Direct answer

Log the cancellations on the record, move the date day for day, and reset the calendar in writing.

The meetings are where this project gets its decisions: the review where work is accepted, the call where the next stage is unblocked, the check-in where a question gets answered in five minutes instead of five days. Your client keeps cancelling them at the last minute, each time reasonably, each time with an apology, and the project is drifting weeks past the date everyone still quotes as if it were real. Drift like this has a quiet danger beyond the lost time: at the end of the job, the delay will be remembered as the project's delay, and unless the causes were written down as they happened, it will be remembered as yours. The fix is not complaining about the cancellations, which changes nothing and spends goodwill. The fix is machinery: a dated cancellation log on the job record, a day for day rule that moves the delivery date by the delay each missed decision causes, and one reset conversation that puts new dates, a cancellation window, and the consequence of the next late cancel into writing. This page gives you the log format, the rule, and the exact wording for the reset, so drift becomes visible, priced in time, and owned by its actual cause while there is still time to recover the job.

See the instrument first: the live demo runs one real call and hands you its verifiable receipt, no signup.

Updated November 22, 2026. First published November 22, 2026. The records quoted below are real and public, read back from the live service on November 22, 2026.

Step 1. See the drift for what it is: decisions not made, on dates you can name

A cancelled meeting is rarely just a lost hour. On a live project, meetings carry decisions, and a decision not made on Tuesday does not vanish, it queues: the stage that needed approval waits, the work that depended on the answer waits behind it, and people downstream either idle or guess. One late cancellation is life. A pattern of them is a schedule change being made by someone who never has to say the words the date has moved, while you carry the original date in every conversation as if it still described reality. That gap, between the quoted date and the real one, is where disputes are born, because at delivery time the client remembers the date they were quoted and you remember the six meetings that moved it, and memory is not evidence. The way out is to make the drift visible while it is happening, neutrally, as a log of facts rather than a complaint about behaviour. Facts are hard to argue with and easy to forgive. A client shown a calm dated log usually recognises the pattern themselves, often before you finish explaining it, and that recognition is what makes the reset conversation in step 5 a planning discussion instead of an accusation.

Step 2. Start the dated cancellation log today, on the job record

Open the log on the job record itself, next to the stages it affects, and give every entry the same five facts: the meeting date, what the meeting was for, the decision or input it carried, when the cancellation arrived, and the new date the decision actually landed, or that it is still waiting. Keep the wording flat and factual: review meeting for stage two, cancelled by client message at 16:40 the day before, approval given by reply on the record four days later. No adjectives, no tally of frustration, nothing you would not want the client to read, because the log works best when the client can read it. Backfill the recent past honestly from your calendar and messages, and mark backfilled entries as backfilled, with the source, so the log never claims a precision it does not have. From today forward, entries are made on the day, in minutes, as part of closing out any cancelled meeting. Within a few entries the log starts doing work no complaint can do: the pattern has dates, the total decision delay has a number, and the delivery date conversation stops being about whose fault anything is and becomes arithmetic both sides can check.

Step 3. Apply the day for day rule: the delivery date moves by the delay the decision caused

The rule is simple and, stated in advance, hard to call unfair: when a client-side decision lands later than the agreed meeting because the meeting was cancelled, the delivery date moves by the same number of days the decision moved. A review that slips four days moves the stages behind it four days. Apply it mechanically and announce it neutrally at the time, in one short line attached to the rescheduled decision: the approval landed on [new date], so the delivery date for [stage] moves from [old date] to [new date], as logged. No blame language, no lecture, just the date doing what dates do when their inputs move. Two disciplines keep the rule clean. First, it only ever moves dates by client-side waiting time, never by your own delays, which the same record shows separately, so the rule can never be used to launder your own slippage. Second, where a cancelled meeting carried nothing time-critical, say so and move nothing; a rule applied only when it truly applies is a rule clients accept as honest. If your current terms do not state the day for day rule, you cannot impose it backwards as a penalty, but you can run the log from today and propose the rule in the reset conversation, where a client looking at their own pattern usually agrees to it on the spot.

Step 4. Let the record separate client-caused drift from your own delay

The question underneath every drift dispute is whose waiting time it was, and a job record answers it structurally: your stages are saved as they complete, with dates, and the decisions sit on the same record with theirs. The record format is AER-1, published as an IETF Internet-Draft (draft-zambo-aer1-26), created by Brennan Zambo, and each step carries a verifiable receipt: a receipt proves the saved result was not changed since it was saved. The two public records below show the two states that distinction is made of:

Accepted: the decision is on the record

Pending: checked, and still waiting on a decision

Read the pending record as drift made visible: 5 steps total, 5 checked, release state hold. The work side is complete and checked; what the job is waiting on is a decision, and the state says so without accusing anyone. The accepted record shows the decision landed: decided_by client, decided_at 2026-10-16T14:31:51.146Z, release state ok_to_release, 4 steps checked. On your project, the cancellation log plus these states give you the sentence every drift conversation needs, provable rather than remembered: our stages were checked and waiting on these dates, the decisions landed on these dates, and the gap between them is the drift, whoever caused it. When some of the gap is yours, the same record shows that too, which is exactly why clients believe it when it shows the rest.

Step 5. Hold the reset conversation: new dates, a window, and the next-cancel rule, in writing

Ask for one meeting whose whole agenda is the calendar, and open it with the log, not with frustration. The shape, word for word in structure: I want to reset the dates so they are real again. The log at [link] shows [number] cancellations since [date], and the decisions behind them landed [total] days late in aggregate, so the current delivery date of [old date] no longer describes the job. Here is the reset I propose. One, new dates: [stage dates and final delivery date], rebuilt from today with the remaining decisions scheduled. Two, a cancellation window: either side can move a meeting up to [the prior business day, or a stated number of hours before], after which the meeting either happens or its decision is made another way. Three, the fallback: if a meeting is cancelled inside the window, its decision is made by reply on the record by [date], so the job never waits on a diary again. Four, the consequence, stated once and kindly: if a decision still lands late, the delivery date moves day for day from the log, so the dates you hold are always the true ones. Confirm the whole reset in writing the same day, on the record, and the drift problem is now governed by machinery both sides agreed, instead of by hope and memory.

Honest limits

This page is general information about managing a drifting schedule, not legal advice: whether a delivery date can be moved, and what a delay means for fees or exit rights, depends on what your agreement says about time, changes, and delay, and a large or disputed engagement deserves advice on its own documents. A day for day rule binds as a term only if your agreement states it or the client agrees it in the reset, in writing; otherwise it is a fair proposal and a clear log, not a penalty you can impose backwards. Some cancellations are genuine emergencies, and the machinery here is aimed at a pattern, applied with judgement, not at invoicing a client for a crisis. A verifiable receipt proves the saved result was not changed since it was saved; the log proves what was recorded and when, and the fairness of any date move still rests on the entries being accurate and the rule being agreed.

Step 6. Put the machinery in your terms, so the next job starts with it

Close by making the reset permanent in your standard terms, while this job's lesson is fresh. Four short clauses do the whole job. A meeting clause: project meetings are decision points, scheduled with the stages they unblock. A cancellation window: meetings can be moved up to a stated point before they start; inside the window, the decision defaults to the written fallback. An async decision rule: any decision a meeting carried can be made by reply on the job record, and counts as made on the date the reply lands. And the day for day clause: where a client-side decision lands late against the agreed schedule, delivery dates move by the same number of days, recorded in the project log. None of these clauses is aggressive; clients who run their own diaries well will never feel them, and clients who cancel often will feel them only as dates that stay honest. Over a few jobs the effect compounds in the direction you want: meetings become rarer and more decisive, because their cost in visible schedule is understood, decisions increasingly arrive in writing on the record, and drift, when it happens, arrives already explained, dated, and owned by its real cause.

Frequently asked questions

My client keeps cancelling our meetings at the last minute and the project is drifting. How do I handle it?

Stop treating each cancellation as an isolated apology and start treating the pattern as schedule data. Keep a dated cancellation log on the job record: meeting date, the decision it carried, when the cancellation arrived, and when the decision actually landed. Apply a day for day rule, announced neutrally each time: when a client-side decision lands late because its meeting was cancelled, the delivery date moves by the same number of days. Then hold one reset conversation built on the log: new realistic dates, a cancellation window after which meetings either happen or their decisions are made by reply on the record, and the day for day consequence stated once, kindly, and confirmed in writing the same day. The log turns a blame conversation into arithmetic, and the reset turns arithmetic into an agreement.

Can I charge my client for last minute cancellations?

Only on a basis that exists: a cancellation or rescheduling term in your agreement, or one the client agrees in the reset conversation for the rest of the job. Without either, a charge invented mid-project reads as a penalty and usually costs more goodwill than it recovers. The stronger tool for most projects is time rather than money: the day for day date move, which costs the client schedule rather than cash and is easy to defend because the log shows exactly which decision waited how long. If cancellations also burn booked resources, travel, a room, a specialist held for the session, those concrete costs can be raised factually in the reset and a term for them agreed going forward. Agree it in writing first, then apply it flatly and without heat.

How do I tell the client the delivery date has moved without sounding like I am blaming them?

Attach the move to the decision, not to the client, and make it at the moment the decision lands rather than in a batch later. The working sentence is: the approval for stage two landed today, [date], so delivery for stage three moves from [old date] to [new date], as logged on the record. It names no fault, it cites the log both sides can open, and it arrives while the cause is fresh and obvious. What creates the blame feeling is the alternative most people default to: silence for weeks, then a large unexplained slip announced at the end, which forces the client to wonder what really happened. Small, dated, neutral moves, made every time, read as professionalism. One big late surprise reads as an excuse, even when it is true.

What if the client says the delay is my fault, not theirs?

Answer from the record, not from memory, and concede your share first if you have one. The job record separates the two kinds of waiting: your stages are saved with the dates they were completed and checked, and the client decisions sit on the same record with their dates, so the gaps sort themselves into work time, your waiting on decisions, and any time that was genuinely yours. Walk the log together: these stages were checked and waiting on these dates, these decisions landed on these dates, and this stretch here was mine, which I own. A client arguing with a person can keep arguing. A client looking at a dated record that honestly shows your delays too has very little left to dispute, and the conversation usually moves to the reset, which is where you wanted it all along.

How does a job record help with a drifting project?

It holds the cancellation log and the stage states in one place, so drift is visible as it accumulates instead of being reconstructed from memory at the end: stages checked and waiting read as a pending state, decisions read with a decider and a date, and the gap between them is the drift, attributed by dates rather than by blame. When the work runs through Zambo, built by Brennan Zambo, every job lands on its own record with a verifiable receipt for each step under AER-1, published as IETF Internet-Draft draft-zambo-aer1-26, so the reset conversation runs from a page both sides can open. Every tool free once a day, plus 3 continue calls. One free wrap-up handover per day, and pickup is always free. For $0.99, run the job on one record, so cancellations, decisions, and date moves are logged where the client can see them too: get the Day Pass. A Day Pass is unlimited for 24 hours, no subscription. Ongoing use is the Zambo Pass at $29 a month, $19 a month for $ZAMBO holders, or Lifetime at $250 one time.

Related