The Signal-Based Outbound Playbook for B2B SaaS
Nine steps from an ICP you can act on to a weekly review that improves the system: signal sources, enrichment, scoring, sending infrastructure, persona messaging, multichannel sequencing, routing and CRM sync.Load more

Signal-based outbound means contacting an account because something changed that makes your product relevant now, and saying so in the message. It replaces the list-and-blast model with a system that sends fewer messages, to better accounts, at better moments. This playbook is the order we build it in. Each step depends on the one before, so resist the temptation to start at sequencing because it feels like progress. If the vocabulary is new, start with what GTM engineering is.
Step 1: Define the ICP you can act on
Most ICP documents are too vague to automate. "Mid-market tech companies with a growing sales team" cannot be turned into a filter. An ICP you can act on is written as attributes a data provider can return.
- Firmographics: headcount band, region, industry codes, business model (B2B, B2C, marketplace), funding stage.
- Technographics: tools that must be present (an integration you depend on) and tools that disqualify (a direct competitor under a long contract, or a platform you do not support).
- Team shape: the presence and size of the function you sell to. If you sell to RevOps, an account with no RevOps hires is not yet in market, whatever its size.
- Exclusions: current customers, open opportunities, partners, competitors, and any account sales has asked you not to touch.
Build it from evidence. Pull your last twenty to fifty closed-won deals and your last twenty closed-lost, and look at what the winners had in common that the losers did not. Then write the personas: for each, the job title patterns, what they are measured on, and what they would need to believe to take a meeting. This work belongs in the baseline and positioning stage of the Surge GTM Engine, and it is the step most often skipped.
Step 2: Choose signal sources
A signal is worth using if it meets three tests: it is observable from data you can get, it happened recently enough to be timely, and it correlates with your past wins. Start with a shortlist and prove each one before adding more.
Common signal families
- People changes: a new leader in the buying function, or a past champion who has moved to a new company. Champion moves are often the strongest signal a B2B SaaS company has, because the person already knows the product.
- Hiring: job posts that describe the problem you solve, mention a competitor's tool, or show the team you sell to is growing.
- Company events: funding rounds, acquisitions, new market launches, new regulatory obligations.
- Technology changes: adoption or removal of a tool that integrates with or competes with yours.
- First-party intent: visits to pricing, comparison or integration pages from identified companies, product sign-ups, webinar attendance, engagement with your founder's LinkedIn posts.
- Third-party intent: research activity on review sites or publisher networks. Useful as a tiebreaker; rarely strong enough on its own.
For each signal, write down the source, how often it refreshes, how you will dedupe it, and how long it stays fresh. A funding round is relevant for a few months; a pricing page visit is relevant for days. That freshness window feeds the scoring in Step 4.
Step 3: Build the enrichment waterfall
Signals usually give you a company. You need the right people at that company, with a verified email address and a LinkedIn profile, plus enough context to personalise. No single data provider has complete coverage, and coverage varies by region, company size and role. The answer is a waterfall.
- Company enrichment: resolve the domain, then fetch firmographics and technographics to confirm ICP fit before spending anything on contacts.
- Contact discovery: find people matching your persona title patterns. Cap it at the two or three most relevant per account to start.
- Email finding: query provider A; if it returns nothing or a low-confidence result, query provider B, then C. Stop at the first confident answer.
- Verification: run every address through a verification service. Drop invalids. Treat catch-all domains with care: send to them in small numbers and watch bounces.
- Context: pull the details your message will reference, such as the job post, the funding announcement or the recent post the person wrote.
Two rules keep this honest. First, record which provider supplied each field, so you can measure accuracy by source and drop the ones that underperform. Second, write enriched data back to the CRM only through defined fields with clear ownership, so enrichment never overwrites something a salesperson entered by hand.
Step 4: Score fit & timing
Keep two scores, not one. A fit score answers "does this look like our best customers?" and changes slowly. A timing score answers "has something happened recently?" and decays as signals age. Combining them into one number hides the reason an account is at the top of the list.
Start with simple, explainable rules: points for each ICP attribute matched, points for each fresh signal, a decay that halves a signal's weight once it passes its freshness window. Then sort accounts into actions:
- High fit, fresh signal: personal outreach from a named person, within a day.
- High fit, no signal: advertising audiences and light-touch nurture, until a signal appears.
- Medium fit, fresh signal: automated sequence with strong relevance to the signal.
- Low fit: excluded, however strong the signal.
Revisit the weights once you have a quarter of results. The question is simple: which combinations produced meetings and opportunities, and which produced nothing? Adjust the points to match.
Step 5: Set up email infrastructure & deliverability
Deliverability is not a setting; it is a reputation you earn and can lose. Outbound should never put your primary domain's reputation at risk, because that domain also carries invoices, support replies and product notifications.
Domains and mailboxes
- Register a small number of secondary domains that are clearly yours (variations on your brand name) and redirect them to your main site.
- Create a few real mailboxes per domain, under real people's names, with profile photos and signatures.
- Warm new mailboxes gradually over several weeks before sending outbound, and keep daily volume per mailbox modest even after warm-up.
Authentication
Every sending domain needs SPF, DKIM and DMARC configured correctly. These are not optional. Google's email sender guidelines require all senders to Gmail accounts to authenticate with SPF or DKIM, have valid forward and reverse DNS, and keep spam rates reported in Postmaster Tools low; bulk senders must also set up SPF, DKIM and DMARC, align the From domain with one of them, and support one-click unsubscribe for marketing messages. Yahoo publishes matching sender best practices. DMARC itself is defined in RFC 7489; start with a monitoring policy and tighten it once reports show your legitimate mail is aligned.
Hygiene that protects reputation
- Send only to verified addresses, and stop a campaign the moment bounces rise.
- Make opting out easy in every message, and honour it across every domain and tool at once.
- Keep tracking light. Open tracking pixels and heavy link tracking add risk and the open data is unreliable anyway.
- Monitor domain health weekly with Google Postmaster Tools and inbox placement tests, and rest any domain that starts to slip.
Check the law where you send. In the UK, the ICO's direct marketing guidance explains how PECR and UK GDPR apply to business-to-business email, including the different treatment of sole traders and partnerships. Take your own advice on the specifics for each market you sell into.
Step 6: Write messaging per persona
A message earns a reply when it shows you noticed something specific and connects it to a problem the reader owns. Write one core message per persona, then a short opening line per signal.
- Lead with the signal, briefly. "Saw you are hiring three SDRs in Berlin" is enough. Do not over-explain how you know.
- Name the problem the signal implies. Hiring SDRs implies ramp time and list quality. Say that, in their language.
- Offer something small. A relevant teardown, a short benchmark you have actually built, a question. Not a thirty-minute demo in the first email.
- Keep it short and plain. A few sentences, no images, one link at most, written the way a person writes to a peer.
Language models are useful for drafting first lines at scale from the context gathered in Step 3, but they need guardrails: a fixed structure, banned phrases, a requirement to cite the specific signal, and a human review of a sample from every batch. An automated line that misreads the signal is worse than no personalisation at all.
Step 7: Sequence across email & LinkedIn
A sequence is a small number of touches over two to three weeks, across channels, that read as one person paying attention. A typical shape:
- Day 1: view the person's LinkedIn profile and engage genuinely with a recent post if there is one worth engaging with.
- Day 2: first email, leading with the signal.
- Day 4: LinkedIn connection request with no pitch.
- Day 6: second email that adds something new (a resource, a different angle), not "bumping this".
- Day 10: if connected, a short LinkedIn message referencing the email thread.
- Day 14: a brief final email that makes it easy to say not now, and offers to point them to the right person.
Do LinkedIn steps by hand or with tools that respect the platform's rules. LinkedIn's User Agreement prohibits bots and unauthorised automated methods, and a restricted founder or salesperson account is a real cost. Running LinkedIn ads to the same account list while the sequence runs means the name is familiar by the time the second email lands.
Step 8: Route & sync to the CRM
The moment a prospect replies positively, the system has done its job and a person must take over. Define this in advance.
- Reply classification: positive, referral ("talk to my colleague"), not now, not interested, out of office, unsubscribe. Each has a different next action.
- Ownership: if the account already has an owner in the CRM, the reply goes to them. If not, assign by territory or segment, with round robin as the fallback.
- Speed: set a target time to first human response and measure it. A positive reply that waits two days often goes cold.
- Context: the owner should see which signal fired, which sequence ran and the full thread, in the CRM record, without opening another tool.
Every touch, reply and meeting writes back to the contact and account records with a consistent source field, so pipeline can be attributed to the outbound engine without manual tagging. If your CRM's lifecycle stages and lead sources are not trustworthy today, fix them first; our RevOps service starts there.
Step 9: Run a weekly review
The system improves only if someone looks at it on a fixed rhythm. Thirty to forty-five minutes a week, with the same agenda:
- Health: bounce rate, spam complaints, domain and mailbox status. Fix anything red before discussing anything else.
- Volume: signals fired, accounts that passed scoring, contacts enriched, sequences started. Look for stalls.
- Response: reply rate by signal, by persona and by message variant; positive reply rate separately.
- Conversion: meetings booked, meetings held, opportunities created, and time to first response.
- Decisions: one or two changes for the coming week, written down, so next week can say whether they worked.
Change one variable at a time where you can. If you rewrite the message, swap the signal and add a channel in the same week, you will not know which change mattered.
What to measure
Measure the system in layers, from leading indicators to the number the board cares about.
- Deliverability: bounce rate, spam complaint rate, inbox placement.
- Relevance: positive reply rate by signal and persona. This is the best early test of whether a signal is real.
- Conversion: meetings held per hundred accounts contacted, and opportunities created from those meetings.
- Pipeline: qualified pipeline value sourced or influenced by the engine, read from the CRM.
- Revenue: closed-won revenue and win rate compared with other sources, once enough cycles have closed.
Ignore open rates. Privacy features in major mail clients make them unreliable, and they tempt teams into optimising subject lines rather than relevance. The one number that matters is pipeline you can trace back to a signal, and the attribution behind it should be reproducible by your finance team from the CRM alone.
This is the system our GTM engineering service builds inside your own accounts and runs with your team each week. For tier-one accounts that justify more effort than a sequence, the ABM workflow guide and our ABM service take the same logic further. Definitions for every term used here are in the glossary.
Want a senior GTM engineer to review your current outbound setup, from ICP to deliverability? Book a call and we will walk through what we would build first.
About Surge45 Team
GTM engineers for B2B SaaS
Surge45 are GTM engineers for B2B SaaS. We build the GTM engine inside your own accounts, run it with your team, and measure what it returns in pipeline and revenue. 30+ B2B SaaS brands served, and 10+ years in-house before that.
Related Articles
Get More SaaS Marketing Insights
Join 10,000+ SaaS marketing leaders who receive our weekly newsletter with actionable strategies.


