Decide Whether to Send Near the End of a Staff Shift
Prepared with AI assistance by Alactic Editorial. Examples identified as hypothetical are illustrative, not customer results.
A campaign sent before staff finish work is not necessarily a campaign the team can support. The message may arrive while someone is at their desk, but recipients may read it after work, discuss it with someone else, or reply hours later.
Send near a shift boundary only when the expected reply load fits the remaining staffed time and the likely unattended enquiries are acceptable for that particular offer. Otherwise, reduce the audience, move the start earlier, or defer. The relevant deadline is not the last moment someone can press send; it is the last start that leaves enough useful coverage for the resulting conversations.
The worksheet below compares those choices using a fictional campaign. It also shows why moving a campaign to tomorrow does not, by itself, eliminate overnight replies.
1. Define what must be true before the team leaves
Start with the consequences of a new enquiry arriving after closing. A general request for information may reasonably wait until the next staffed period. An enquiry about an offer that expires that evening has a different consequence: the recipient may be unable to act without an answer.
Ask what someone is likely to need after reading the message. Can they understand the offer and decide without assistance? Or does the campaign invite questions about availability, suitability, booking details, or a decision that cannot wait?
Use that distinction to set an internal overnight tolerance. This is a planning limit, not a service promise. It might allow a small number of low-urgency enquiries to wait, while rejecting any send likely to create time-sensitive enquiries without staff coverage.
- Lower consequence: an informational announcement with no immediate deadline and enough detail to stand alone.
- Higher consequence: an invitation that requires a staff answer before the recipient can complete the intended action.
- Reason to defer regardless of volume: the message implies immediate assistance that the team will not be available to provide.
Do not let a reassuring average hide a serious individual case. A small expected queue can still be unsuitable if the campaign creates urgent dependencies.
2. Work backwards from a protected finish
Separate the end of the shift from the end of useful intake. Staff need time to finish work already under way, handle unexpected demand, and leave the queue in a known state. Treat that time as a protected buffer rather than spare campaign capacity.
A useful starting rule is:
Latest start = shift end − protected buffer − chosen reply-coverage horizon − launch allowance
The reply-coverage horizon is how long after launch you want staff available to receive the main reply wave. The launch allowance accounts for the difference between the planned start and the likely distribution of messages. Neither is a guarantee that replies will stop afterward.
Hypothetical rule: if staff finish at 17:00, protect 30 minutes, want two hours of useful reply coverage, and allow 15 minutes for launch uncertainty, the latest planned start is 14:15.
This time rule is only the first gate. A campaign can start early enough and still produce more conversations than the available team can handle. Conversely, a small, low-consequence wave may fit where the full audience does not. Check both timing and volume before approving an exception.
3. Estimate the reply tail, not just the first burst
For comparable past campaigns, group replies by elapsed time from actual delivery where that information is available. Useful groups distinguish immediate replies, replies later in the staffed period, and replies after closing. Keep recipient-local time in view: the sender's afternoon may not be the recipient's afternoon.
Measure new enquiries separately from total message count. One person sending several short messages is not several independent enquiries, although the conversation may still require substantial work.
If delivery timestamps are unavailable, use launch time as an approximation and record that limitation. A spread-out delivery period can make replies appear slower than they were relative to receipt.
With little history, use explicit scenarios rather than a confident-looking average. Test a higher reply rate and a heavier delayed tail. If the decision changes under a plausible alternative, the campaign is too close to the boundary to treat as routine.
There are three quantities to estimate:
- Replies before the protected buffer: the demand the remaining productive time should absorb.
- Replies during the buffer: work that could consume the intended safety margin.
- Replies after shift end: new enquiries likely to wait unattended.
These estimates do not prove that earlier conversations will be finished at closing. Carry-over work needs its own allowance in the capacity figure.
4. Filled fictional worksheet: full send, early wave, or defer
Fictional worked example: all times, quantities, thresholds, and reply patterns below are hypothetical planning assumptions, not customer results or benchmarks.
Shared assumptions
- Decision time: 13:30 today.
- Staffed hours: 09:00–17:00, with no evening coverage.
- Protected buffer: 16:30–17:00.
- Campaign: a non-urgent information offer that remains useful tomorrow.
- Eligible audience: 400 recipients with appropriate permission, excluding opt-outs.
- Expected response: 10% of recipients start an enquiry.
- Reply timing: 40% of enquiries arrive in the first hour after launch; 35% in the second hour; 25% arrive between four and sixteen hours after launch.
- Calculation assumption: replies are evenly distributed within each stated band. There are no additional replies in this simplified model.
- Launch assumption: delivery is concentrated close enough to the stated start to use it for this comparison. A material delay requires recalculation.
- Available capacity today: 12 additional campaign enquiries before 16:30, after existing work and likely carry-over are allowed for.
- Available capacity tomorrow: 40 additional campaign enquiries before 16:30.
- Overnight tolerance: no more than eight expected new unattended enquiries for this low-urgency offer, with uncertainty checked separately.
Fractional enquiry counts below are expected values, rounded for readability. They are not predictions of exact people or a guarantee of a manageable queue.
Option A: send all 400 at 15:30 today
- Total expected enquiries: 400 × 10% = 40.
- Before 16:30: the first-hour share, or 16 enquiries.
- During the buffer: half the second-hour share, or seven enquiries.
- After 17:00: the remaining seven second-hour enquiries plus ten delayed enquiries, or 17.
- Decision: do not send. The expected pre-buffer load exceeds today's capacity of 12, and the expected overnight load exceeds the tolerance of eight.
The failure is not merely that the send is late. It consumes the closing buffer and leaves a substantial part of the response wave without coverage.
Option B: send a smaller wave of 100 at 14:00 today
- Total expected enquiries: 100 × 10% = 10.
- Before 16:30: both early bands have finished, producing 7.5 expected enquiries.
- During the buffer: none under the simplified timing model.
- After 17:00: the delayed band begins at 18:00, producing 2.5 expected enquiries.
- Decision: a reasonable candidate if the assumptions survive a stress check. Leave the other 300 recipients unsent pending a separate decision.
This option reduces exposure; it does not establish that the remainder should go later today. A fast initial response is not evidence that the delayed band has disappeared.
Option C: defer all 400 until 09:00 tomorrow
- Total expected enquiries: 40.
- Before 16:30: 30 early enquiries plus approximately 2.9 from the delayed band, or 32.9.
- During the buffer: approximately 0.4 enquiries.
- After 17:00: approximately 5.8 enquiries, as seven hours of the twelve-hour delayed band fall after closing.
- Decision: the base case fits tomorrow's capacity and overnight tolerance. Approve only after checking tomorrow's actual staffing and uncertainty.
Deferral provides more staffed exposure to the reply curve. It still leaves an evening tail. If all 25% of delayed replies arrived after closing rather than being evenly spread, the full campaign would produce ten unattended enquiries and fail this worksheet's tolerance.
5. Choose the option that survives uncertainty
The worksheet supports rejecting the full afternoon send. It leaves a genuine choice between a small early wave and a full next-day send, rather than pretending one is universally better.
The smaller wave limits today's exposure but creates a second scheduling decision for the remaining audience. Deferral preserves a single larger launch and more working time, but depends on tomorrow's capacity and the delayed-response assumption.
Hypothetical stress check: increase the response rate from 10% to 15%. The small wave would generate 15 enquiries: 11.25 before the buffer and 3.75 after closing under the same timing pattern. It still fits the example's limits, though with little remaining capacity today. The full next-day send would generate 60 enquiries, including approximately 49.4 before the buffer, exceeding tomorrow's capacity of 40.
That changes the decision: the small wave remains a candidate, while the full next-day campaign needs a smaller audience, more verified capacity, or a different day. Waiting until tomorrow is not a substitute for checking the numbers.
Do not automatically choose the largest wave that just fits the base case. Preserve room for estimation error, especially when the audience or offer differs from the campaigns used to estimate reply timing.
6. Keep readiness separate from permission to launch
A prepared campaign can remain unsent. Finishing the content or obtaining template approval does not mean the operational timing is suitable.
For WhatsApp, recipient permission and opt-outs remain relevant regardless of the chosen send time. Approved templates are used to initiate Platform conversations and outside the customer-service window, which opens or resets with a user message and lasts 24 hours. Template approval does not guarantee delivery. Check the current WhatsApp Business Messaging Policy for precise requirements.
Keep campaign preparation, template review, and actual delivery as separate states in the launch decision. A campaign that becomes ready after its latest-start time should return for a timing decision, not proceed simply because preparation is complete.
An automated acknowledgement also does not turn an unattended period into staffed coverage. If automation is involved, maintain a clear human escalation path and do not count it as resolving enquiries unless that outcome has actually been established.
7. Recalculate when the launch slips or the wave changes
A latest-start rule needs a stop condition. Before launch, recheck the actual clock time, remaining staff capacity, and audience size. If any have changed materially, rerun the worksheet.
If a smaller wave has already started, use observable demand to decide whether to release any additional audience. Count new enquiries, unfinished conversations, and remaining productive time. Do not expand solely because the initial reply count is low: the whole problem may be delayed replies.
- Launch slips: calculate from the revised likely delivery period, not the original scheduled time.
- Replies exceed the estimate: withhold any additional authorised waves and reassess the remaining capacity.
- Staff availability falls: reduce the capacity figure immediately rather than borrowing the closing buffer.
- The delayed tail is unknown: choose a smaller exposure or defer to a period with more coverage; do not record unknown demand as zero.
These are recommended operating decisions, not assumptions that a particular scheduling system can pause or alter a campaign already in progress.
8. Turn the next send into a measurable rule
Before the next campaign, record one completed worksheet with the planned start, shift end, protected buffer, audience size, expected response rate, reply-time bands, available capacity, and acceptable unattended exposure. Name the person authorised to approve a late exception.
Afterward, compare three observed counts with the estimate: new enquiries before the buffer, during it, and after closing. Also record unfinished campaign conversations at shift end, so timely arrival is not mistaken for completed work.
Use those observations to revise the next latest-start rule. The useful outcome is a defensible decision: a start time and audience size that fit the team's actual coverage, with the remaining overnight exposure made explicit.
Put the next conversation to work.
Explore Alactic Communications for WhatsApp conversations, campaigns and AI-assisted workflows. Choose the product and plan that fit your team.
Explore Communications