Skip to main content

The Amazon data warehouse your team would have built.

Schema-versioned raw layer. Query it with whatever you already use. You keep the keys.

See integrations

daily_sales_and_ads.sql
-- Daily sales and ad spend per SKU, one marketplace.
-- Tables ship documented; you write the marts your team needs.

with sales as (
  select local_date_first::date as day, sku,
         sum(total_excl_tax) as revenue
  from amazon_orders_by_day_and_sku
  where marketplace_id = 'ATVPDKIKX0DER'
  group by 1, 2
),
ads as (
  select "date" as day, sku,
         sum("totalCost") as ad_spend
  from "amzagg_profit__adSkuDay"
  where "marketplaceId" = 'ATVPDKIKX0DER'
  group by 1, 2
)
select s.day, s.sku, s.revenue,
       coalesce(a.ad_spend, 0) as ad_spend
from sales s
left join ads a on a.day = s.day and a.sku = s.sku
order by s.day, s.sku;

The Amazon raw layer is a graveyard

Every brand we talk to has a half-built SP-API job. Token rotation works most of the time. The retry logic was added after the last incident. The schema breaks twice a year. Nobody owns it.

We own that layer. Your team owns the marts.

Engineering choices we made on purpose

Idempotent loads, not "eventually consistent"

Every report has a deterministic primary key. Re-running a window produces the same rows, never duplicates.

Raw archive in your S3

Every original report is stored as JSON in your bucket. You can re-derive any table, forever.

Profit aggregates and your own SQL

Use the supplied profit aggregates — daily summaries of ledger entries and advertising data — or define your own margin in SQL.

Agent-ready schema

Documented columns and sample queries — the metadata an LLM agent needs to query without hallucinating.

Skip building the SP-API pipeline