Home / Glossary / ICP rubric
Glossary entry

ICP rubric

A weighted scoring model that turns "is this account a good fit?" from a vibes call into a number you can sort, filter, and defend. Mama uses one rubric per workspace to score every account in your CRM along four dimensions.

First written: March 12, 2026 · Last revised: May 4, 2026 · Category: scoring Reading time: 6 min
01

Plain English

TLDR
Your sales team already knows what a "good" account looks like — they just can't agree on the rules. An ICP rubric writes those rules down as weighted scoring dimensions, so every account in your CRM gets a number from 0–100 and you can stop arguing about which list to work first.

Most outbound teams have an Ideal Customer Profile (ICP) document. Almost none of them have a working rubric — a structured, weighted set of criteria that turns the ICP definition into an actual score.

The difference matters. An ICP doc tells you what a good account looks like. An ICP rubric tells you how good this specific account is, in a way you can compare to the 2,000 other accounts in your pipeline.

02

Why it matters

If you don't have a rubric, you have one of two problems:

  • The CRM-owner problem. Whoever loaded the CRM picks the working list, usually by gut. Reps work whatever's at the top. Half the booked meetings are out-of-ICP and waste your AE's time three weeks later.
  • The "everyone is a top account" problem. Marketing tiers everything as A/B/C. Eight months in, 60% of the database is tagged A. The tier means nothing.

A rubric fixes both by forcing every account through the same calculation. You don't need to agree on whether Acme Co is a good account — you just need to agree on the dimensions and weights. The math handles the rest.

It also gives you a defensible answer to the question every CRO eventually asks: "why is your team working this account?" The answer becomes a score and four lines of evidence, not "felt right".

03

The four dimensions

Mama's rubric uses four dimensions, weighted to sum to 100. You can rebalance the weights per workspace — these are the defaults we recommend after testing across our consultancy clients.

01
ICP fit
Weight · 35%
How closely the account matches your firmographic and technographic ICP — industry, headcount band, revenue band, tech stack, geo, business model.
Inputs: CRM enrichment data, public registry data (LinkedIn, Crunchbase, registry of companies), tech detection from /product/tech-lookup. Scored 0–100 based on how many ICP criteria are met and which ones (must-haves count 3× more than nice-to-haves).
02
Signal recency
Weight · 25%
When the most recent buying signal fired. A signal from yesterday is worth roughly 4× a signal from 90 days ago.
Inputs: the timestamp on each detected signal — funding event, exec move, hiring spike, tech change, voice cluster. Decay curve: half-life of 14 days. After 60 days, a signal contributes less than 10% of its peak weight unless it's been refreshed by a new signal of the same type.
03
Signal strength
Weight · 20%
How loud the signal is. A Series B with a press release is louder than 3 new LinkedIn job posts in the same week.
Inputs: structured event data — funding round size, number of net-new hires, seniority of exec move, the press footprint of a launch. Strength is per-signal-type: a "loud" hiring signal looks different from a "loud" funding signal. Mama's catalog of 47 signal types each have their own loudness curve.
04
Intent depth
Weight · 20%
How many independent signal types fire on the same account in the same window. Three different signals beat one big one almost every time.
Inputs: the count of distinct signal categories firing in the last 30 days for this account. Depth multiplier: 1 signal = 1.0×, 2 = 1.4×, 3 = 1.7×, 4+ = 2.0×. Hits a ceiling at 4 deliberately — we don't reward noise.
04

A worked example

Here's how the rubric applies to a real-looking account. The ICP for this hypothetical workspace: B2B SaaS, 100–500 employees, US headquarters, has a CRM + sequencer in production, sales-led GTM.

Northbeam AnalyticsAccount
Rubric score: 82 / 100
B2B analytics SaaS · 280 employees · NYC HQ · Series B closed Apr 2026 · Salesforce + Outreach detected
Dimension Evidence Score
ICP fit B2B SaaS ✓ · 280 emp ✓ · NYC US ✓ · Salesforce + Outreach detected ✓ · sales-led inferred from hiring mix ✓. All 5 must-haves cleared. 95 / 100
Signal recency Series B announced 11 days ago. Within high-weight window (half-life 14 days). 78 / 100
Signal strength $45M round · 4 press hits · CEO interview in Bloomberg. Loud by funding-signal standards. 82 / 100
Intent depth 3 distinct signal types firing: funding, hiring spike (12 net-new AEs in 30d), exec move (new VP RevOps). Depth multiplier 1.7×. 68 / 100
Weighted total (35/25/20/20) 82.0

What this score actually means in practice: Northbeam goes to the top of the working list this week. The brief Mama writes will lead with the Series B + VP RevOps move (highest-strength signals), and the email opener Mama drafts will anchor on the AE hiring spike (most specific, most personal to the buyer).

05

How Mama uses it

The rubric isn't a one-time scoring step. Mama re-runs it on every account in your CRM every 6 hours as new signals are detected, so the working list naturally re-sorts itself as the world changes.

Where you actually see the score

  • In your CRM — Mama writes the score back to a custom field on each Account record, plus a 4-line evidence summary. You can build views and reports off it like any other Salesforce/HubSpot field.
  • In the brief — every account brief shows the current rubric score at the top, with the per-dimension breakdown one click away. See /product/account-briefs for what a brief looks like.
  • In the working list — Mama generates a daily "score-jumped" list of accounts where the rubric score rose by 10+ points in the last 24 hours. That's where most booked meetings come from in practice.

Tuning your rubric

Defaults work for ~70% of B2B SaaS workspaces. Where defaults fail and you should rebalance:

  • If you sell to a very narrow ICP (e.g. seed-stage AI startups only): bump ICP fit to 50% and reduce signal weights — the floor on out-of-ICP accounts should be brutal.
  • If you sell to a long tail of midmarket (where most accounts technically qualify): bump signal strength + intent depth to combined 55% — fit becomes a filter, not a sorter.
  • If you sell into an industry that doesn't fund-raise much (manufacturing, services): drop the weight on funding signals inside strength, raise hiring + exec move weights.

Weights are editable in the workspace settings. We log every change so you can see whether last month's tuning actually moved booked meetings up or just rearranged the deck chairs.

06

Common mistakes

Three things we see teams do wrong with rubrics, repeatedly, including teams we worked with in our consultancy days.

  • Adding too many dimensions. Six dimensions feels more rigorous than four. It isn't. Each new dimension has to earn its weight in predictive power, and most don't. Stay at four unless you have data showing a fifth is doing real work.
  • Refusing to penalize stale signals. A funding round from 14 months ago is not a buying signal — it's history. If your rubric scores it the same as last week's round, your working list is full of accounts that already finished the buying process.
  • Letting the score replace the brief. The score tells you who to work. The brief tells you what to say. Teams that look at the score and skip the brief end up sending generic "saw your Series B" emails to every funded account, and reply rates collapse within a quarter.
Found something off? Glossary entries are written by humans and reviewed in batches. If a definition is wrong, incomplete, or missing the angle that would actually help you, email [email protected] and reference this entry — we'll revise and credit your input in the next revision.
Build your own

Want a rubric that scores your CRM in 90 seconds?

Connect a CRM, pick the dimensions you care about, and watch Mama score every account against your ICP. Free up to 100 accounts — no card required.