Identify the actual product
Keep model, configuration, condition, and source evidence together. Flag uncertain matches before treating two listings as comparable.
For small cross-border sellers
Know your margin.
Before you buy.
A price gap is only the beginning. Margin Desk is building a clearer way to compare products, account for costs, and decide what deserves your inventory budget.
Early prototype · Initial focus: Canada and the United States
Purchase decision / 001Illustrative example
The price is not the profit.
One product. Every cost in view. All amounts in CAD.
Expected sale revenue$200.00
Purchase cost− $95.00
Shipping & packaging− $30.00
Platform & currency costs− $28.00
Loss reserve− $10.00
Estimated contribution
After the listed costs$37.00
After the listed costs$37.00
Example inputs, not a live listing or a forecast. Other applicable costs must be included before a real decision.
Product identity before price comparisonCosts before inventory allocationEvidence before confidence
The problem we solve
Small omissions.
Expensive decisions.
Different bundles can look identical. Shipping can erase the spread. An old price can turn a promising purchase into unwanted stock. We are bringing these checks into one decision workflow.
Make the margin explicit
Evaluate purchase price, shipping, platform fees, currency conversion, and other applicable costs with a visible loss reserve.
Set a purchase boundary
Compare estimated contribution, break-even price, and maximum purchase price. Test less favorable assumptions before committing funds.
Product workflow
From a candidate
to a considered decision.
Our prototype focuses on evaluation. Data integrations and AI assistance are being developed and validated separately.
Collect evidence
Record the product, source, currency, and date behind each assumption.
Resolve differences
Review model and bundle mismatches; keep missing information visible.
Evaluate costs
Calculate the contribution after known costs and compare a downside case.
Decide & review
Let the seller make the decision; use outcomes to improve the workflow.
Try the basic margin calculation
Illustrative calculator · CAD · Inputs stay in your browser
Estimated contribution
CA$37.00
With sale revenue 10% lower: CA$17.00
Revenue − purchase cost − other costs − reserve. Include all relevant costs in your inputs. This simplified example excludes income tax and does not predict sales.
PLANNED CLAUDE API INTEGRATIONAI for messy information.
AI for messy information.
Code for the numbers.
We plan to use Claude to help structure product information and explain evidence gaps. Financial calculations and budget rules remain in deterministic code.
Claude integration is on our development roadmap. We do not claim an Anthropic partnership or program acceptance.
- Structured product attributesExtract models, specifications, and bundle details from user-provided documents and permitted sources.
- Evidence reviewFlag possible inconsistencies and missing fields, with source references where available.
- Clear scenario explanationsExplain calculated results and uncertainty without presenting generated text as verified market data.
- A human decision pointRequire seller review before purchase or listing decisions. No autonomous buying or guaranteed returns.
Built around trust
Useful analysis starts
with accountable inputs.
Trace the assumption
Keep the origin, age, and currency of evidence attached to the evaluation.
Show what is missing
Unknown costs and uncertain identity should remain visible rather than become invented certainty.
Explain the calculation
Make inputs and calculation rules reviewable so a seller can understand why a result changes.
About Margin Desk
A focused start.
A practical ambition.
Margin Desk is an early-stage software project for small cross-border sellers. Our initial focus is evaluating products sourced in Canada for sale in Canada and the United States.
Our long-term goal is to connect sourcing evidence, cost assumptions, and actual sales outcomes in an auditable decision-support platform.
Our operating model spans business strategy, product engineering, and seller operations. We are developing the prototype and seeking conversations with prospective pilot users.
Development priorities
- PrototypeRefine product evaluation, cost inputs, and conservative margin scenarios.
- ValidationTest workflows with prospective sellers and verify permitted data integrations.
- AI supportEvaluate Claude-assisted attribute extraction and evidence review against representative examples.
- BusinessValidate demand for a subscription product through a small pilot.
Make the next purchase
a better-informed one.
Building a cross-border selling business? Tell us where your product research and cost checks get difficult.