Product Brief: Auto-Reminder SMS for Upcoming Shifts

Kestrel -- auto-reminder SMS for upcoming shifts. A compact product brief a team could act on directly. Kestrel and every number, quote, and interview finding below are illustrative, constructed examples written to teach this topic's techniques -- not a real company, real usage data, or a real transcript.

1. Problem statement

Employees on a Kestrel-scheduled team currently have no way to be reminded of an upcoming shift other than actively checking the schedule themselves. For a shift published several days out, in the middle of a busy week, this is easy to lose track of -- and a forgotten shift produces the same operational damage as a shift someone consciously blew off: an uncovered position, a scramble to find last-minute coverage, and a manager's trust in the schedule eroded. We want the rate of shifts missed specifically because the employee forgot -- as distinct from shifts missed due to a genuine availability conflict -- brought down. This statement does not yet commit to SMS, email, push notifications, or any other specific reminder mechanism; which channel (if any) is worth building is addressed in Section 4.

2. Job story and evidence

When I have a shift three days out that's easy to forget in a busy week, I want a well-timed nudge before it happens, so I show up on time without having to actively remember to check the schedule myself.

Evidence: this problem traces directly to Worked Scenario 19's opportunity-solution tree, where "employees forget upcoming shifts" was named as the first of three opportunities under the "reduce no-show shift incidents by 30%" outcome. In that same illustrative dataset, roughly 40% of no-show incidents tagged that quarter were tagged "forgot," as distinct from "conflict" (the employee became unavailable after the shift was published) or "no longer employed" -- making "forgot" the single largest named cause and the one this brief's MVP targets.

3. MVP scope and non-goals

MVP scope: send an automated SMS reminder to the assigned employee 24 hours before a shift, and a second SMS reminder 2 hours before the shift. Both reminders are sent automatically once a schedule is published; no manager action is required per shift.

Explicit non-goals (deliberately out of scope for this MVP):

• No per-shift or per-employee customization of reminder timing -- the 24h/2h cadence is fixed for every shift in this MVP.
• No reminders for shifts published more than 7 days in advance -- a reminder sent 10 days early provides little value and adds unnecessary message volume.
• No email channel -- SMS only for this MVP; email is a possible later addition, not part of this scope.
• No per-shift opt-out granularity -- an employee can disable reminders globally in their profile settings, but cannot selectively silence reminders for one specific shift while keeping others active.

4. RICE-ranked backlog

Reference units: Reach in employees receiving at least one reminder per quarter; Impact on Intercom's discrete scale (3 massive, 2 high, 1 medium, 0.5 low, 0.25 minimal); Confidence as a percentage; Effort in person-months.

Backlog item	Reach	Impact	Confidence	Effort	RICE
24-hour-before SMS reminder	2000	2	0.8	1.5	(2000×2×0.8)÷1.5 = 2133.3
2-hour-before SMS reminder	2000	1	0.7	1	(2000×1×0.7)÷1 = 1400
In-app reminder banner (no SMS)	1500	0.5	0.9	0.5	(1500×0.5×0.9)÷0.5 = 1350

Ranked order: 24-hour SMS reminder (2133.3) > 2-hour SMS reminder (1400) > in-app reminder banner (1350).

Defended ordering: the 24-hour reminder ranks first because it is the highest-impact, highest- confidence item -- it targets the specific "forgot entirely, days in advance" failure mode named in Section 2's evidence, at a cost proportional to its reach. The 2-hour reminder ranks second: it targets a narrower failure mode (remembered the shift existed but lost track of it on the day itself) with lower estimated impact and confidence, since Kestrel has no prior data on same-day reminder effectiveness. The in-app banner ranks third and lowest: it only reaches employees already actively in the app (smaller effective reach among people who need a reminder in the first place, since the point of a reminder is reaching people who aren't already looking), giving it both a lower estimated impact and a narrower true audience than either SMS item, despite its low cost.

5. Metrics: north-star and AARRR mapping

North-star metric (unchanged from this topic's earlier north-star work): number of teams with a published schedule active in the last 7 days -- this feature does not change the north-star itself, but is expected to support it indirectly.

Supporting metric for this feature: no-show rate -- the % of published shifts with no clock-in recorded and no approved swap on file. This is the feature's direct target metric, distinct from the north-star, and maps to the Retention stage of the AARRR funnel (Worked Scenario 9): a team whose employees reliably show up for their shifts is a team with fewer operational fires, and fewer operational fires is a team more likely to keep renewing its Kestrel subscription.

6. A/B experiment design

Hypothesis: sending an SMS reminder 24 hours and 2 hours before a shift reduces the no-show rate for shifts tagged "forgot," without meaningfully increasing the rate at which employees opt out of Kestrel's SMS notifications generally.

Primary metric (OEC): no-show rate on shifts eligible for the reminder (published more than 24 hours in advance).

Guardrail: the SMS notification opt-out rate must not increase by more than 1 percentage point during the experiment window -- a proxy for notification fatigue. A reminder feature that reduces no-shows while also driving enough employees to disable all Kestrel SMS (including shift-swap notifications, Worked Scenarios 28-30) would be trading one problem for a different, possibly worse one.

7. Internal consistency

• Scope serves the problem: the MVP (Section 3) is scoped specifically around the "forgot" failure mode named in Section 1 and Section 2's evidence -- fixed-timing SMS reminders, nothing broader. The non-goals explicitly exclude scope creep (per-shift customization, additional channels) that would not change whether an employee remembers a shift they'd otherwise forget.
• Backlog reflects the problem's evidence: the highest-RICE item (the 24-hour reminder) is the one most directly aimed at the "days in advance" pattern in the "forgot" evidence from Section 2, not merely the cheapest item to build.
• Metrics measure the outcome: the supporting metric (no-show rate) is the direct, measurable version of Section 1's stated outcome ("fewer shifts missed because the employee forgot"), not a proxy metric several steps removed from it.
• Experiment tests the hypothesis: the primary metric (no-show rate on eligible shifts) is exactly what Section 1's problem and Section 6's hypothesis both target, and the guardrail (SMS opt-out rate) guards against the specific, plausible harm this particular mechanism (SMS reminders) could cause that the primary metric alone would not reveal.

This brief is executable without hand-waving: a reader could hand it to a real product trio, and the team would know exactly what to build, in what order, and how to know whether it worked.