Multi-person scheduling stalls because you cannot see the other side's calendar

Meeting scheduling and room booking take an average of 26.5 minutes a day. The most common difficulty reported was "finding a time when the participants are free", at 72.0% (source below). Each instance is short; in a week where internal and external meetings stack up, it adds up to half a working day.

Adding more candidate dates does not move that 72.0%, because what we call "scheduling with multiple people" is really three different problems. This article identifies which one you are in with a single question, then covers what to do in each, and the shape of message that lets the other side settle it in one reply.

For the wider set of scheduling problems in sales and hiring, see the sales and recruiting resource hub.

"Multiple people" covers three different situations

Headcount is not what makes it hard. The number of organisations involved is. Five people from one company settle it in minutes; three people from three companies start a round trip.

The scale of the load is measurable. Time spent holding and arranging meetings comes to roughly 30% of the working day, of which scheduling, finding a room and booking it take an average of 26.5 minutes. The most-cited difficulty was "struggling to find a time when participants are free", at 72.0% (source: Acall Inc., "Survey on Meetings", published 10 July 2024, fielded 18–20 June 2024, online panel survey, n=600; the figure includes time spent booking meeting rooms).

What stands out is that the top answer was not "the back-and-forth is tedious" but "we cannot find a time that is free". The problem is missing information, not effort. And whether that information is obtainable at all is decided by which type you are in.

One question tells you which one you are in

Ask: how many organisations other than ours are taking part? Zero is type A, one is still type A, and at two it becomes type B. If the other side is a single person but you are running dozens of these, you are in type C.

Type	Who is involved	What works	Round trips
Type A	Several of us, at most one outside party	Match our own calendars automatically, show only the surviving slots	One
Type B	Two or more outside parties	Collect answers on candidate dates (a poll), or lock parties in one at a time	Two or more, usually
Type C	One person at a time, but many of them	Automatic assignment to whoever is free	One, repeated

The table says something specific: automatic matching works up to the edge of the calendars you can read, and no further. Once two or more outside parties are involved there is nothing to match against, so changing tools does not help. From there it stops being an information problem and becomes a question of how many times you make the other side answer.

Type A: several of us, one outside party

This is the one that automates cleanly: three of our salespeople and one client, two interviewers and one candidate. Work out the times when everyone on our side is free, show only those, and the exchange ends in one round trip.

Four things decide whether a tool can carry this type for you.

1. Does it check everyone's availability? A tool that reads one calendar will offer times that collide with the other participants
2. Does the calendar sync both ways? If events booked elsewhere do not block, the slots you offered fill up twice
3. Does the other side avoid signing up? Requiring an account from an external guest adds a round trip on its own
4. Is the meeting link generated automatically? Creating and sending a URL afterwards costs the same as another round trip

Below, as one example of a tool that meets those four points, we look at SailLab.

SailLab

SailLab is a Japanese scheduling tool that compares several people's calendars and publishes only the times that all of them are free, behind a single link.

• Only slots where everyone is free are shown: name multiple hosts and the guest sees the intersection, nothing else (Light plan and above)
• Two-way sync with Google and Outlook: anything booked outside SailLab blocks the time automatically
• No sign-up for the guest: they open the link, pick a slot, done
• Meeting links are issued automatically: Google Meet, Zoom and Microsoft Teams are supported
• Confirmation lands in everyone's calendar: both sides get an email, and the guest can reschedule or cancel

Plan	Monthly (annual billing, ex-tax)	What matters here
Free	¥0	One scheduling link, Google / Outlook sync, automatic meeting URLs
Light	¥800	Unlimited links, multi-host scheduling (only slots where everyone is free), team management
Standard	¥1,200	Round-robin assignment, booking analytics

Running type A through one link needs multiple hosts, so it starts at the Light plan (as of August 2026). See pricing for details and sales teams for how this is used on client meetings.

One thing it does not do: poll-style scheduling, where candidate dates are listed for everyone to mark. If type B is your main business, the next section is the more useful one.

Type B: two or more outside parties

In this type, the other side's availability never reaches you. Even inside Google Calendar, the most that sharing exposes to someone you have shared with is "See only free/busy (hide details)" (source: Google Calendar Help, "Share your calendar", as of August 2026; on Google Workspace an administrator may restrict sharing further). For an external party who has shared nothing with you, there is nothing to see at all.

So type B is not a matching problem, it is an agreement problem, and there are only two moves.

• Collect answers (a poll): list candidate dates and have everyone mark what works. Effective with larger groups, but nothing is settled until the answers are in
• Lock them in one at a time: secure the party with the least movable calendar first, then offer that slot to the rest. No extra round trips, but the later participants get no choice

In practice the second settles faster. The rule is to confirm the person whose calendar cannot move — an executive, a client's decision-maker — before anyone else. Offering everyone an equal choice is fair and does not converge: with six people at six companies, the chance that three candidate dates suit all of them is low, and everyone's calendar keeps moving while you re-propose.

Polls do have their place. If you are on Microsoft 365, Outlook's built-in scheduling poll is already there; scheduling in Teams and Outlook covers which built-in option fits which case. When the other parties are outside your organisation and not on the same stack, pick something they can answer without an account — a no-signup scheduling poll, for instance. The moment you ask a respondent to register, your reply rate drops.

Type C: a lot of one-to-ones

The other side is always one person, but there are twenty of them a week: interviews, recurring one-to-ones, first calls after an enquiry. No single instance is hard; the volume is what makes it hard.

What works here is not matching but assignment. Register the people who can take the meeting and let each request go to whoever is free. Instead of choosing an owner and then finding a time, the time and the owner are decided together, which removes an internal step.

In this type the deciding factor is how much of the correspondence you can remove. For the wording of interview and meeting scheduling messages, see interview and meeting scheduling emails.

In every type, close your own side first

Once you know the type, one step is common to all of them: settle your own side's availability before you ask the other side anything. Do it the other way around and you get one more round trip, because your slot has gone by the time their answer lands.

When everyone sat in the same room you could walk over and hold a time verbally. That assumption has weakened: 47.3% of companies have adopted telework (source: Ministry of Internal Affairs and Communications, "Results of the 2024 Communications Usage Trend Survey", published 30 May 2025, covering companies with 100 or more regular employees). In close to half of companies, the person you are scheduling with may not be at the next desk.

How you close your own side depends on the type. Type A: build a link from the matched availability. Type B: block the candidate windows in your own calendar as tentative holds. Type C: register each owner's bookable hours in advance. In all three, the goal is that this is finished before the first email goes out.

The message shape that gets one reply

Round trips do not multiply because a message was insufficiently polite. They multiply when the information does not let the other side finish in one reply. Four elements are enough.

1. At most three options, given as windows: "any time between 14:00 and 17:00 on the 5th" leaves them room to arrange their own side
2. A by-when: without a deadline it slides. "Could you let us know by Friday" changes the response rate on its own
3. The next move if none of them work, stated up front: "If none of these suit, please send two times that do." Without it, "none of these work" costs a whole round trip
4. Duration, purpose and attendees: 30 minutes and 60 minutes are different searches on their calendar

In type B the third element carries the most weight: having the fallback already in the message removes the re-proposal round trip when everyone declines. For the wording itself, see scheduling email templates.

What to change on the next one

Nothing already in flight needs to be redone. On the next one that comes up, change only the order.

1. Same day: count how many organisations other than yours are involved, and name the type
2. Before writing anything: close your own side. Type A, build the link; type B, put tentative holds in your calendar
3. In the first message: options, deadline, and the fallback — all three, once

Those three steps turn two round trips into one. The part of the 26.5 minutes that disappears is the searching.

If you schedule inside Microsoft 365, see scheduling in Teams and Outlook. If you want to publish open slots from Google Calendar, see appointment booking with Google Calendar.

SailLab's Free plan needs no credit card — create an account to try it.