B2B outbound: how a modern outbound system works, stage by stage

B2B outbound is reaching companies that have not come to you yet, by cold email, LinkedIn and sometimes the phone, to start a sales conversation. A modern B2B outbound system runs it as one loop. A scored list feeds warmed sending infrastructure, a message built on a signal goes out on the right channel, and every reply is written back to the CRM.

We build and run these systems for clients every week, and the order of the parts matters more than any single tool. This guide walks through each stage in the order we build it. It also answers what buyers ask us most on sales calls: how it works phase by phase, how many leads or emails a month it takes, and how long results take.

01How a B2B outbound system works, end to end

There are seven stages, and each one feeds the next. Skip one and every stage after it inherits the problem, so the table shows what breaks as well as what happens.

Two rules hold the order together. The list is scored before any channel is picked, because who you are writing to decides how to reach them. The infrastructure is built before any campaign, because warm-up takes weeks you cannot buy back later.

02The three levels of outbound execution

How much of that loop a person runs by hand varies a lot. When we mapped our own process, it fell into three levels. The tools change at each level, and so does the person's job.

The person stays in the loop at every level. Their role moves up, from operator to orchestrator to advisor. At level three the system can build a thousand messages, and someone still signs off before any of them goes out. Every stage below applies at any level. What differs is how much of it is wired together.

03Start with the ICP and a scored list

The list decides more of the result than anything else in B2B outbound. Start from the customers who bought, closed quickly and stayed. Their shared traits become the ideal customer profile you source against: industry, size, region, the tools they run and the roles that signed.

Source accounts first, then people. We default to Prospeo, which sources contacts and verifies emails in one pass, with Apollo alongside it for company search. Clay adds the enrichment and signals for the outbound motion. Claude scores each account against the ICP and writes the reasoning next to the score. The full set, stage by stage, is on our tools page, and the scoring step runs on our account scorer skill.

Inside each account, look for a small set of people: the person who signs, the person who feels the problem, and sometimes the one who uses the product. Then tier the scored list. Top-tier accounts justify more channels and more research per contact, and the long tail gets a lighter touch. The channel mix comes out of this step.

04Sending infrastructure and deliverability

We send every cold email from secondary domains. A secondary domain is a close variant of your brand's domain, bought only for outbound. A reputation problem there never touches the inbox your clients write to. Each domain carries a few inboxes, and every sending domain passes SPF, DKIM and DMARC.

Each inbox warms up before it sends, and warm-up stays on for its whole life. Each address is verified before it enters a sequence, so bounces never reach your domains. Rotation then rests inboxes before they wear out. Provider limits, warm-up presets and the weekly health check are all in our cold email deliverability checklist.

How many emails or leads a month

Volume is a calculation with two halves: what your infrastructure can send, and what each account needs. Work out both before you buy a single domain.

Capacity per month = inboxes x sends per inbox per day x sending days
Need per account    = contacts per account x emails per contact
Accounts per month  = capacity per month / need per account

Example: 3 contacts x 4 emails = 12 emails per account
23,100 emails a month / 12 = 1,925 new accounts a month

The 23,100 is a real reference point. The StryvRevenue build ran 30 inboxes across 10 domains to support 23,100 emails a month. On LinkedIn the same build used four profiles, with a capacity of 2,000 messages a month.

The right number of leads is the number your market can supply without padding the list. A small market with large deals needs fewer accounts and more research per contact. Past the accounts that fit, extra volume only adds poor fits, and it costs you sender reputation to send them.

05Channels: email, LinkedIn and the phone

The channel decision follows the tiers. Email carries the volume. LinkedIn suits the accounts where a few touches from a real person's profile are worth more than reach. The full comparison, with the sequence for running both against one list, is in our breakdown of LinkedIn outreach vs cold email.

The phone is a third layer. It catches contacts the other channels miss and adds weight on the top tier. In the StryvRevenue system, Smartlead handled the email sends, and contacts without a valid email went to Kixie for calling instead of being dropped. The top tier also received handwritten letters alongside LinkedIn and email.

06The message: a reason to write today

Good outbound copy tells the reader why you are writing to them now. The strongest reason is a signal, something true about their company this week, followed by what you do about the problem it points to. A small ask closes the email. Follow-ups stay short and add something new each time, instead of repeating the first ask louder.

Personalisation comes from the enrichment, so it scales with the list. In the BIGVU build, every prospect was a real estate agent or broker with live Zillow listings. Clay layered in their YouTube, LinkedIn, Instagram and podcast presence. Claygent then wrote a message for each contact from that footprint, so every email referenced the prospect's own listings.

Subject lines and the layers of cold email personalisation each deserve a breakdown of their own. So do the pattern-breaking plays, from physical mail to screenshot-led openers, that sit on top of this base once it works. The cold email playbook in the guides below covers the copy rules in the meantime.

07Signals and timing

A signal is an event that makes a problem urgent: a hiring push, a funding round, a new leader, a technology change, a visit to your pricing page. Signals decide who enters the sequence this week, and they hand the message its opening line. Building the whole motion around them is signal-based selling.

The obvious signals are the ones every vendor pulls. A funding announcement reaches everyone with the same alert on the same morning, and the inbox fills with the same congratulations. The signals that work best are specific to one vertical. For a staffing firm, a burst of sales roles posted this week says more than a funding round. We map each client's triggers from scratch, and our guide to buying signals covers which ones to track and where they come from.

08Replies and CRM write-back

Replies get classified before anyone reads them. Each one gets one of nine tags: super interested, interested, question, not now, not interested, out-of-office, wrong person, unsubscribe or spam. Positives go straight to a person, with the thread and the account context attached. Not-nows get a dated follow-up task, and unsubscribes are suppressed on every channel. Our reply triage skill does that first pass.

Every outcome is then written to the CRM against one record and one owner, and a reply on any channel stops the other channels for that person. In the StryvRevenue system, replies from every channel landed in one Slack channel for the SDRs. HubSpot records were created and updated automatically, including live updates from the Kixie dialer.

Any CRM with an API can take the same write-back. Build the structure before the first send, using our CRM setup checklist, because records a live campaign has already written to are harder to fix than an empty structure.

09Measuring outbound sales: positives per touch and cost per meeting

Emails sent and open rates tell you how busy a B2B outbound system is. Two other numbers tell you whether it works. Positives per touch is interested replies divided by messages sent, with out-of-office and auto-replies removed first. Cost per meeting is total system spend, covering tools, data and people, divided by the qualified meetings booked.

Track both by segment and by channel, never as one blended figure. A segment that books meetings cheaply earns more volume. A segment with no positives after a fair test gets paused or rewritten. Cost per meeting is also how we judge a GTM engineer, because it measures the system they built instead of their activity. It is the fairest way to compare a system against hiring another SDR, using your own salary and tool costs.

10The phases, and how long until it works

Every B2B outbound system we build moves through the same phases. Weeks 1 to 2 are extraction: sessions with your best sellers, your won and lost deals, and an agreed target definition. Weeks 3 to 6 are build and launch, where the infrastructure is set up and warmed and the first campaigns go live. Weeks 7 to 12 are run and tune, with weekly changes to targeting, copy and sequencing.

The early weeks are build time because warm-up has to run before volume does. The run-and-tune stretch is where the message gets proven at controlled volume, and where you find out which segments pay. Scaling comes after that, onto whatever has earned it. A good list and clean infrastructure from day one are what shorten the wait.

For StryvRevenue, the system shortened time to revenue generation from about a year to a couple of months. The two engines it ran produced these results:

The week-by-week shape of an engagement, including what we need from your team, is on our process page. Our outbound service is this system, built and run for you.