57 channels · 43 markets · one database

Margin you can
audit, not
estimate.

Ad spend in one tool. Stock in another. Marketplace settlements weeks late, minus deductions nobody reconciles. ZERPHIX holds all of it on the same rows — so contribution is computed, not assembled after the fact.

Schema-per-workspace isolationMulti-currency booksPluggable tax by jurisdiction
amazon.comCS-SERUM-VITC-30 ×2$47.98ingested · journal posted
Contribution · one SKU, one month
Net revenue$761,663
Cost of goods−$428,900
Channel fees & payment−$121,866
Shipping and returns−$68,550
Ad spend−$54,300
Contribution$88,047
Gross margin says 43.7%. Contribution says 11.6%. The difference lives in four other tools.
Landed cost · day 63
$4.12
100% of components actual
The gap

What your dashboard says. What it actually made.

Same month, same SKU, same orders. The right-hand column is the left-hand column with the four lines your reporting stack cannot reach — and it is the only one you can bank.

Your dashboard
Channel exports joined on SKU, once a day.
Revenue$761,663
Cost of goods−$428,900
Channel feesnot visible
Shipping and returnsnot visible
Ad spendnot visible
Gross profit$332,763
43.7% gross margin
ZERPHIX
The same rows the warehouse and the ledger wrote.
Revenue$761,663
Cost of goods−$428,900
Channel fees$0
Shipping and returns$0
Ad spend$0
Contribution$332,763
43.7% contribution margin
0
Console screens on one database
0
Channels modelled end to end
0
Markets in the catalogue
0
Allocation bases per cost component
Market-aware, not translated

A US seller should never be offered a Blinkit connector.

Most “global” software is one product with a currency switcher. Commerce is not like that: the channels, the tax, the settlement shapes and the return economics are different per country, and a connector list that ignores it tells your customer you have never sold in their market. Pick one and watch the console change.

Connect a channel · United States15 available · 42 hidden as not applicable
Amazon.comMarketplace
Walmart MarketplaceMarketplace
eBayMarketplace
EtsyMarketplace
TikTok ShopMarketplace
Target PlusMarketplace
NeweggMarketplace
ShopifyStorefront
BigCommerceStorefront
WooCommerceStorefront
Walmart Drop Ship VendorB2B / dropship
WayfairB2B / dropship
FaireB2B / dropship
InstacartRetail media
DoorDashRetail media
What else changes
Books kept in
USD
Tax engine
Sales tax · destination-sourced, per state
Modules
Gated by market, plan and an operator override

A module that does not apply to a market is hidden, not locked. “Not applicable here” and “not in your plan” are different sentences, and showing the second when you mean the first is how software reads as foreign.

One pipeline

Every channel. One ingest path. No copies.

A marketplace seller API, an EDI document set, a vendor-panel report and a storefront webhook all run the same ingest → normalise → store path: dedupe, tax derivation, fee model, journal, stock. What you watch on the right is what happens on a sale morning.

AmazonSeller API · 6hWalmartSeller API · 6heBaySeller API · 6hEtsySeller API · 6hTikTok ShopSeller API · 6hWayfairEDI 850/810InstacartVendor report · dailyShopifyWebhook · realtimeONEDBNORMALISE
channel SKU → master SKU · tax · fees
STORE
orders · stock ledger · journal · attribution
ANSWER
contribution · POAS · landed cost — live
sync log · livescheduled · workspace timezone
06:00:02amazonpull started · 412 rows
06:00:03amazondedupe on (channel, external_order_id) · 7 skipped
06:00:03amazontax resolved · destination state · 6.35%
Landed cost

A cost that gets more right as the paperwork arrives.

Five documents, five counterparties, sixty-three days — each allocated on its own basis, because one basis for a mixed shipment misprices half a catalogue. Scroll: this is what converging looks like.

True unit cost
$0.00
Estimated from the lane, the forwarder and this supplier’s history.
0% of components actualfinal at day 63
day 0Supplier invoiceGoods value, ex-worksdirect
day 34Freight invoiceNingbo → Long Beach, FCLby weight
day 41Customs entryDuty by HS code; import VAT recoverableby HS code
day 47Terminal & brokerHandling, documentationby volume
day 63DemurrageNine days, 40% already soldby volume
The product

Four screens, one set of rows.

Nothing here is joined after the fact. The number on the dashboard is the number in the report is the number in the ledger.

CONTRIBUTION · LAST 30 DAYS
$88,047
11.6% marginon $761,663 net revenuegross margin would say 43.7%
WHERE THE REVENUE WENT
BLENDED POAS
1.62×
ad spend
RETURN RATE
4.4%
of delivered orders
CHANNEL FEES
$121,866
net revenue 16.0%
ALERTS
47
2 zero-sales · 45 low stock
01 / 04

The Monday morning answer

Contribution first, immediately undercut by what gross margin would have told you. Six figures, a watchdog against your own baseline, and where the revenue went — every one a query on the operational tables, not an overnight extract.

OrdersContributionAd spendReturnsStock
CHANNEL
SKU
REVENUE
CONTRIB
MARGIN
Amazon.com
CS-SERUM-VITC-30
$214,880
$31,420
14.6%
Shopify
CS-DERMA-SPF50-50
$168,220
$38,100
22.6%
Walmart
CS-PRO-ONION-300
$141,055
$9,880
7.0%
TikTok Shop
CS-MEN-BEARD-50
$118,400
$4,210
3.6%
eBay
CS-KIDS-SHAM-200
$74,310
$2,980
4.0%
02 / 04

Ask, then act — in the same screen

Pick a model, pick measures, pick dimensions. Every measure declares in code what it means and which operational screen it reconciles with, so two people asking the same question get the same number.

CREATIVE
SPEND
ROAS
CONTRIB
VERDICT
UGC · before/after · 15s
$18,400
4.1×
$12,880
ok
Static · founder quote
$9,200
3.8×
$7,410
ok
Carousel · bundle offer
$21,600
3.9×
−$2,140
discount eats it
UGC · unboxing · 30s
$6,400
2.2×
$1,180
thin
Static · price-led
$14,900
4.4×
−$880
returns eat it
Two of the five best ROAS rows lose money. That is the whole screen.
03 / 04

Creatives ranked by what they earned

Not by ROAS. By the contribution the units actually generated, net of cost of goods, channel fees and returns — with a flag on the rows whose ROAS looks fine and whose economics do not.

Amazon.com
Seller API
settled to 14 Aug
Walmart
Seller API
variance $214 open
Shopify
Webhook
realtime
eBay
Seller API
settled to 15 Aug
Wayfair
EDI 850/810
3 POs unacknowledged
Etsy
Seller API
commission +2.1pt
TikTok Shop
Seller API
settled to 12 Aug
Instacart
Vendor report
daily 06:00
Faire
Seller API
net-60 terms
04 / 04

Every channel, one order lifecycle

Marketplace, storefront, EDI vendor and retail media land in the same orders, the same stock ledger, the same books — each with its fee model and settlement shape already built, and each hidden entirely in markets where it does not exist.

Where it fits

Replacing three tools, or sitting beside them.

What you use nowWhat it does wellWhat ZERPHIX adds
An OMS or WMS
Channel sync and warehouse suites
Runs the warehouse and the channel sync properly.Ad spend, contribution, landed cost and the ledger on the same rows the warehouse moved.
An analytics layer
Profit dashboards on top of your channels
Good margin maths, fast to switch on.The ability to act — pause the creative, reprice the SKU, raise the replenishment — without leaving the screen.
Spreadsheets
For landed cost and settlement reconciliation
Nothing. It is a spreadsheet.A cost that converges as documents arrive, and a claim raised before the window closes.
Before you ask

The questions we would rather answer now.

Not out of the box. The fee models, settlement formats, report parsers, EDI documents and claim windows are built and tested for 57 channels across 43 markets. The seller-API client for each marketplace — the OAuth app, the credentials, the rate limits — is engineering work we do during implementation. It is the last mile, and we would rather say so than let you find out in week two.

See it on your own numbers.

Forty minutes. Bring one settlement report and we will show you the deductions nobody reconciled.