The Always‑On Outbound Engine

$2M in pipeline in four months, from one ranked list that scores every account and every contact, then decides who gets a rep and who gets an email.

Get the play
Video Poster

Outbound programs usually run on a list somebody built once. It ages from the day it ships, and reps work down it in whatever order feels right, so the accounts worth real effort get the same treatment as the ones that were never going to buy.

This is the system ZoomInfo built to replace that. The ICP definition refreshes continuously. Every account gets a fit score, every contact gets an engagement score, and the two multiply into one ranked list. That list splits into two paths: one where a fully briefed rep owns the account, and one where cold contacts get warmed by email until they show real intent. The first four months produced $2M in pipeline.

1. Define the ICP accounts

Every company that matches the profile of your best customers. Store this as a definition that re-runs rather than a list you export, so accounts growing into the ICP get picked up on their own.

Filters

  • Industry

  • Location

  • Company size

  • B2B

  • Always-on refresh

Why always-on - A static list decays from the day it's exported. Storing the ICP as a filter that re-runs keeps the top of the engine full without a quarterly rebuild.

2. Scrub the accounts that should never get outbound

Strip out everything where cold outreach would be wrong or embarrassing, so no customer, live deal, or rep-owned account ever receives it. Each of these is a check you run against an individual account, which is why they sit here instead of inside the ICP definition in step 1.

Exclusions

  • Existing customers

  • Open opportunities

  • Rep-owned accounts

  • No sales team

Order of operations. Scrubbing at the end means you've already spent scoring and research effort on accounts that were never eligible.

No sales team. A company can match on industry, location and size and still have no sales function, which leaves nobody to use what you are selling. Sales headcount is not one of the firmographic fields you filter on in step 1, so it has to be caught here.

3. Give each account a priority score

Score the surviving accounts against criteria that estimate whether they will open an opportunity in the next six months.

Scoring criteria

What belongs in the score. Each criterion should move the likelihood of an opportunity in the next two quarters up or down. Criteria that only describe the company add weight to the score without improving it.

4. Layer live signals on top

Factor recent activity into the score. This is what separates accounts that are warm this week from accounts that simply match the profile.

Live signals

  • Review site visits

  • New CRO hire

  • Recent tech install

Recency. A G2 visit from last week is worth acting on, where the same visit from last year is mostly noise. Let signals age out instead of accumulating.

5. Map the buying committee

Find every decision maker on the deal. These are not won through one person, so a single contact per account is a gamble.

Functions to cover

  • Sales

  • RevOps

  • Marketing

Function before names. Start from which teams have a stake in the decision, then find the named contacts inside each one. Working the other way around gets you whoever happened to be in the database.

6. Score each contact

Rate how likely each individual is to engage a rep, based on their own signals rather than their account's.Warmth values

  • Pricing page - 100

  • Other pages - 50

  • Past user - 50

  • New hire - 0

Why contact scoring exists. A high-fit account full of cold contacts still produces cold calls, and contact scoring is what stops reps spending their best accounts on the wrong person.

Past users. Somebody who used the product at a previous employer needs no explanation of what it does, which is why they score level with a warm page visit.

Why a new hire scores zero. These values measure warmth, meaning what the individual has already done that suggests they will reply, and someone three weeks into a job has no history with you yet. The hire itself is already counted at the account level, where leadership changes feed the fit score in step 3 and a new CRO hire is a live signal in step 4.

Zero is not a disqualifier. A contact with no warmth signals still moves through the system. They land on the cold path in step 9, get warmed by email, and graduate to a rep the moment they engage.

7. Scrub the contacts who should not be touched

The same discipline as step 2, applied to people. Nobody gets double-touched, re-sequenced, or emailed after opting out.

Exclusions

  • Already in a sequence

  • Inactive 2 months

  • Opted out

Deliverability. Mailing someone after they have opted out costs more than the extra volume is worth.

8. One score decides

Combine the two scores into a single number, and let it produce one ranked list.

Account fit × Contact score = Overall score

Why the scores multiply. Adding them lets a strong account carry a hopeless contact through to a rep. Multiplying requires both to be true.

Volume follows capacity. The ranked order decides who gets worked first, and rep capacity decides how far down the list that reaches.

9. Route the list, and let engagement move people between paths

The ranked list splits in two. The dividing question is whether a human picks up the phone.

SDR path

One rep owns the account and gets a full brief before making contact.

• Why this account • Who they are • Opening line

Cold path

Warms cold contacts by email until they show real intent.

• No calls • Email only • Pause after 5 ignored

Engagement graduates cold contacts to the SDR path

The two paths need very different first messages. The SDR path leads with the signal that surfaced the account, while the cold path makes no ask at all.

Every outcome feeds back into the scoring, so replies and closed deals keep sharpening the model. ZoomInfo's own build of this engine produced $2M in pipeline over a four month runtime.

FAQs for always-on outbound

Why score accounts and contacts separately?

Account fit tells you whether the company is worth pursuing. Contact score tells you whether this particular person is likely to reply. Multiplying them requires both to be true before a rep spends time on the account, where adding them would let one strong half cover for a weak one.

What does always-on actually mean here?

The ICP is stored as a definition that re-runs, so it is never exported and frozen. Companies that grow into the profile get picked up automatically and companies that fall out of it drop away, which means the top of the funnel maintains itself.

How do you decide who gets a human and who gets email?

The overall score sets the order and rep capacity sets the cut-off. The volume reaching the SDR path is whatever your team can work with a full brief behind each account, and everything below that line goes to the cold path.

What happens to cold path contacts who never engage?

They pause after five ignored emails. Contacts who do engage graduate to the SDR path, so a human only ever picks up the conversation once the person has shown real intent.

Why do new hires score zero when new hire signals drive so many other plays?

Because this score measures individual warmth rather than account opportunity. A new hire has no engagement history with you on the day they start, so they add nothing on that axis. The opportunity a new hire represents is captured at the account level instead, through leadership changes in the fit score and a new CRO hire as a live signal. Contacts scoring zero are routed to the cold path and graduate to a rep as soon as they engage.

Do I need my own CRM data to run this?

For the two scrub steps, yes. Removing customers, open opportunities and contacts already in a sequence all depend on your own records. Fit scoring and live signals run on third-party data.

← Back to the GTM Plays hub