Apply for access ↗
MARKETPLACE INTELLIGENCE

Amazon

Understand what sells, what needs restocking and which costs sit behind your contribution.

MELTH brings your Amazon products, sales and inventory together with your own cost files. Give your commerce, operations and finance teams the context to investigate a product and review the evidence behind a decision.

The sale is in Amazon.
The cost is in a spreadsheet.
The question is yours.

Which product needs attention? How much stock is really available? Can we trust the cost behind this calculation?

Useful answers need more than another export. They need product identity, the right period, a meaningful stock context and source values that someone has actually reviewed.

01 /

Platform data

Your own Amazon connection, with marketplace and period context.

02 /

Your sources

Excel and CSV files, with inspectable versions and evidence.

03 /

AI assistance

Ask questions, explore patterns and understand the limits of an answer.

04 /

Your judgement

Review proposed cost values before they become active.

05 /

Traceable insight

Follow a finding back to the source and the decision behind it.

One product.
A connected line of enquiry.

Start with a business question. Move through the context. Keep the evidence with the answer.

  1. 01Recognise itProduct & listings
  2. 02Understand itSales & stock context
  3. 03Question itCosts & contribution
  4. 04Verify itSource & review

Workflow explanation. This is not a live dashboard or a claim that MELTH executes actions in Amazon.

Different listings.
A recognisable product.

Follow a product across its listing, inventory and sales context, without rebuilding the story from separate exports.

A canonical catalogue brings ASIN identity and associated SKU/listing context together. Product images, titles and normalised brands help your team recognise what they are looking at.

  • Focus the view. Search and filter products, choose columns and move through server-side pages.
  • Choose the period. Investigate Sales & Traffic history within the dates and coverage available.
  • Open the context. Use product detail to explore linked information and its origin.

A marketplace filter describes sales context. It does not prove that every country-analysis module or every historical period is available.

Where you sell
is not always
where stock sits.

Understand inventory through fulfilment pools, not a misleading sum of country totals.

Multiple marketplaces can draw on a shared European stock pool. That stock should not be counted again for every marketplace. EU and UK networks need their own context where applicable.

  • Sold per week adds the measurement period to the sales pace.
  • Weeks cover relates stock to available sales evidence.
  • Reorder advice prepares a discussion, with its assumptions and completeness visible.

Local and remote pool labels need the product’s actual definition. Advice is not autonomous purchasing or a guaranteed forecast. Missing history stays incomplete.

See what you can calculate.
And what you cannot. Yet.

A contribution figure is only useful when you can see what went into it.

Connect revenue with available Amazon financial events and applicable, approved cost prices. Keep currencies separate and gaps visible.

01 / AVAILABLE EVIDENCE

After Amazon costs

Understand proceeds after the Amazon costs available for the relevant period.

02 / APPLICABLE COST BASIS

After cost of goods

Calculate contribution where approved unit costs, currency and coverage support it.

03 / WIDER BUSINESS PICTURE

Not complete net profit

Advertising, other business costs or missing evidence may sit outside the calculation.

THE RULE

Missing cost ≠ zero cost.
Different currencies ≠ one euro total.

Without a matching cost currency or governed conversion, contribution after COGS is not calculable. A high coverage percentage does not remove a currency mismatch.
Understand unit cost, COGS and coverage +

VVP: the managed cost per sellable unit in MELTH. The product, unit, currency and effective date all matter. A single item, multipack and gift set are not interchangeable.

COGS: the cost of the units sold. Coverage: how much of the relevant data has applicable evidence. Product coverage and revenue coverage have different denominators.

Start with a spreadsheet.
Keep the story behind it.

Your files hold decisions, assumptions and knowledge. Bring them into a workspace where that context can be inspected.

EXPLORE BEFORE YOU IMPORT

Understand what is there.

Add Excel or CSV as a business source. Inspect worksheets, search the contents, download the original and retain its version. Set a general analysis goal or specific intended uses, and revise those uses later.

Suggested purposes can follow recognised columns. A purpose label is context, not an automatic write to another system.

ASK MELTH

Ask a better question.

Investigate supported questions such as duplicate SKUs or which currencies are present. Review the source version, worksheet and analysis coverage with the answer.

Open questions may use column structure and a labelled sample. An AI explanation does not imply that every row was analysed.

AI helps you understand.
You decide what becomes valid.

The concrete cost-price review chain within MELTH:

  1. 01

    Original source

    The file, version and evidence rows.

  2. 02

    Prepared proposal

    Product match, unit, value and validity. A normalised source if needed.

  3. 03

    Human review

    Inspect the evidence. Approve eligible values or leave questions open.

  4. 04

    Active cost

    The approved basis flows into product detail and applicable calculations.

Individual, selected and bulk approval +

Approve one suitable proposal, explicitly selected suitable rows, or all ready and eligible proposals. A filter is not approval of hidden rows. Unresolved and batch exceptions do not silently become accepted costs.

Reject, correct or withdraw +

Reject a pending proposal. Correct or withdraw an individual accepted value. Preserve the previous value, reason and source history. A replacement is different from a withdrawal; bulk withdrawal is not claimed.

Return to the evidence +

Follow an active value through the decision, evidence rows and source version. Where an intermediate normalised source exists, retain the path back to the original file. Open cost questions stay visible.

A useful answer.
An explainable decision.

Know which data you are using, how recently it was updated and what has actually been approved.

↗

Your connection

Customer-owned Amazon authorisation and customer-scoped data and sources. The described Amazon integration reads data; it does not execute changes.

◷

Visible freshness

Automatic synchronisation brings platform data into MELTH. Check the relevant update time and available history before drawing conclusions.

⤴

A trail you can follow

Source versions, evidence and review history help reconstruct the decision. No blanket claim that every external system is audited.

Different responsibilities.
Shared evidence.

OWNER / FINANCE

“Where does this
number come from?”

Move from contribution to product detail, then to the applicable cost source and original version. Review which costs and currencies the figure covers.

Follow the calculation ↗
E-COMMERCE / BUYING

“Which products
need a closer look?”

Start with a period and product filter. Inspect performance, the fulfilment pool, weekly sales and coverage before your stock discussion.

Understand the context ↗
OPERATIONS / MANAGEMENT

“Can we use
this cost price?”

Inspect the file, product match and unit. Review suitable proposals. If an approved value is wrong, correct or withdraw it with history retained.

Explore the review chain ↗

Begin where your
business works today.

Make your current files understandable first. Build a more deliberate process from there.

CURRENT FOUNDATION

From files to governed costs.

Excel/CSV → inspection and analysis → supported mapping and proposals → human review → managed cost prices inside MELTH → traceable product and contribution calculations.

DEVELOPMENT DIRECTION

From process knowledge to connected systems.

Over time, this foundation can support a transition towards systems such as Odoo or Zoho. Migration and external execution require their own validated connectors, permissions, checks and recovery paths.

Next-stage work: Buy Box, expanded marketplace/country analysis, Deals, Advertising, a full alerts module and additional source connectors. These are not presented here as completed features. Scope is confirmed for each selected workspace.

A little more
clarity.

Does MELTH change my Amazon prices or listings?+

No. The Amazon connection described here is read-only. It does not change prices, listings, inventory, shipments or advertising. Data synchronisation is separate from permission to execute business actions.

Is the data real-time?+

The connection is periodically updated. Freshness information helps you check the latest update for the relevant data. Coverage and available history depend on the source; a connection is not a real-time guarantee.

Can we use our existing spreadsheets?+

Yes. Excel (.xlsx) and CSV files can be added as business sources, inspected and analysed. You can select or change their intended uses. A purpose such as CRM or marketing is source metadata, not proof of an implemented import or connector for that process.

What do VVP, COGS and coverage mean?+

VVP is the managed cost per sellable unit in MELTH, with a relevant currency and effective date. COGS means the cost of the units sold. Coverage describes how much of the relevant dataset has usable cost evidence. Product coverage and revenue coverage measure different things.

Does contribution mean complete net profit?+

No. Contribution uses the available Amazon costs and applicable approved cost prices. Missing costs, incompatible currencies and incomplete coverage limit what can be calculated. Advertising is a future integration phase; full business profit is not claimed.

Does Ask MELTH analyse every row?+

Not for every question. Supported calculations can answer specific questions such as duplicate SKUs or currencies present. Open-ended analysis may use column structure and a labelled sample. Check the source version, coverage and limitations shown with the answer.

Can an approved value be corrected?+

An individual approved value can be corrected or withdrawn with its history retained. Rejecting a pending proposal and withdrawing an accepted value are different actions. Bulk withdrawal is not part of the scope described here.

What about other marketplaces, Buy Box and advertising?+

Amazon is the concrete application described on this page. Buy Box, expanded country analysis, Deals, Advertising and a full alerts module are next-stage work in the supplied product checkpoint. Other channels and external execution require their own verified integrations.

How do we get access?+

MELTH works with selected B2B clients. Submit your business context, challenge and motivation through the application form. Every application is reviewed personally; access is not automatic.

Bring your
e-commerce question.

Tell us about your business, the problem you want to solve and why now. We personally review each application and choose the clients we work with.

Discuss your e-commerce case ↗Application → personal review → invitation if there is a fit.