How Skimmr computes your profit and estimates

This is the trust anchor for Skimmr. It describes what the app actually computes, in the same order the code runs. Nothing here is marketing. If a number appears in the product, you can trace it back to one of the formulas below.

Net profit per order

Skimmr starts from the sale price of one order and subtracts every cost that touches that order. The result is your real net profit per order, not your revenue. Each deduction below is a named line you can see and check.

First Skimmr computes profit before ad spend by subtracting these from the sale price:

• Product cost: what you pay your supplier for one unit.
• Shipping cost: the estimated cost to ship that one unit.
• Packaging cost: the cost of the box, mailer, or insert per order.
• Platform fee: the sale price multiplied by your platform fee percentage. For example, a 5 percent fee on a $30 order is $1.50.
• Creator commission: the sale price multiplied by your creator commission percentage. This is the cut paid to the creator or affiliate who drove the sale.
• Refund reserve: the sale price multiplied by your estimated refund loss percentage. This sets aside money for orders that get refunded so the per-order number stays honest.

So profit before ads equals sale price minus product cost minus shipping minus packaging minus platform fee minus creator commission minus refund reserve. The three percentage costs (platform fee, creator commission, refund reserve) are each a percentage of the sale price, while product, shipping, and packaging are flat dollar amounts per order.

Then Skimmr subtracts your estimated ad cost per purchase to get profit after ads. Ad cost is entered as a flat dollar amount per order, not a percentage. Profit after ads is the bottom line Skimmr treats as your real net profit per order.

Margin percentage is profit after ads divided by sale price, times 100. If the sale price is zero, margin is reported as zero rather than dividing by zero.

From the same inputs Skimmr also derives three planning figures so you can see your headroom: the break-even ad cost (how much you can spend acquiring one order before profit before ads reaches zero), the break-even creator commission as a percentage, and a recommended maximum product cost that would keep you at a 25 percent target margin after all the other costs. These are guides for pricing and sourcing decisions, computed from the same line items.

Product score and verdict

A product score is a 0 to 100 number built from three parts, and the parts are always shown so you can see why a product scored the way it did.

• Margin strength: how strong the profit margin is, measured against a 45 percent reference. A 45 percent margin reads as full strength on this part.
• Upside: the average of four ratings you set from 1 to 10 (impulse buy potential, content potential, problem intensity, and visual demo potential), expressed as a percentage.
• Risk: the average of four risk ratings from 1 to 10 (saturation risk, supplier risk, shipping complexity, and refund risk), expressed as a percentage. Higher risk lowers the score.

The final score weights margin strength at 45 percent, upside at 40 percent, and the inverse of risk at 15 percent, then clamps the result to the 0 to 100 range. Ratings are kept within 1 to 10 before they are used, so an out-of-range entry cannot distort the math.

Alongside the score, Skimmr writes a short verdict made of plain-language reasons drawn directly from the same numbers: whether net profit after ad cost is healthy, whether the margin has room for commissions and testing mistakes, whether content potential looks useful for short-form testing, and whether the risk ratings are elevated. The verdict explains the score; it does not add hidden inputs.

Estimates for products you do not sell yet

For a product you are not selling yet, Skimmr has no real order history, so it produces a labelled estimate of monthly demand. This is a heuristic model. Its output is clearly marked as an estimate and it exposes the basis it was built from.

The demand estimate takes the sale price and, where you have provided them, a signal strength from 0 to 100, an impulse buy rating from 1 to 10, and a saturation risk from 1 to 10. Stronger market signal raises the estimate, higher impulse appeal raises it, higher saturation lowers it, and a higher price dampens unit volume. The model multiplies these factors to produce estimated monthly units, then multiplies units by price for estimated monthly revenue.

Every demand estimate ships with a confidence value between 0 and 1 and a short basis list naming the signal strength, the impulse and saturation ratings, and the price volume factor that went into it. Confidence rises with stronger signal and falls with higher saturation. This is the confidence range, shown next to the estimate so you can weigh it.

Calibration: estimates improve as real orders arrive

When you connect a shop and real orders come in, Skimmr calibrates its estimates against what actually happened. This is the flywheel: the more real order records exist, the more the estimate is pulled toward reality.

Calibration compares estimated units to actual units and produces a factor that adjusts the estimate. With few samples the factor stays close to 1, so a handful of orders cannot swing an estimate wildly. As the number of real order records grows toward 30, the actuals are trusted more and the factor moves further from 1. With no real orders yet, Skimmr keeps showing the raw estimate and says so plainly.

Where each number comes from

Skimmr labels the provenance of values so you always know whether a number is real, entered, or modeled. The labels are:

• Estimated: produced by a Skimmr model or a default assumption, not from a real order. Carries a confidence range where applicable.
• Manual: a value you typed in yourself, such as your supplier cost or packaging cost.
• Imported: a value brought in from a file or export you provided. Raw imports are kept separate from normalized data.
• API synced: a value pulled directly from a connected shop or platform through its API.
• AI generated: a value or piece of text produced by an AI assistant. It is treated as an assumption to review, never as a confirmed fact.

When a real, manual, imported, or API-synced value becomes available, it can replace an estimate without erasing where the original came from. The point of the labels is that you can always tell which kind of number you are looking at.

Our honesty statement

Skimmr does not fabricate numbers. It does not invent sales, orders, imports, or performance you did not have. Example and seed values are shown as examples, never as real operating performance. Estimates are always labelled as estimates and carry a confidence range. AI-generated content is marked as such and treated as an assumption to check.

If you can see a number in Skimmr, you can see how it was made. Connect a shop and walk the math yourself.