Guide

Fit and intent scoring in 2026: the practical taxonomy, and exactly how to turn it into booked meetings

May 14, 2026Updated June 24, 202614 min read2,734 words

Fit and intent scoring works on two axes: fit asks whether an account matches your ICP, intent asks whether they are buying now. Score them separately, add decay windows and stop rules, then act daily on the high-fit, high-intent accounts.

Most teams miss meetings because they score the wrong thing. They chase “hot” accounts with terrible fit, or they obsess over fit and ignore the buyer who is shopping today.

Fit and intent scoring fixes that. Not as a dashboard you check on Mondays. As the rule set that decides what outreach happens next, every day.

The gap is rarely the model. Plenty of teams can build a decent scoring sheet. What they never build is the loop that turns a score into a sent email, a paused sequence, or a human escalation. This guide gives you the taxonomy and the loop, and shows where an autonomous operator runs the loop for you so the scores never just sit there.

Fit and intent scoring in 2026: the only definition that matters

Fit and intent scoring is a dual scoring system that predicts two different truths:

Fit score: how closely an account matches your ICP.
Intent score: how likely that account is to engage or buy in a near time window.

If your pipeline depends on outbound, you need both. Otherwise you do what most teams do: send 10,000 emails to prove nobody wants your product.

Fit vs intent scoring, and why a single score lies

Fit scoring: “should we sell to them?”

Fit stays mostly stable. It moves when:

the company grows or shrinks
the stack changes
the team structure changes
your ICP shifts

Fit comes from:

firmographics: industry, size, geo, revenue, growth stage
technographics: tools in use, cloud, data warehouse, CRM, marketing automation
structural reality: business model, sales motion, compliance needs

Intent scoring: “should we sell to them now?”

Intent is perishable. It spikes, then dies.

Intent comes from:

hiring
funding
leadership changes
web activity
review-site research
content consumption
email engagement
negative signals that say “stop”

Why a single blended score lies

A single “lead score” hides the reason a lead looks good. Two classic failures show why that matters:

High intent, low fit: students, consultants, tiny teams, competitors, job seekers. Lots of clicks, zero deals.
High fit, low intent: perfect accounts, bad timing. You burn them with spammy follow-ups and poison the well.

A blended 70 could be either case, and the right move is opposite in each: aggressive on the first looks reckless, patient on the second looks lazy. Two-axis scoring keeps the reason visible so the action is obvious.

The practical taxonomy: fit signals

Firmographics (fit)

Best for fast filtering and clean ICP boundaries. Score:

Industry or vertical (you sell to B2B SaaS, not restaurants)
Company size (use employees as a proxy if revenue is messy)
Geography (time zones, language, regulatory reality)
Business model (PLG vs sales-led vs services-heavy)
Sales motion (a company with no outbound motion is a “future maybe,” not a “now”)

Scoring rule: big points for must-have traits, zero for nice-to-have, negative for never.

Technographics (fit)

Technographics are fit signals with teeth because they map to integrations, switching costs, nameable pain, and budget patterns. Score:

CRM (HubSpot, Salesforce, Pipedrive)
Sales engagement (Apollo, Instantly, Outreach)
Data tools (Clay, ZoomInfo, enrichment vendors)
Email infrastructure (Google Workspace vs Microsoft 365, sending tools)
Analytics or warehouse (if your product depends on it)

Use them honestly. If you integrate with Salesforce, Salesforce presence is fit-positive. If you replace it, Salesforce presence is closer to an intent signal, but only when you also see a switching trigger: a new RevOps leader, a cost-cutting push, a CRM migration.

The practical taxonomy: intent signals

Intent is where teams get religious and weird. Don’t. Treat every intent signal as a probability with a half-life.

Hiring (intent)

Hiring is a budget leak with a paper trail. Score net-new roles in the function you sell into (SDRs, RevOps, demand gen), a first hire that signals a new pod, and “building outbound” language in job posts. A team hiring SDRs while reply rates fall will need automation or it will churn people, fast. Decay window: 30 to 60 days.

Funding (intent)

Funding changes risk tolerance and priorities. Score the stage: seed to Series A is building the first repeatable pipeline, Series B and C is process and tooling consolidation, growth equity is efficiency and cost control. Funding is not “buying,” it is “can buy,” which is still useful. Decay window: 45 to 90 days.

Leadership changes (intent)

New leaders make changes because they have to justify the hire. Score a new CRO, VP Sales, or Head of RevOps, a new Head of Growth, a new CMO at sales-led orgs. Tie the change to a reason: “new RevOps leader plus a messy stack” beats “congrats on the new job.” Decay window: 60 to 120 days.

Web activity (intent)

First-party web intent is real because it is your property. Score pricing-page visits, integration-page visits, multiple sessions in a short window, visits from target geos during business hours, and repeat visits to compare pages and case studies. Web intent without identity is just vibes; you need enrichment and matching, or it is a ghost story.

Review-site intent (intent)

Review sites are late-stage behavior. Buyers go there to shortlist, not to learn what a category is. G2’s Buyer Intent signals cover explicit evaluation actions like viewing product profiles, pricing pages, alternatives, and comparison pages, not inferred topics (documentation.g2.com). Score compare-page views, pricing-page views, alternatives-page views in your category, and competitive research heavily. Decay window: 7 to 21 days, because this is the most perishable signal you have.

Content consumption (intent)

Content intent is tricky because half your “traffic” is AI scrapers, students, internal employees, competitors, and agencies doing research. Score bottom-of-funnel assets (implementation guides, migration docs, pricing PDFs, ROI calculators), webinar attendance rather than registration, and repeat consumption across multiple days. Decay window: 14 to 30 days.

Email engagement (intent, handle with gloves)

Open rates got cooked years ago by privacy features, so treat opens as directional at best. Apollo’s benchmarking guidance notes that privacy proxies inflate opens and recommends prioritizing inbox placement, reply rate, and positive reply rate (apollo.io). Score human replies, positive replies, and forwarded or introduced signals highest; treat link clicks and opens as light signals only.

Score deliverability health too. Validity highlights spam complaint rates as a direct driver of sender reputation, naming the under-0.1% band as best practice while bulk-sender requirements reference a 0.3% threshold (validity.com PDF). Decay windows: replies 30 to 90 days, clicks 7 to 14 days, opens 3 to 7 days and only if your tracking is clean.

Negative signals (stop rules, not intent)

Negative signals are not intent. They mean “stop bothering people.” Score these as hard brakes: hard bounce, spam complaint, unsubscribe, domain block, repeated “not interested,” “remove me” replies, and role mismatch (student, recruiter, vendor, competitor). Validity calls spam complaint rates a key driver of reputation degradation, so treat them as an existential risk, not a metric (validity.com PDF). No decay window: these are terminal.

The scoring model: simple, weighted, decayed, operational

You want a model your team can run weekly without a PhD.

Step 1: build two 0-100 scores

Fit score (0-100): stable traits
Intent score (0-100): time-bound signals

Do not blend them.

Example fit weights (0-100)

For a B2B SaaS selling outbound automation:

Firmographics (60 points)

Industry match (0-20)
Employee-size band match (0-15)
Geo match (0-10)
Sales-motion match (0-15)

Technographics (40 points)

Uses a CRM you support (0-10)
Uses outbound tools you complement or replace (0-10)
Uses data tooling that indicates outbound maturity (0-10)
Stack complexity, meaning too many tools (0-10)

Fit stop rules

“Never” industry: fit = 0, stop.
Employee count below minimum viable: fit capped at 30.

Example intent weights (0-100)

High-intent (60 points)

Review-site compare or pricing activity (0-35)
Pricing or integrations page visits (0-15)
Reply with interest (0-10)

Mid-intent (30 points)

Hiring in function (0-15)
Leadership change in function (0-10)
Funding event (0-5)

Light (10 points)

Content consumption (0-5)
Email click (0-5)

Intent stop rules

Spam complaint: intent = 0, suppress contact and domain
Unsubscribe: suppress contact
Hard bounce: suppress address, re-enrich if the account is strategic

Add decay windows so old “intent” stops pretending

Intent signals die, and your model should say so. Use stepped decay; it is easier to run in ops than a continuous curve.

Suggested windows:

Review-site intent: full value 0-7 days, half 8-21, zero after 21
Web pricing or integrations visits: full 0-7, half 8-14, zero after 30
Hiring: full 0-30, half 31-60, zero after 90
Leadership change: full 0-60, half 61-120, zero after 180
Funding: full 0-45, half 46-90, zero after 180
Email click: full 0-7, half 8-14, zero after 21

This keeps your hot list honest.

Thresholds and the decision matrix

Your score only matters if it triggers an action.

Fit bands

80-100: core ICP
60-79: adjacent ICP
40-59: long tail, test only
under 40: suppress

Intent bands

80-100: active buying window
60-79: warming
40-59: watchlist
under 40: low priority

The decision matrix

Fit 80+, intent 60+: sequence aggressively now.
Fit 80+, intent under 60: light sequence, then pause and monitor.
Fit 60-79, intent 80+: qualify fast, do not over-automate.
Fit under 60: only engage if intent is extreme and the deal size justifies it.

Stop rules: protect deliverability and your brand

You do not get bonus points for persistence. You get blocks.

Hard stop rules, non-negotiable:

Spam complaint: suppress contact, consider suppressing the domain
Unsubscribe: suppress contact permanently
Hard bounce: suppress address permanently
Confirmed role mismatch: suppress contact, route the account to a different persona if relevant
Three negative replies across one domain in 14 days: pause the domain, review targeting

Validity’s report names spam complaint rates as a top driver of sender reputation damage (validity.com PDF). Treat stop rules as pipeline insurance.

The operational loop: score, prioritize, sequence, pause, escalate, book

This is the part most teams never build. They score, they stare, they do nothing.

1) Score (daily refresh)

Refresh fit weekly or monthly, refresh intent daily, apply decay daily, apply stop rules in real time. If your scoring does not refresh daily, it is not intent scoring. It is historical trivia.

2) Prioritize (today’s call sheet)

Three queues:

Now: fit 80+ and intent 60+
Watch: fit 80+ and intent 40-59
Trash: fit under 60 or a stop rule triggered

3) Sequence (match message to signal)

Triggered outreach beats list blasts because relevance goes up. Tie the opener to the signal category, and keep it account-level rather than claiming you watched a person:

review-site intent: “saw your category is in a comparison cycle”
hiring: “you’re hiring SDRs, outbound volume is about to jump”
leadership change: “new RevOps leader, stack-audit season”
web pricing: “pricing question, here’s the two-line answer”

For the compliance and deliverability discipline that keeps these sequences landing, start with Chronic’s cold email compliance SOP and cold email deliverability in 2026.

4) Pause (when intent drops or negatives appear)

Pause when intent falls below 40 after decay, when there is no engagement after N steps, when a domain shows multiple negative replies, or when deliverability metrics degrade. Pausing is not giving up. It protects future inbox placement.

5) Escalate to a human

Escalate when fit 90+ meets intent 80+, when a review-site compare signal hits twice in 7 days, or when a real reply asks a real question about security, pricing, or timeline. The human tasks are the high-touch ones: two-minute account research, a custom teardown, a short Loom, a call plus voicemail when the persona matches.

6) Book (and feed the loop)

When a meeting books, record which signals preceded it, raise the weights on signals that correlate with meetings, and lower the weights on signals that correlate with noise. That is how scoring compounds instead of drifting.

Where an autonomous operator runs this loop for you

Most tools own one slice of the loop. Clay builds data flows, powerful and complex. Instantly sends email. Salesforce stores objects and charges per seat for the privilege. You are left wiring them together and being the loop yourself, which is exactly the work that never gets done.

Chronic is an autonomous revenue operator: you set the goal, and the agent runs the loop end to end until the meeting is booked. It defines the ICP, enriches accounts, scores them on both axes, applies decay and stop rules daily, generates and sends signal-matched email from managed, warmed mailboxes, handles replies, and surfaces approvals only for the decisions that need a human. The score does not land in a dashboard you forget to open; it becomes the next action automatically.

For the broader picture, see the 2026 outbound stack that produces booked meetings.

Implementation: set this up in 7 steps

Write your ICP as rules, not adjectives. Not “mid-market tech.” Instead: 50-500 employees, B2B SaaS, North America and UK, an outbound motion exists, uses HubSpot or Salesforce or is actively switching.
Pick 10-15 fit fields and 10-20 intent events. Pick 80 and you will never ship.
Define point values and stop rules first. Weights are easy to tune later. Stop rules prevent damage now.
Define decay windows per intent type. Start with the windows above and adjust based on meetings booked.
Create the three queues. Now, Watch, Trash.
Build 3-5 signal-specific sequences. One per major trigger: review-site evaluation, hiring, leadership change, web pricing visit, and generic ICP fit with light intent.
Run weekly weight reviews tied to booked meetings. Not clicks. Not opens. Meetings.

Common mistakes that kill fit and intent scoring

Blending fit and intent into one number. You lose the reason, so you lose the action.
Scoring only positive intent. Negative signals matter more; they prevent domain death.
No decay. You chase last month’s hot lead who already bought.
No stop rules. You keep sending into bounces and complaints, and deliverability collapses.
No operational loop. You built analytics, not pipeline.

FAQ

What is fit and intent scoring?

Fit and intent scoring is a dual scoring model that ranks prospects on two axes: fit (ICP match) and intent (time-sensitive buying signals). Fit answers “should we sell to them?” Intent answers “should we sell now?”

Which matters more, fit or intent?

Fit matters more for long-term efficiency, intent matters more for short-term meetings, and the highest conversion comes from high fit plus high intent. If you must choose, prioritize fit for deliverability and brand safety, then use intent to time the push.

What intent signals are strongest in 2026?

Late-stage evaluation signals: review-site comparisons and pricing-page behavior. G2’s Buyer Intent includes profile, pricing, alternatives, and compare-page views (documentation.g2.com).

Should I use email open rates as an intent signal?

Only lightly. Open rates are distorted by privacy proxies and bot opens, so treat them as directional inside your own program, not a primary indicator, and prioritize replies and positive replies (apollo.io).

What are the non-negotiable stop rules for outbound?

At minimum: spam complaint, unsubscribe, hard bounce. Complaints damage sender reputation and trigger deliverability problems, which Validity’s benchmarking emphasizes (validity.com PDF).

How often should we refresh intent scores?

Daily. Intent decays fast, and review-site and pricing signals can be useless after two or three weeks. A weekly model shows up after the buyer already chose someone else.

Build the loop, then let it run

Stop arguing about scoring theory. Ship a taxonomy, add weights, add decay, add stop rules, then run the loop every day: score, prioritize, sequence, pause, escalate, book.

The model is the easy half. The hard half is running it daily without dropping a signal or burning a domain. That is the work an autonomous operator does on its own. Set the goal and let Chronic run the loop end to end, until the meeting is booked.