Signal-based selling: what it is, the signals we use, and how we route each one to a play

Signal-based selling is outreach timed by something that happened at an account. That can be a visit to your pricing page or a new head of sales. It can also be a funding round, or a role posted for the problem you solve. The signal decides when you write and what the first line says.

We run it for our clients as one layer of the GTM engineering system we build for them. The starting point is always the same. We look at the signals in the market that tell us why these leads are the best ones to go after. Then we build the route from each signal to a play.

This guide covers where signals come from and how we tier them. It then follows one signal from capture to a sent message, and shows where ABM and account scoring fit. Prospects asked us how signals work on 13 of our last 48 sales calls, so their questions are answered here too.

01What signal-based selling is

A signal is something an account does, or something that happens to it, that makes your offer more relevant this week than last week. On its own it is a piece of information. It turns into selling when three things are attached to it.

• The account it belongs to, checked against your ICP, so you know the signal comes from a company worth selling to.
• The reason it matters to your offer. A new CRO matters a lot to a sales consultancy and much less to a payroll provider.
• The action that fires next, with a named owner. Without a next step, a signal sits in a dashboard until it goes stale.

Signals are not a magic bullet, and the word itself has become a bit of a buzzword. The signal setup is always individual to the company. What matters is identifying which signals make sense for you, then wiring those few properly. For an HR agency the signal is new job postings and roles opening. For a construction firm it is something else entirely.

02The three types of signals

We group signals by where they come from, because the source decides how warm a signal is and how fast you need to act. Our list of buying signals that actually work goes signal by signal. Here we stay at the level of type.

First-party signals

First-party signals happen on channels you own. They are the cheapest to catch and the hardest for a competitor to copy. RB2B identifies website visitors at person level, so a visit becomes a named person at a named company. The page matters. A visit to a pricing or service page says more than a visit to a blog post.

Trigify watches LinkedIn engagement: who reacted to or commented on your posts, and on posts about the problem you solve. Our post engagers skill turns them into a weekly channel.

The third first-party source is your own CRM. Closed-lost deals, old opportunities and past contacts who come back to the site are signals too. They arrive with history, which makes the message easier to write.

Third-party signals

Third-party signals are events at the account that anyone can see if they look. The ones that come up most on our calls are hiring and job changes, leadership changes, funding rounds and changes to the tech stack. A leadership change might be a new head of sales or a new CRO.

An open role tells you a budget decision has already been made, which is why we treat hiring signals as strong ones. A tech install tells you whether a company runs the tool your pitch depends on. Our tech stack detection skill checks that one domain at a time. Funding is the loudest of the group. Every vendor sees the same announcement on the same morning, so it rarely carries a message alone.

These come from databases that capture each event, and Clay enriches them onto the account inside the outbound motion. Prospeo and Apollo then find the right people at the account, once the signal says it is worth working.

Intent signals

Intent covers what a buyer researches away from your site. Specific intent sits with the strongest signals there are, like an account comparing you with a competitor or reading reviews in your category. When it is a broad surge of interest in a topic, we use it to confirm another signal before anyone writes.

03Signal-based selling examples from our own system

The clearest example we can show is the system we built for KlientBoost, a performance marketing agency. They had high website traffic, a large database of closed-lost and past opportunities, and consistent LinkedIn engagement. Each of those ran on its own. Website visits were not worked systematically, and outbound had no signal-based prioritisation.

We designed and deployed a signal-based ABM engine around their Marketing Plan 2.0 offer, organised into five play groups.

The full system went from design to deployment in three months and runs 11+ automated ABM plays. HubSpot, Clay and the intent sources now feed one RevOps workflow. The KlientBoost case study walks through the whole build.

What we take from it applies to any signal-based selling setup. Each play group starts from one signal type and ends in one defined action. That pairing is what lets the plays run automatically.

04How we tier signals

We sort signals into four tiers by how close each one sits to a buying decision. Tier 1 is an account already looking, and it gets worked first. Tier 2 carries context worth writing around, like a relevant new hire. Tier 3 is accounts that fit your ICP with no event yet, which run as steady targeted outbound. Tier 4 is hypotheses we are still testing, sent in small batches.

The tier sets two things: how fast the system moves, and how much personal work goes into the message. The example signals in each tier, and how we treat them, are in the table in what a GTM engineer does day to day.

We also read signals together. One account read from several directions before anyone reaches out gives the first line more to work with than a single event does.

05Timing and routing: from signal to play

Every signal follows the same route, whichever type it is. Channel choice comes late, after the account is enriched and scored, because the right channel depends on what the data says about that account.

1. 01Catch it in one placeRB2B, Trigify, the CRM and the third-party sources all write to one record per account, stored in Supabase. Every new signal lands on an account that already has its history attached.
2. 02Enrich before deciding anythingClay adds company context, role and seniority, competitors and other market signals. Prospeo or Apollo finds the right people at the account and verifies their emails.
3. 03Check the CRMBefore anything goes out, the system checks whether the account is already in a conversation, is a customer, or was lost before. Each answer sends it down a different path.
4. 04Score fit and signal separatelyClaude scores the account against the ICP and scores the signal on its own. An account below the bar stops here. Our account scorer applies a locked model, so the same list always gets the same tiers.
5. 05Route to a play with an ownerThe tier picks the play. A tier 1 account reaches its owner in Slack and the CRM the same day, with the context attached. Lower tiers join sequences in Smartlead or HeyReach.
6. 06Write around the signalThe first line names the reason for writing this week. The follow-ups belong to the same play, so they read as one conversation instead of four more first emails.

Timing is where signal-based selling earns its value. A pricing-page visit is worth most on the day it happens, which is why a pricing-page visit reaches its owner the same day. Third-party signals have a longer window. They are public, though, so other vendors are reading them too.

The channel follows the signal. A tier 1 account with a person-level visit usually justifies LinkedIn and email together, while a tier 3 account gets email first. Layering signals across channels lifts response rates 18% to 22%. We cover how to sequence the two in LinkedIn outreach vs cold email.

06How ABM fits with signal-based selling

ABM and signal-based selling use the same route in a different order. ABM starts from the account: a short list chosen up front, the whole buying group mapped, then signals watched on those accounts only. Open-market signal work starts from the signal and asks whether the account fits afterwards.

We run both. The ABM short list comes from what your won deals had in common, and the signals decide when each account gets worked. That is how our account-based marketing service is built, with five to eight named people mapped per account.

Signals also sit inside a wider B2B outbound system. The data, the sending infrastructure, the copy and the reply handling all have to be in place before a signal can turn into a meeting.

07Account scoring and TAM come first

Signal-based selling only works on accounts worth selling to. A pricing-page visit from a company outside your ICP is noise, however hot it looks. So before any signal is wired, we size the market and score it.

That starts with knowing your TAM, from your own CRM as well as external data. We have sat with companies past $100M in revenue who could not tell us their own TAM. Duplicates, broken account hierarchies and missing scores were behind it every time. Then each account gets a fit score.

Fit and signal are scored separately for a reason. Fit tells you whether an account is worth working at all. The signal tells you whether now is the time. Appetite to buy shows up where the two meet.

08How to start signal-based selling

1. 01Pick the two signals your buyer emits. Look at your last won deals and write down what happened at those accounts before they bought.
2. 02Wire first-party signals first. Website visits and post engagement come from people already paying attention to you.
3. 03Give every signal an action and an owner. Decide what fires, on which channel, and who gets told.
4. 04Make every channel read the same signal. Email, LinkedIn and calls should all know why this account is being worked this week.
5. 05Review which signals earn replies. Count positive replies per signal, keep the ones that work and retire the rest.

When we set up signal-based selling for a client, we keep a library of signals inside their GTM brain. When one looks interesting, we implement it and it runs from there. The setup is custom every time, because every company sells to a different buyer.