When real-time matters.
Real-time matters during specific events: launch day for a new product, a marketing push you are validating, a flash sale where you might run out of stock. In those moments, knowing in the next hour whether you are on track changes what you do.
Real-time also matters for high-velocity QSR and grocery where a stock-out by 4pm means lost evening sales. If your business is normally steady but spikes on weekends, real-time is overkill on Tuesdays and essential on Fridays.
When end-of-day is enough.
For most retailers, end-of-day is enough for 95% of decisions. Yesterday's sales by branch, by category, by cashier, with comparisons to last week and last month — that is what tells you what is working. You cannot fix yesterday today, but you can fix this week tomorrow.
End-of-day reports are also more accurate. Real-time data has settling artifacts: refunds processed minutes later, end-of-shift adjustments, Z-report finalization. The number you see at 6pm and the number you see at 11pm are slightly different. End-of-day is final.
The metrics that need real-time.
Sales per hour during a launch. Stock level on hot SKUs during a flash sale. Active terminals (is one of your branches offline?). Cash drawer balance during a high-volume shift. Active staff (did the closer show up?). These are operational signals, not strategic ones.
Real-time has to be visible somewhere staff actually look. A dashboard buried in a back-office app that nobody opens is useless. The right place for real-time is a wall display in the back office, a phone notification when thresholds breach, or a daily morning summary on chat.
- Sales velocity per hour vs forecast
- Stock level on featured/promoted SKUs
- Branch terminal availability
- Cash drawer balance approaching cash-pull threshold
- Staff clock-in versus schedule
The metrics that need end-of-day.
Gross margin by category. Discount rate by cashier. Average ticket per branch. Inventory turn by SKU. Refund rate by reason. Customer count by hour. These are pattern questions, and patterns need at least a full day of data to be meaningful.
End-of-day reports should be automated and arrive in your inbox by 9am the next morning. Not in a dashboard you have to log in to — in your email. Owners who have to actively pull reports stop pulling them. Push beats pull.
Weekly and monthly cadences.
Weekly: branch-level P&L comparison, cashier performance summary, slow-moving inventory, customer acquisition by channel, refund and shrinkage trends. Monthly: full P&L by branch with last month and last year comparison, full inventory valuation, AR aging, vendor spend.
Most owners check daily reports, miss the weekly summary, and get blindsided at month-end when the bigger trends emerge. Reverse the priority — weekly summaries are where strategy lives. Daily is for operations only.
The illusion of "we need everything live."
Many owners initially demand real-time everything because it sounds modern. After a month they admit they only check the live dashboard at 11pm to feel reassured. That feeling is the product they really want, and it is not strategic value.
A better solution is a clear daily summary that arrives reliably with all the answers, plus on-demand queries when something specific needs investigation. The real-time dashboard becomes a fallback for unusual days, not a daily ritual. Mental health improves and decision quality stays the same.
Alerts beat dashboards.
Instead of staring at a dashboard, configure alerts for the few things that need immediate action: branch terminal offline for more than 10 minutes, a single transaction over S$10,000, a refund without manager approval, drawer variance over threshold, daily sales below 50% of forecast.
In Nonari alerts go to the branch manager via email or chat depending on configuration. The owner gets a daily digest at 9am. Nobody is staring at a dashboard except during specific events. The signal-to-noise ratio is better than any dashboard can deliver.
Building the right reporting cadence.
Start with daily by-email at 9am, weekly review on Mondays for an hour, monthly close on the 5th of the next month. That is the rhythm. Real-time is for events, not for daily life. Get the rhythm right and your reporting becomes a tool instead of a distraction.
A retailer doing S$5 million per year should spend at most 2 hours a week looking at reports — 30 minutes per morning to scan the daily, an hour on Monday for the weekly. If you are spending more than that, your reporting is too noisy or your processes are not working.