Two staff members work the same Saturday shift. One consistently upsells sides and drinks without being asked; the other spends half the night on their phone in the pass-through. The owner suspects this, vaguely, from walking past a few times — but suspecting isn’t the same as knowing, and a performance conversation built on “I have a feeling” doesn’t land well with either the staff member who’s coasting or the one who’s actually earning their keep, and it gives the manager nothing concrete to point to when it’s time to actually have the conversation. The gap wasn’t a lack of trust. It was a lack of any data to check the feeling against.
Why watching more closely doesn’t fix it
Standing over staff, checking in constantly, or reviewing camera footage after the fact all have the same problem: they’re either intrusive, or they’re after-the-fact and prove nothing about a specific shift’s sales. What’s actually missing isn’t supervision — it’s a record, one that exists whether or not anyone was watching, showing what each person actually did on their shift: what they sold, what was in the till when they closed out, and what hours they actually worked versus what the roster says they were meant to work.
What actually needs to happen
Let shifts track themselves
Shift tracking where staff start, pause and close their own shift means the record exists automatically — sales, cash in the drawer, and hours worked, per person, in real time, without a manager needing to stand there with a clipboard. It’s not surveillance; it’s the same data a till already has, just attached to the person who was actually on it. This works the same way whether it’s a food truck at a Saturday market or a nightclub bar running until 3am.
Make cash discrepancies visible the same night, not at the next stocktake
When cash in the drawer is tracked per shift rather than per till for the whole day, a discrepancy shows up the same night, tied to the person who closed out — not three weeks later during a stocktake when nobody can remember what actually happened on any given Tuesday.
Give each role its own view
Role-based dashboards mean a casual server sees their own shift and sales, a manager sees the whole floor, and an owner sees it across every shift and every team member — nobody’s drowning in a report meant for a different job, and nobody feels like they’re being watched by a system built for someone else’s role. A ghost kitchen running three brands out of one site, or a food hall with a dozen independent stalls, needs this separation just as much as a single-site restaurant does.
Let end-of-day reports do the arithmetic
Sales by item, hours worked, cash variance — finalised the moment a shift closes rather than reconstructed from memory the next morning. A manager checking Monday’s numbers on Tuesday afternoon is already a day behind; a report that’s ready the second the till closes means the conversation can happen while the shift is still fresh for both people in it.
Turn the numbers into a conversation, not a verdict
The point of this data isn’t to catch anyone out — it’s to make a performance conversation specific instead of vague. “You averaged $340 in upsells last week versus $190 the week before, nice work” is a completely different conversation from “you seem to be doing better lately,” and it’s the kind of feedback staff actually respond to, because it’s clearly about the work, not a mood.
Data replaces guesswork, not trust
None of this is about watching your team more closely — it’s about not needing to. Once sales, cash and hours are tracked automatically, per shift, per person, you stop managing on gut feel and start managing on what actually happened — which, more often than owners expect, ends up backing up the staff member they suspected was doing a great job all along, not just catching the one who wasn’t.