Shopify Profit Tracking

Most stores do not have a profit problem. They have a latency problem.

The arithmetic is rarely wrong. What goes wrong is that it runs once a month, so every decision made in between is made against a number that was assembled weeks ago, from a period that has already ended.

What “continuous” actually requires

Continuous tracking is not a faster spreadsheet. It is a different arrangement, and it needs three things a monthly rebuild does not.

Data that arrives on its own. Orders, refunds, carrier charges, gateway fees and ad spend all have to flow in without anyone exporting a file. The moment a step requires a person, the cadence collapses to whenever that person has time.

A stable join key back to the order. Costs arrive from systems that know nothing about Shopify order numbers. Carrier invoices carry tracking numbers, gateway records carry transaction IDs, 3PL charges carry their own references. Keeping those keys attached as data lands is what makes attribution possible later.

Cost settings with effective dates. COGS, shipping rates and fee percentages change. Without a date attached, a supplier price rise silently recosts every historical order, and last quarter’s margin changes because you updated a number today.

Miss any of the three and you have faster reporting, not tracking.

The four flows and their cadences

Sales and refunds — effectively immediate. Orders can be processed as they land, which means revenue and cost of goods are known within seconds of a sale.

Ad spend — daily. Platforms restate the last few days as attribution settles, so the most recent figures are provisional by nature. Pulling more often than daily adds noise, not accuracy.

Carrier and gateway costs — days to weeks. Labels are billed on the carrier’s cycle, fees settle inside payouts, adjustments arrive later still. Recent orders are always partially costed and should be marked as such rather than presented as complete.

Your own cost settings — whenever they change, with the date they took effect.

The consequence is that a continuously tracked number is a blend: fully costed for older orders, partially costed for recent ones. That is not a flaw to be hidden. It is the honest state of the data, and showing which part is still settling is what stops people treating a three-day-old margin as final.

What to watch daily, weekly, monthly

Daily — contribution margin rate and cost per acquired customer, as rates rather than totals. These are the two numbers that move fast enough for a day to matter, and the two that give early warning of everything downstream.

Weekly — margin by product, by discount code and by channel. Weekly is the shortest window with enough orders in each bucket for a comparison to mean anything, and it is where mix shifts become visible.

Monthly — the full P&L, with fixed costs and comparison columns. Fixed costs arrive in lumps; looking at them more often measures invoice timing rather than performance.

The mistake is inverting this — watching net profit daily, where it is pure noise, and margin by product monthly, where the finding arrives too late to act on.

Thresholds worth alerting on

Alerts should fire on rates crossing lines, not on totals changing.

Contribution margin rate below a floor, set a few points under your normal operating range. This catches promotions running longer than intended and free-shipping thresholds set too low.

Acquisition cost above the margin it has to clear. The comparison that matters is CAC against contribution margin per order, not CAC against a target from last quarter.

A discount code performing below store-average margin. This is where the fortnight in the example above would have surfaced on about day two.

Orders shipping with no cost of goods attached. Uncosted orders count as pure margin and quietly inflate every number above.

Four alerts is close to the practical maximum. Beyond that they get muted, and a muted alert is worse than none because it creates the impression of coverage.

Marking what is still provisional

The hardest part of showing a live number is being honest about which parts of it have settled.

An order placed this morning has known revenue and known cost of goods, an estimated shipping cost, and no chargeback history at all. An order from three months ago has all of those resolved. Presenting both at the same confidence invites people to treat today’s margin as though it were final, then to be surprised when it drifts.

The workable convention is to show the figure with an indication of how much of the period is fully costed — how many orders still have estimated shipping, how much spend is still provisional. It costs one line of interface and it prevents the single most common failure of live dashboards, which is not inaccuracy but false certainty.

Why staleness costs more than imprecision

Given a choice between a number that is accurate to the cent and arrives in three weeks, and one that is within a percent and arrives today, the second is worth more — because a decision informed at the point it is made can change an outcome, and an audit of a period that has already closed cannot.

This is not an argument for tolerating sloppy inputs. Uncosted products and stale rates need fixing, and precision matters enormously for the P&L that goes to an accountant or a lender.

But the operational value of profit data is almost entirely a function of how quickly it can change what you do next. The example above cost $2,256 not because anything was calculated incorrectly, but because the correct calculation ran fourteen days after it would have been useful.

Which tools do this

Continuous tracking is a category rather than a feature, and the products in it meter and model costs differently. Best Shopify profit tracking apps ranks them against the criteria above and names what each one, ours included, is worse at.