Built for grocery operators · est. 2026

All channels.
One system.
From raw feeds to decisions you can act on.

The data platform grocery operators wish they had. Every banner, every channel, every POS, on one governed read. Sub-minute refresh. Days, not quarters.

116

stores live

11

verticals

99.1%

schema health

ICInstacartDDDoorDashUEUber EatsSTShiptPOSPOSAMZAmazonORIGINOne read
HGG · Cárdenas Markets · El Rancho · Tony's · Mi Tienda

Designed for

Heritage Grocers Group and the regional banners that run alongside it.

  • Cárdenas Markets
  • El Rancho
  • Mi Tienda
  • Sun Fresh
  • La Tropicana
  • Tony’s Fresh Market

What ORIGIN does

Four pieces. One governed read of the grocery business.

01

Connect every channel

Instacart, DoorDash, Uber Eats, Shipt, your POS, the warehouse, Amazon Fresh. One read across all of them.

  • Marketplace APIs, POS schemas, ad platforms, web traffic
  • No flat-file uploads, no nightly batch lags
  • Schema drift handled at connect, not on a Friday afternoon
02

One governed read

Every banner exports differently. ORIGIN normalizes on connect so executive views compare like-for-like across stores, weeks, and channels.

  • Canonical SKU, banner, region, and category dimensions
  • Lineage on every metric, audit trail per row
  • No spreadsheet glue, no copy-pasted KPIs
03

The dashboard your team opens

Aggregate, per-banner, per-platform, operations, promo. Filters that stick. Sub-minute refresh. CSV export that matches the screen.

  • Channel × banner breakdown with anomaly cards
  • Funnel: orders, fulfilled, on-time, no-sub, no-reschedule
  • Weekly review that takes 8 minutes, not 2 hours
04

Ask in plain English

Origin Intelligence is wired to the same data layer the portal renders from. Ask "what was Cárdenas WoW" and it answers with the number, not a hallucination.

  • Tool-calling against the typed query API, never invented SQL
  • Cites the row, the period, the channel
  • No agent hand-waving, no "I don't have access"

What we actually fix

Grocery BI breaks in four predictable places.

Every operator we talked to said the same things. The numbers below are what they meant.

  1. 01

    Five marketplaces, five exports, five spreadsheets.

    Instacart, DoorDash, Uber Eats, Shipt, and the POS each ship a different schema. Reconciling them by hand on a Monday morning is the work.

    ORIGIN normalizes on connect. One canonical channel × banner × period table feeds every view.

  2. 02

    The dashboard your team actually trusts is 48 hours stale.

    Most grocery BI stacks ship overnight batches. By 8 AM Monday, the WoW number is already lying.

    Sub-minute refresh on every metric. Anomalies surface before the standup, not after.

  3. 03

    Two reports. Two numbers for the same week.

    No lineage means no answer to "where did $847K come from." Trust collapses.

    Every metric carries its row, its period, its source. Audit trail per number, exported with the CSV.

  4. 04

    The good answers live in someone's head.

    Ad-hoc questions wait for the analyst with the JOIN memorized. The director gives up and asks the operator.

    Origin Intelligence runs the question against the typed query API. Cárdenas WoW? It returns the number, the period, the source.

Master Portal

One operating picture. Every banner, channel, POS.

Built for grocery. Refreshed every minute. Queryable in plain English.

Demo: sign in, explore Overview → Analytics → Operations.

Operator deck · WTD 2026 W18

The dashboard your team actually opens at 8 AM Monday.

Every banner. Every channel. Every POS. Refreshed every minute. The figures below are sourced from the same pipeline that powers /portal - jittered for public viewing.

Open portal
Banner A WoW-11.8%
Banner B WoW-9.1%
Banner D WoW-5.4%
Uber Eats YoY+52.5%
Instacart / Banner D YoY+33.7%
Instacart / Banner A YoY+53.7%

Banner A

-10.2% YoY

$869.4k

9.3k orders$90.38 basket

Banner D

+24.7% YoY

$711.3k

5.4k orders$135.05 basket

Banner C

$413.9k

7.2k orders$63.53 basket

Banner B

-9.7% YoY

$144.1k

1.7k orders$77.54 basket

Source: jittered sample·5 banners ·13 channels ·updated every minute

Compare vs Instacart / Looker / Tableau →

How it ships

From raw feed to Monday morning, in three moves.

No data-engineering team required. No nine-month rebuild. Days, not quarters.

  1. 01 · Connect

    Plug in the channels.

    Instacart, DoorDash, Uber Eats, Shipt, the POS, the warehouse. Live APIs, not nightly CSVs.

    Typically: 1–2 hours per banner.

  2. 02 · Normalize

    One canonical read.

    ORIGIN handles schema drift, SKU dedupe, banner reconciliation, period alignment. You stop maintaining spreadsheet glue.

    Lineage and audit trail land with the first row.

  3. 03 · Open the portal

    The team picks it up Monday.

    Aggregate, per-banner, per-platform, operations, promo. Anomalies surface before the standup. Ask in English when the answer isn’t obvious.

    Sub-minute refresh on every metric.

Get started

Make every decision data-backed.

Connect every banner, channel, and POS. Ship clean, sub-minute analytics in days, not quarters.

Origin Intelligence