits from the line that ranks
There are three different numbers in circulation, and only two of them come from the platform.
| Line | Where it comes from | What it buys you |
|---|---|---|
| 24 hours | The definition of response rate itself | Nothing positive — it simply keeps you out of trouble |
| 90% or higher | The published eligibility requirement for the badge | Eligibility at the next review, alongside every other criterion |
| 95%–100% | Trade reporting (industry claim, see note below) | The band that operators ranking well are reported to work to |
Below those two disclosed lines sits a set of figures that circulate constantly and come from somewhere else entirely: that well-ranked hosts target 95%–100%; that replies landing inside an hour during active hours convert better, one report putting that advantage at roughly 25%; and that a host moving from 89% to 100% saw bookings rise by as much as 116%. Those three figures belong to emeraldcoastbyowner.com and similar trade sources. They are one-party claims, not platform disclosures, they have not been independently audited, and they should be read as ambition rather than as a rule.
Even taken loosely, the direction they point is useful. A target somewhere above a disclosed 90% floor is not a vanity setting: it is headroom. Eligibility is reviewed on a natural cycle, so the month you let a thread rot is not a month you can undo with two strong weeks afterwards — it sits in the window until it rolls out of it. Ask what costs more, an extra ten minutes a day or a quarter of your eligibility resting on one bad fortnight.
Replies slip for a boring reason far more often than a dramatic one: whoever is answering has to go and look something up. Availability and rates for Airbnb, Booking.com, Agoda and Trip.com sit on one grid at localsbnb.com, so the answer to "is it free, and how much" is a glance rather than four logins — and if the channels disagree, that is visible too, because the grid shows each one's source rate and source status.
Three replies that protect the number
None of these require you to be faster. They remove the need to think when thinking is least available.
The placeholder. One sentence, sent within minutes, that says you have seen the message and when the real answer is coming. It does three jobs at once: it stops the 24-hour clock being the thing you manage, it makes the guest feel answered, and it converts "reply now" from a task into a scheduled one. Write it once, keep it somewhere reachable from a phone lock screen.
The template set. Count your last fifty inquiries and you will find six questions make up most of them: parking, earliest arrival, luggage before check-in, pets, how far it is to something specific, and whether there is a cot or a highchair. Draft each answer once, and leave the variables obvious — the actual earliest arrival time, the actual nearest car park — because a template with the variable left blank reads worse than no answer at all.
The escalation. Name one person who answers when you cannot, and be explicit about what they may decide without you: whether they can move a guest, whether they can approve money, and how much. An escalation with no ceiling generates exactly the messages you were trying to avoid, so write the number down rather than saying "use your judgement".
| Reply | Job it does | Fails when |
|---|---|---|
| Placeholder | Buys the 24 hours and calms the guest | It promises a time you then miss |
| Template | Answers the recurring half in one tap | Variables are left unedited |
| Escalation | Covers the hours you are genuinely unreachable | The stand-in has no authority to decide |
Keep the written versions somewhere that survives a change of phone, and rewrite one the first time a guest complains about it.
Reading your own numbers before they bite
Ten minutes a week is enough, and the questions are three. First: how many threads went unanswered past the twelve-hour mark — those are the ones that became emergencies later. Second: which questions took longest because the answer had to be looked up, because each of those is a template waiting to be written. Third: which recurring question arrived twice this week, since anything asked twice is a listing problem, not a correspondence problem.
Do the same read monthly, but on the trend rather than the snapshot. Because the window looks backwards, the visible figure lags your behaviour by weeks; judging yourself on one week's percentage is how people conclude that a bad month cost them nothing, right up until eligibility day. Write the number down monthly with the date, and after three months you will have what no dashboard gives you: your own baseline.
Looking things up across channels is the part worth deleting rather than speeding up. Asking questions in one sentence is another route to the same facts — LOCALSBNB connects Claude, ChatGPT and Cursor to your property data, so you can ask today's arrivals, in-house guests and departures, order detail, room status, the room-type calendar, channel rates, and occupancy, ADR and RevPAR, in the language your property runs in. It installs once and adds no new app, and those answers are read-only.
Replies stall most often when the facts behind them are scattered across four back offices; keeping availability and rates current together at localsbnb.com removes that lookup from most answers.
FAQ
Does declining a request hurt my response rate?
Answering it does not. Leaving it open does, because unanswered requests stay in the denominator. If the answer is no, say no quickly — that is both the politer and the cheaper option.
Do automated replies count?
Anything you send inside the thread counts as a reply; check the current definition shown in your dashboard before relying on any particular mechanism. The reason the placeholder works is not automation but timing: it arrives before the guest has decided to message three other properties.
How long does one bad month last?
Until it ages out of the measurement window, which is exactly the point about recovery. Treat eligibility as reviewed on a natural cycle, and assume the month you neglected is with you for a while.
Response rate is one of the few host metrics that punishes neglect rather than slowness, which makes it unusually cheap to protect: a placeholder sent now, six answers written once, and one person who can decide something without you. Do those three, then check the number weekly — and put the four channel calendars that most answers depend on onto one grid inside LOCALSBNB.
Metric definitions, eligibility criteria and how long look-back windows run are set by each platform and change from time to time; check what your own dashboard shows today. Figures attributed here to trade sources are their claims, not platform disclosures and not independent measurement.
確認担当
Localsbnb 編集チーム