~/scoring/

Tier every account by fit and timing

Name: Cargo
SKU: cargo
Availability: InStock

Your coding agent builds the scoring on Cargo, with the rubric in your repo. Cargo runs it every hour and writes the result onto the account.

Start building

NNorthwindAccount

Employees140

IndustryB2B SaaS

Last roundSeries B, 12 days ago

Careers pageHiring a GTM engineer

AFit tier

Fits the ICP, hiring the persona

Timing 74 / 100

Trusted by GTM teams at

Augment Code
Weights & Biases
Dust
WorkOS
Braintrust
Descript
Alaan
Engine
Scaleway
Preply
DataDome
Oneflow

Two scores, two questions

Fit says whether an account is worth working. Timing says whether now is the moment.

Fit

An agent tiers each company A, B, C or disqualified against your ICP and rubric.

TAM companies

All records

FilterSort6 records4/4 columns

Name+3

Tier

Tier rationale

Tiered at

Add column

Northwind

Fits the ICP band, and its careers page lists a GTM engineer.

Nov 3, 2026

Contoso

Fits the firmographics; the persona is inferred, not seen.

Nov 3, 2026

Globex

Fits the ICP band, and runs the motion in public.

Nov 2, 2026

Initech

disqualified

Acquired: the sourced record is wrong about the company.

Nov 2, 2026

Fabrikam

Fits the firmographics and nothing else.

Nov 1, 2026

Umbrella

Fits the firmographics, with a growth pattern but no persona.

Nov 1, 2026

Timing

Recent signals add to a score, and each one counts for less as it ages.

Unify Accounts

All records

FilterSort5 records3/3 columns

Domain+2

Readiness score

Readiness stage

Add column

northwind.com

11.6

ready_now

contoso.com

8.4

ready_now

umbrella.com

9.1

ready_now

globex.com

3.2

cold

initech.com

2.5

cold

Every tier comes with its reason

Two sentences naming the rubric lines that decided it, and the page that proved it.

The judgment

The agent reads your rubric, searches once when it must, and decides.

Account tier analyst

TriggerEditorActivities

Deploy

Tier Northwind (northwind.com) against the ICP and the tiering rubric in the workspace context.

Reasoned for a few seconds

Read context: icp.md, tiering-rubric.md

Search the web: does Northwind employ a GTM engineer

Tier A.

Northwind is in the ICP's headcount band and industry, which meets the firmographic bar for every tier. Its careers page lists an open GTM engineer role, the persona the rubric's tier A bar asks to see employed.

Ask agent...

What gets written

The rubric line it cites, and what it hands back. The evaluator fails any tier with no grounded reason.

## Tiers
 
A   Work it now
    Fits the firmographics AND shows strong
    evidence of the motion the ICP sells into:
    the persona employed, or the practice
    visible in public.

{
  "tier": "A",
  "rationale": "Northwind is in the ICP's headcount
    band and industry, which meets the firmographic
    bar for every tier. Its careers page lists an
    open GTM engineer role, the persona the rubric's
    tier A bar asks to see employed.",
  "evidence_url": "northwind.com/careers"
}

When an account crosses the line

Scores land in segments, in Slack and in your CRM, where the team already works.

A ping, once per climb

An account that enters ready_now is posted once. If it cools and climbs back, it is posted again.

ready-accounts

CargoAPP8:20 AM

contoso.com is ready (readiness score 8.4)

Its recent signals crossed the ready line this hour. The events behind it are on the account's Account Events.

CargoAPP10:20 AM

northwind.com is ready (readiness score 11.6)

Its recent signals crossed the ready line this hour. The events behind it are on the account's Account Events.

CargoAPP2:20 PM

umbrella.com is ready (readiness score 9.1)

Its recent signals crossed the ready line this hour. The events behind it are on the account's Account Events.

The tier in your CRM

Tier, reason and scored-at update the matching HubSpot or Salesforce company. Nothing is created.

Tier accounts

TriggerEditorRecordsMetrics

Hourly

TAM companiesUntiered or stalestart

Account tier analystAgenttier_analyst

Write tierTier, reason, tiered atwrite_tier

Update companyHubSpotupdate_hubspot

Update accountSalesforceupdate_salesforce

HubSpotSalesforce

~/scoring-as-code/

Change the rubric with a pull request

What tier A means is a markdown file in your context repo, beside the ICP. The agent reads it on every run, so a merged edit changes the next tier with no deploy.

Reviewed like code. A rep can read the same file the agent read, and git keeps the history of why a tier changed.
Timing in TypeScript. Weights, half-lives and thresholds live in your project. Edit them, run the plan, and read what moved.
Your coding agent does the edit. It works on the same files, in the repo you own, and Cargo runs what you deploy.

$ git diff context/tiering-rubric.md
 
  ## Ties
 
  Fit, but the disqualifier is uncertain:
    tier C, and say what would disqualify it.
 
  Strong evidence, headcount one band off:
-   tier B.
+   tier A when the persona is observed,
+   tier B when it is only inferred.

Cookbooks built on scoring

Accounts tiered by fit, ranked by timing, ready to adapt.

Account scoring

Score every account against your ICP on a schedule.

Inbound qualification

Qualify and route inbound demo requests from your website form.

Readiness scoring

Keep a sales readiness score on every account.

See all cookbooks

Frequently Asked Questions

Can't find the answer here? Contact our team.

Fit asks whether an account is worth working at all: an agent tiers each company A, B, C or disqualified against your ICP and tiering rubric. Timing asks whether now is the moment: a readiness score adds up the account's recent signals, each weighted by type and halved every half-life, and sorts the account into ready_now, warming or cold.

In your context repo, as a markdown file beside the ICP. The agent reads both before every judgment, and nothing in its prompt overrides them. Edit the rubric and the next run tiers differently, with no deploy and a git history of why. The timing weights, half-lives and thresholds live in your project's TypeScript.

Yes. Every tier is written with a rationale of at most two sentences naming the rubric lines that decided it, the URL of the page that decided it, and a timestamp. The agent's evaluator fails a tier that has no grounded rationale.

Anything that lands as an account event: Cargo Signals feeds such as funding rounds and job posts, and your own models such as demo requests, signups or CRM meetings. You set a weight and a half-life per signal type. A signal with no weight counts for nothing, so a new source never moves scores until someone has decided what it is worth.

The tier can. Name HubSpot or Salesforce in the settings and each run updates the matching company with tier, reason and scored-at, in three properties you create. It updates and never creates, and when a domain matches more than one company it writes to none of them. When an account becomes ready, a play posts it to a Slack channel once.

The tiering play runs hourly and only judges accounts that are new or older than your re-score window, six months by default. That is one agent run per company, so pilot ten rows before the backfill. The readiness score uses no model at all: its columns are recomputed hourly in your workspace's warehouse, and Cargo Signals bills per event delivered.

Build the self-driving GTM engine

Your coding agents build it on Cargo.Or our team builds, operates and improves it with you.

Start building

//////// /////// //////// //////// /////// /// /// /// /// /// /// /// /// /// /// /// /// /// /// /// /// /// ///////// //////// /// //// /// /// /// /// /// /// /// /// /// /// /// /// /// /// /// /// /// /// /// /// //////// /// /// /// /// /////// ///////

{

"tier": "A",

"rationale": "Northwind is in the ICP's headcount

band and industry, which meets the firmographic

bar for every tier. Its careers page lists an

open GTM engineer role, the persona the rubric's

tier A bar asks to see employed.",

"evidence_url": "northwind.com/careers"

}

$ git diff context/tiering-rubric.md

## Ties

Fit, but the disqualifier is uncertain:

tier C, and say what would disqualify it.

Strong evidence, headcount one band off:

- tier B.

+ tier A when the persona is observed,

+ tier B when it is only inferred.