Most small retailers know their busy days. They have a feel for the morning rush, a sense that Saturdays beat Tuesdays. But that intuition usually lives in the manager's head and gets baked into a schedule that hasn't been revisited in months. Underneath the daily totals is an hourly pattern that explains where revenue actually comes from — and where margin quietly evaporates when the floor isn't matched to the curve. Day-part sales analytics turns that hidden pattern into a working tool, and the data you need is already sitting in your POS.

What Day-Part Analytics Actually Is

Day-parting is the practice of breaking your operating day into time blocks — typically morning, midday, afternoon, evening, and late — then measuring revenue, transactions, average ticket, and (where you can) conversion within each block. The chain QSRs have done this for decades. They don't run breakfast and dinner with the same staffing, the same menu emphasis, or the same promotions, because the customer in front of them isn't the same person.

Independent retailers have access to the same data, just at a smaller scale. Every transaction in your POS already carries a timestamp. Joining 90 days of those timestamps into a day-of-week × hour grid gives you a full picture of when your store actually earns. The grid is the foundation; everything else in this guide is how to read it and what to do with it.

A useful baseline:

  • Day-part block: typically a 2–4 hour window (e.g., "lunch" = 11am–2pm)
  • Cell: a single intersection of weekday and hour (e.g., "Saturday 1pm")
  • Money hours: the cells generating the top 20–25% of your weekly revenue
  • Thin hours: the bottom 25% of cells by revenue

You're going to spend most of your operational attention on those two extremes.

The Three Patterns Every Day-Part Curve Reveals

When you plot 90 days of hourly revenue, the curve almost always falls into one of three shapes. Knowing which one you have changes what levers actually work.

Concentrated Peaks A concentrated curve has 50–65% of weekly revenue compressed into 8–12 hours. Think Friday 4pm–8pm and Saturday 11am–6pm carrying the week. This pattern is most common in apparel, specialty retail, and dispensaries near nightlife corridors.

The risk here is capacity ceiling masquerading as demand ceiling. When the line is out the door at 5pm, you can't tell whether you've reached the natural top of demand or whether you're losing customers who walked past, saw the wait, and kept going. Concentrated patterns reward investment in throughput at the peak: a second register, a queue manager, a faster checkout flow.

Steady Plateau A plateau curve spreads revenue more evenly — say, 70% of revenue distributed across 30+ hours of the week. Convenience-driven categories (pet supply, hardware, neighborhood grocery) often look like this. The pattern looks healthy on a daily total but masks a real problem: you're staffing every hour roughly the same, and your payroll-to-revenue ratio is dragged up by the average hour, not the peak hour.

Plateau retailers usually have the most room to recover margin from scheduling alone. Cutting one labor hour from each of the bottom 12 cells in the week — without losing a sale — is a quiet 3–6% reduction in payroll cost.

Twin Peaks Twin-peak curves show two distinct daily highs, usually a midday bump and an after-work bump. Common in coffee, food retail, and dispensaries that serve both daytime and evening customer segments. The trap here is treating the two peaks as the same business. Customer mix, basket composition, and service expectations differ. A twin-peak curve almost always benefits from a deliberately different floor presence between the two — different signage, different recommended add-ons, sometimes different staff.

Where the Margin Hides: Conversion at Off-Peak vs Peak

Revenue is one layer. Conversion is the layer underneath, and it's where most operators are most surprised.

If you have door-counter data — or even rough proxies like web visits to a "today's deals" page, or call volume — pair it against POS transaction count by hour. A common finding:

  • Peak hours convert at 18–25% of foot traffic
  • Off-peak hours convert at 35–50%

The pattern isn't universal, but it shows up often enough to be worth checking. The reason is intuitive once you see it: peak-hour visitors are more likely to be browsers driven by social momentum, off-peak visitors are more likely to be intentional buyers who came specifically to transact, and off-peak staff have more time to actually help.

The actionable takeaway is uncomfortable. You may be over-investing in peak hours and under-investing in conversion at the shoulder. A small staffing addition at 10am — when your one closer is also receiving a vendor — could lift conversion 5–8 points on a window that's already efficient. The same dollar at 6pm probably gets absorbed by a line that was going to clear anyway.

Building the Day-Part Report

You don't need a BI platform to build a usable day-part view. A spreadsheet and a clean POS export will get you there. Here's the minimum viable workflow:

  1. Pull the timestamp data. Export every transaction for the last 90 days with: date, time, gross sale, net sale, transaction count, item count, and (if available) customer ID and salesperson.
  1. Bucket by weekday and hour. Add a "day-of-week" column (Monday–Sunday) and an "hour" column (0–23). You now have a long table where each row is a transaction, tagged with its time slot.
  1. Aggregate into a grid. Pivot the data so rows are weekday and columns are hour. Compute per cell: total revenue, transaction count, average ticket, items per ticket. That's your base grid — typically 7 rows × 12–14 hours of operation.
  1. Identify the money hours. Sort all cells by revenue. The top 8–12 cells almost always carry 40–55% of your weekly revenue. Mark them.
  1. Identify the thin hours. Sort the bottom 8–12 cells. Together they often represent under 5% of revenue but 20–30% of your open hours. Mark them too.
  1. Layer in conversion if you have it. For each cell, compute conversion = transactions ÷ visitors. Look for the gap between your highest-revenue cells and your highest-conversion cells. They are almost never the same set.
  1. Set a refresh cadence. Day-part curves shift with seasonality, marketing changes, neighborhood foot traffic, and competition openings. Rebuild the grid quarterly. Anything older than 90 days is decoration, not data.

What to Do With What You Find

The grid is diagnostic. The real work is reshaping operations around it.

Schedule against the curve, not the calendar. Most retailers schedule by shift template — a 9-to-5 and a 1-to-9 every weekday — because that's how shifts have always looked. The grid says some hours need three people and some need one. Build the schedule from the grid backwards: start with the money hours, fully staff them with your strongest closers, then taper.

Use thin hours as production time. The bottom-quartile cells are when you do the work that doesn't require selling: restocks, transfers, vendor receiving, training, deep cleans, system updates. Most retailers use peak-day morning for this and burn their best staff capacity on stocking shelves at 10am Saturday.

Test in the quiet windows. A new merchandising display, a price change, a script change for upsells — test it in a thin cell where the downside is small. Once you have a read, roll it forward into a peak window.

Restructure promotions to lift, not transfer. A 15% off promotion that runs all day usually transfers margin to customers who would have bought anyway during peak hours. The same promotion limited to a specific shoulder window — say, 10am–noon Tuesday/Wednesday — has a chance of lifting incremental volume into hours that need it. This is the single biggest day-part-driven margin recovery for many small retailers: roughly 15–25% of promotional spend reallocated rather than added.

Order receiving against the curve, not against the truck schedule. If your distributor offers Tuesday morning or Thursday afternoon delivery, take the one that lands in your thinnest cell. The cost of an aisle being half-blocked at 10am Tuesday is much less than the same aisle being blocked at 2pm Saturday.

Common Pitfalls

A few traps to avoid as you build this practice.

  • Don't read totals through a single outlier. One promotional Saturday or one weather-disrupted Wednesday can warp the curve. Either drop the obvious outliers or use median rather than mean for cell values.
  • Normalize for holiday weeks. Memorial Day weekend and a normal Memorial Day-adjacent weekend should not sit in the same average. Tag them.
  • Compare like store to like store. If you have multiple locations, run separate grids. A pattern that's true downtown is rarely true in the suburbs.
  • Don't confuse demand with capacity. A flat ceiling at 5pm Saturday isn't necessarily the top of demand — it might be the top of what your current floor can serve. The signal is whether transaction count plateaus while average ticket also plateaus. If both flatten, capacity is binding.
  • Refresh quarterly. Curves drift. Neighborhoods change. A grid built off Q1 data and used to schedule Q4 will get most of the close hours wrong.
  • Don't optimize so hard you become brittle. Some buffer in the schedule absorbs the small surprises — sick calls, an unexpected rush. A grid is a guide, not a constraint.

The Bottom Line

Daily totals tell you whether the week was good. The day-part grid tells you where the good came from — and where the leak is. For most small retailers, the gap between operating against intuition and operating against the grid is worth a measurable point or two of payroll cost, a few points of conversion at the shoulders, and a quieter restock day that doesn't burn peak-hour staff.

Three takeaways worth acting on this week:

  • Build the 90-day grid. A spreadsheet pivot of POS timestamps gets you 80% of the value of a BI tool.
  • Find your money hours and your thin hours, then schedule and merchandise differently against each. The grid is only useful when it changes a decision you were going to make anyway.
  • Refresh quarterly. Curves shift, and a stale grid quietly leads to stale schedules.

At Chapters Data, we help small retailers and dispensaries pull the patterns hiding in their POS timestamps into a working operational picture — so the floor, the schedule, and the promotion calendar all line up with when revenue actually shows up.