The Clay Outbound Playbook: What 50+ Deployments Taught Us

Table of Contents

• Why Manual Outbound Breaks at Scale

• The Playbook: Nine Steps from Raw List to Qualified Pipeline

• Module 1: Source Your Target Accounts

• Module 2: Enrich with AI Research

• Module 3: Score and Qualify with ICP Tiering

• Module 4: Find the Right Contacts

• Module 5: Multi-Provider Email and Phone Waterfall

• Module 6: AI-Personalized Sequences

• Module 7: Push to Sending Tools

• Module 8: Reply Monitoring and Classification

• Module 9: Measure What Matters

• The Stack and What It Costs

• Who This System Works For

• Frequently Asked Questions

Why Manual Outbound Breaks at Scale

Most outbound teams operate the same way. An SDR opens LinkedIn, manually researches a company, copies details into a spreadsheet, writes a semi-personalized email, and sends it through their sequencer. Multiply that by 50 prospects a day. That is 60% of their time spent on research and data entry instead of selling.

The results reflect it. Generic cold emails land at 0.5 to 1% reply rates. Domain reputation suffers. Pipeline stays flat even as headcount grows.

The core issue is not effort. It is architecture. When prospecting runs on manual research and template-based outreach, every additional rep adds cost but not proportional output. The system does not scale because there is no system. There are just people doing repetitive work.

We have deployed Clay-based outbound systems for over 50 companies. Every deployment follows the same 9-step workflow. The specifics change per client (different ICP, different messaging, different tools), but the architecture stays consistent.

This post walks through that architecture module by module.

The Playbook: Nine Steps from Raw List to Qualified Pipeline

The system runs as a linear pipeline where each step feeds the next:

1. Source target accounts

1. Enrich with AI-powered research

1. Score and qualify against your ICP

1. Find the right contacts at qualified accounts

1. Verify emails and phones with multi-provider waterfalls

1. Generate AI-personalized sequences per contact

1. Push to sending tools

1. Monitor replies and classify intent

1. Measure qualified outcomes

The entire pipeline runs inside Clay, with two external tools for delivery: Smartlead for email and HeyReach for LinkedIn.

Research, scoring, sequence writing, and campaign launch all happen in one place. That is the difference between a tool and a system.

Module 1: Source Your Target Accounts

Every campaign starts with a list of companies to pursue. There are four ways to build that list inside Clay:

CRM Import. Pull existing accounts from HubSpot or Salesforce. Best for re-engaging stalled deals or running outbound against accounts your team already knows. Filter by deal stage, last activity date, or lifecycle stage before importing.

CSV Upload. Import a spreadsheet from any external source. Conference attendee lists, investor portfolio pages, industry directories. Clean the data before uploading: remove duplicates, standardize company names, verify domains.

Lookalike Audiences. Start with your best customers and let Clay find similar companies based on industry, headcount, tech stack, and funding stage. This is the fastest path to a qualified list when you know your ICP but need fresh accounts.

Clay Native Search. Use Clay's built-in company search with filters for industry, employee count, location, and technology. Useful for building targeted lists from scratch when you have a clear ICP definition.

For most B2B teams, we start with a combination of CRM import (re-engage what you have) and lookalike audiences (expand into new territory). A typical starting list is 100 to 500 accounts.

Module 2: Enrich with AI Research

Raw account data is not enough to make outreach decisions. A company name and domain tells you nothing about whether they are a good fit or what pain points to reference in your messaging.

Clay connects to over 100 data providers. For each account, we enrich with:

• Headcount (current employee count and growth rate)

• Funding (last round, amount, date, investors)

• Tech stack (CRM, marketing automation, sales tools they use)

• Hiring signals (open roles for SDR, BDR, RevOps, GTM positions)

Then we run Claygent, Clay's built-in AI research agent. Claygent visits the company's website, reads their LinkedIn page, and generates a structured summary of what the company does, who they sell to, and what signals suggest they might need outbound infrastructure.

The output is 3 to 5 enrichment signals per account. These signals feed directly into scoring (Module 3) and sequence personalization (Module 6).

The difference between good and bad outbound starts here. If your enrichment is shallow (just company name and headcount), your scoring will be inaccurate and your sequences will be generic. If your enrichment captures real signals (they just raised Series B, they are hiring 3 SDRs, they adopted HubSpot last quarter), everything downstream improves.

Module 3: Score and Qualify with ICP Tiering

Not every company on your list is worth pursuing. Module 3 separates signal from noise.

Step 1: Hard disqualifiers. These are binary rules that immediately remove a company from the pipeline. Examples:

• B2C company (not your market)

• Fewer than 15 employees (too small to buy)

• More than 500 employees (too large for your offering)

• Government, nonprofit, or education sector

If a disqualifier triggers, the account is removed. No exceptions.

Step 2: Multi-factor ICP scoring. Accounts that pass disqualifiers get scored on four weighted factors:

• Company-Market Fit (industry, size, geography alignment)

• Growth Momentum (funding recency, headcount growth, expansion signals)

• GTM Maturity Gap (do they lack dedicated RevOps or GTM engineering? this is the opportunity)

• Intent Signals (hiring for relevant roles, adopting relevant tools, posting about relevant pain)

Each factor contributes to a composite score. The score maps to a tier: A (75-100, pursue immediately), B (50-74, engage with tailored sequence), C (25-49, nurture only), D (0-24, skip entirely).

Only A-tier and B-tier accounts proceed to contact discovery. This is how you avoid spending outreach budget on companies that will never buy.

We use Claygent to run the scoring logic inside Clay. The AI evaluates each account against the scoring criteria and returns a tier classification with a brief justification. This scales to thousands of accounts without manual review.

Module 4: Find the Right Contacts

Once you know which companies to pursue, you need to identify the right people inside those companies.

Clay's People Search lets you target by title, seniority, and department. For a typical B2B outbound campaign, we target:

• VP of Sales / CRO

• Director of Revenue Operations

• Head of Growth / Demand Gen

• Founder / CEO (at smaller companies)

For each contact, three checks run before they enter the sequence:

1. Persona match. Does their actual role match your buyer persona? A "VP of Sales" at a 10-person startup is a different buyer than a "VP of Sales" at a 500-person enterprise.

1. Employment validation. Are they still at this company? LinkedIn data can lag behind reality. Verify before sending.

1. Recent activity. What are their last 5 LinkedIn posts about? This becomes personalization fuel in Module 6.

The third check is often overlooked, but it is the single biggest lever for reply rates. A prospect's recent LinkedIn activity tells you what they care about right now. Reference that in your email, and you immediately stand out from the templates flooding their inbox.

Module 5: Multi-Provider Email and Phone Waterfall

This module solves one of the biggest problems in outbound: email coverage.

If you rely on a single data provider for email addresses, you get roughly 50% coverage. Half your qualified contacts never receive an email because you could not find their address.

A waterfall approach fixes this. Clay lets you chain multiple email providers in sequence:

1. Provider 1 runs first

1. If it returns empty, Provider 2 fires

1. Then Provider 3

1. Then Provider 4

Each provider checks a different database. Where one fails, another succeeds. The result is 80%+ email coverage across your target list. This is a drag-and-drop setup inside Clay. No code required.

The same waterfall pattern works for phone numbers, typically reaching 40 to 60% coverage across multiple providers.

Verification is mandatory. After the waterfall completes, every email address runs through a verification step. Invalid addresses, catch-all domains, and temporary emails get flagged or excluded. Sending to unverified addresses damages your domain reputation, which undermines every future campaign.

Module 6: AI-Personalized Sequences

This is where most outbound systems fail. Companies invest in data and tools but then send the same template to every prospect. The personalization is a mail merge token: "Hi {{first_name}}, I noticed {{company_name}} is growing."

That is not personalization. That is a template with variables.

Real personalization means each email references something specific about the prospect: their recent funding round, the RevOps role they posted last week, the LinkedIn post they wrote about pipeline challenges. Claygent makes this possible at scale.

For each contact, Claygent generates:

A 5-step email sequence

• Email 1: Lead with their specific signal (hiring, funding, tech adoption)

• Email 2: Share a relevant insight or case study

• Email 3: Reference something from their LinkedIn activity

• Email 4: Social proof from a similar company

• Email 5: Clean breakup with a direct ask

A 3-step LinkedIn sequence

• Connection request with a one-line contextual note

• Value message after connection (insight, not pitch)

• Soft ask for a conversation

Rules we enforce in every Claygent prompt: each email under 100 words, subject lines under 7 words in lowercase, no generic openers, no superlatives, and every email must reference a specific data point about the prospect.

The difference in results is measurable. Template-based outbound typically produces 0.5 to 1% reply rates. Signal-based personalized outbound consistently lands in the 3 to 6% range.

Module 7: Push to Sending Tools

Clay generates the sequences. Dedicated sending tools handle the delivery.

Email goes to Smartlead. Clay has a native Smartlead integration. Map your fields (email, first name, company, sequence steps) and push contacts directly into a Smartlead campaign. Configure sender rotation across multiple mailboxes, set daily sending limits, and schedule your send windows.

LinkedIn goes to HeyReach. Same pattern. Push contacts and sequences from Clay to HeyReach for LinkedIn outreach. HeyReach handles connection requests, messages, and follow-ups while rotating across LinkedIn accounts to stay within platform limits.

The coordination matters. Email and LinkedIn sequences run in parallel but with deliberate timing. A prospect receives a LinkedIn connection request on day 1 and the first email on day 2. The channels reinforce each other without overwhelming the recipient.

Module 8: Reply Monitoring and Classification

Sending is half the job. The other half is knowing what happens next.

Both Smartlead and HeyReach support webhooks. When a prospect replies, opens, clicks, or bounces, the event fires back to Clay (via n8n or a direct webhook).

An AI classification layer reads each reply and categorizes the intent:

• Interested (wants to talk, asks for more info, suggests a time)

• Not now (timing is wrong, but not a hard no)

• Not interested (clear rejection or unsubscribe request)

• Out of office (auto-reply, needs follow-up later)

• Referral (forwarded to someone else in the org)

Interested replies route immediately to Slack (for speed) and CRM (for tracking). Not-now replies get tagged for re-engagement in 60 to 90 days. Unsubscribes are removed permanently. Out-of-office replies pause the sequence and resume after the return date.

This classification runs automatically. No rep needs to read and sort through hundreds of replies manually. They see only the qualified, interested conversations.

Module 9: Measure What Matters

The output of this system is not "emails sent" or "open rates." Those are vanity metrics.

What matters:

• Qualified replies. How many prospects responded with genuine interest?

• Meetings booked. How many conversations turned into scheduled calls?

• Pipeline generated. What is the total value of opportunities created?

• Cost per meeting. How does this compare to the cost of hiring an SDR?

We track these per campaign and per module. If reply rates drop, the issue is usually in Module 6 (sequences) or Module 3 (targeting the wrong accounts). If email deliverability drops, the issue is in Module 5 (verification) or the sending infrastructure.

The system is designed to be debuggable. Every module has a clear input and output. When something underperforms, you know exactly where to look.

The Stack and What It Costs

The core tools for this system:

• Clay (research, enrichment, scoring, sequences): Launch plan at $185/mo, Growth at $495/mo

• Smartlead (email campaign delivery): Starting at $39/mo

• HeyReach (LinkedIn outreach delivery): Starting at $79/mo

• CRM (HubSpot/Salesforce): Free tier available

Total tooling cost: roughly $300 to $600 per month depending on your plan tiers and volume. Clay recently restructured their pricing (March 2026), replacing the old Starter/Explorer/Pro tiers with Launch and Growth plans while significantly reducing per-credit data costs.

Compare that to the fully loaded cost of an SDR: $70,000 to $90,000 per year in base salary alone, before commissions, benefits, management overhead, and ramp time.

This system does not replace SDRs. It replaces the manual research, data entry, and template-writing work that consumes most of their day. An SDR supported by this system spends their time on conversations, not spreadsheets.

For teams exploring this stack, we maintain a curated list of GTM tool recommendations and exclusive deals on the tools we use in production.

Who This System Works For

This playbook is not for every company. It works best when:

• You sell B2B and your ICP is definable (industry, size, signals)

• You have enough volume to justify automation (targeting 100+ accounts per campaign)

• Your sales cycle involves outbound prospecting (not purely inbound or PLG)

• You have a CRM in place, even if the data is messy

It works across verticals. We have deployed this for B2B SaaS companies, fintech platforms, professional services firms, staffing agencies, and managed service providers. The ICP definition and messaging changes. The architecture stays the same.

If you want to see the full system built live with a real client dataset, we published a step-by-step masterclass that walks through every module with a working Clay table.

Frequently Asked Questions

How do I set up outbound with Clay?

Start by sourcing target accounts using Clay's Find Company search or importing from your CRM. Build an account table with enrichment columns for headcount, funding, tech stack, and hiring signals. Score each account with Claygent using a structured ICP prompt. Find contacts at qualified accounts, run a multi-provider email waterfall for 80%+ coverage, and generate AI-personalized sequences. Push to Smartlead for email and HeyReach for LinkedIn.

What is ICP scoring in outbound?

ICP scoring is a system for ranking target accounts by how closely they match your ideal customer profile. It typically includes hard disqualifiers (binary rules that exclude a company immediately) and a multi-factor scoring model. Factors usually include company-market fit, growth momentum, technology maturity gap, and intent signals. The output is a tier classification (A, B, C, D) that determines which accounts receive outreach and what priority level they get.

How does waterfall enrichment work in Clay?

Add multiple email-finding providers in sequence inside Clay. Provider 1 runs first. If it returns empty, Provider 2 fires, then Provider 3, then Provider 4. Each provider checks a different database. Single-source email finding gives roughly 50% coverage across a target list. A waterfall approach pushes that to 80% or higher. Clay makes this a drag-and-drop configuration with no code required.

What reply rates does signal-based outbound produce?

Template-based outbound typically produces 0.5 to 1% reply rates. Signal-based outbound with AI personalization consistently produces 3 to 6% reply rates across our deployments. The difference comes from three factors: better targeting (only A and B-tier accounts get outreach), richer data (5+ signals per account feed the messaging), and genuine personalization (each email references specific prospect context instead of using generic templates).

How does Clay integrate with Smartlead and HeyReach?

Clay has native integrations with both tools. For Smartlead, configure the integration in Clay's settings, map your fields (email, first name, company, sequence steps), and push contacts directly into a campaign. For HeyReach, the same pattern applies for LinkedIn sequences. Both integrations support pushing enriched data and AI-generated sequences without manual export or import steps.

What does a Clay outbound system cost compared to hiring an SDR?

The core tooling (Clay, Smartlead, HeyReach) costs roughly $300 to $600 per month depending on plan tiers. A fully loaded SDR costs $70,000 to $90,000 per year in base salary alone. The system does not replace SDRs entirely. It replaces the manual research, data entry, and template work that consumes most of their day.

We build these systems for B2B teams as part of our Outbound Systems service. If you prefer the DIY path, start with our Clay Outbound Automation Masterclass.