Most enrichment advice reads like it was written by someone who's never opened a messy CRM. Generic tips. Vendor marketing. Theory without battle scars.

This handbook comes from the trenches. Every tactic has been tested on real systems with real problems. We share the failures as openly as the wins.

If you want "buy data and you're done," this isn't for you. If you want a system that gets smarter over time, keep reading. The further you go, the bigger the insights become.

A few years ago, Marina met an SDR leader whose Salesforce looked immaculate. Every field filled. Dashboards gleaming. She said:

"Your CRM looks top-notch."

She smiled. 

"Looks is the key word."

Six months earlier, they'd done a full enrichment project with a major data provider. Every empty field got filled. From the outside, it looked like a win. Inside, it was a disaster. Nobody trusted the data. Sales only used the CRM for compliance. The project didn't just fail - it destroyed the one thing that matters most: trust.

This isn't rare. After a decade working in CRM infrastructure, I've seen the same mistake repeated over and over. Teams confuse filling fields with building intelligence. They chase volume over quality. They skip the foundation work, then wonder why everything falls apart.

Here's the truth: bad enrichment is worse than no enrichment at all. Done right, enrichment turns noise into signal and gives your sales team real conversation weapons.

Most enrichment advice reads like it was written by someone who's never opened a messy CRM. Generic tips. Vendor marketing. Theory without battle scars

This handbook comes from the trenches. Every tactic has been tested on real systems with real problems. We share the failures as openly as the wins.

If you want "buy data and you're done," this isn't for you. If you want a system that gets smarter over time, keep reading. The further you go, the bigger the insights become.

Your CRM should be a competitive advantage. For most Series A to C companies, it's a liability.

Data goes stale within 90 days. Your go-to-market strategy rides on information that's 60% accurate at best. SDRs call people who left six months ago. Marketing targets the wrong decision-makers because company information is outdated. Win rates slip, but the data is too dirty to diagnose why.

Most teams treat enrichment like spring cleaning: do it once, feel good about it, then watch it decay. Then they wonder why their CRM is useless.

Save Time

Stop chasing ghosts. When contact data is current, SDRs stop wasting 15 hours per week on manual verification. Account executives walk into calls with context. Revenue operations stops firefighting and can focus on strategy.

Here's a simple model: a 10-person SDR team saving 15 hours weekly frees up 150 hours for selling. At a loaded cost of £75,000 per SDR, that time has real cash value.

Save Money

Lower customer acquisition cost by targeting people who can buy. One client spent six figures on tools to enrich anyone vaguely in their target market. We flipped the approach: qualify first, enrich second. Their CAC dropped 41% because sales time stopped burning on non-buyers. 

Marketing stopped funding personas that don't convert. The savings compound over time.

Make Money

Conversion rates improve at every stage. Contact-to-meeting rates go up with good enrichment. Meeting-to-closed deals improve when you reach decision-makers with real pain points.

Why it works: Sales walks into conversations with better intelligence. Gong analysis shows teams using enriched context convert at higher rates. You reach people who can move budget. Leads are qualified, so you hit fewer dead ends after discovery calls.

If you want a one-time data dump, just buy a CSV file. Strategic enrichment builds a closed feedback loop that improves every quarter.

Here's how it works:

1. Start with hypotheses about buyer signals
2. Put that intelligence in your sales team's hands
3. Use Gong or similar tools to capture what happens in conversations
4. Analyse what worked and what fell flat
5. Feed those learnings back into your enrichment logic
6. Repeat

Gong transcripts give you two layers of insight:

• Signal Discovery: Why did they take the call? What pain did they mention? Which signals should we track going forward?
• Quality Control: Did sales use the enriched data? Was it accurate? What resonated with buyers?

That feedback loop transforms "we bought data" into "we built a learning engine."

Because it's hard. Here's what goes wrong:

• No Patience: Teams want results immediately. They skip three weeks of discovery work and jump straight to "enrich everything."
• Wrong Ownership: Marketing owns the CRM but sales won't buy in. Or sales owns it without data discipline. Misalignment creates a foundation built on sand.
• Tool-First Thinking: Buying a data provider solves maybe 10% of the problem. The other 90% is strategy, governance, and continuous improvement.
• No Feedback Loops: Enrich once, never revisit the logic. Markets change but the enrichment strategy doesn't.

All of these problems are solvable. That's what the rest of this handbook covers.

A great enrichment strategy on broken infrastructure is worthless. Most failures aren't data failures - they're foundation failures.

Skip the groundwork and you get: duplicates multiplying, conflicting enrichment rules, broken attribution, garbage pipeline forecasts, and multiple people contacting the same prospect with different messages.

Before touching anything, you need to understand how your system works today. This isn't a quick audit. Plan for three weeks. No building. No enrichment. Just discovery.

Find What Works

What data is currently good and used? Don't rely on what people claim - observe their work.

Sit with an SDR or sales rep. Watch where they click. If they copy information from LinkedIn into Salesforce, you've found a workaround. If they always use "Company Domain 1" and ignore "Company Domain 2," that's the field they trust.

Most CRMs have multiple domain fields and phone number fields. Users know from experience which ones burn them. Believe their behaviour.

Understand Your Real Workflows

What fires when a lead enters your system? Score, route, add to sequence. Then you might discover that scoring is broken, routing is random, and sales ignores 80% of leads anyway. The documented process and the actual process rarely match

Identify Your No-Contact Lists

Who must never be contacted? This is critical:

1. Current customers (unless it's upsell context)
2. Active opportunities
3. Partners
4. Investors
5. Recently churned customers in their cool-off period
6. Companies in litigation with you
7. Competitors

Interview your revenue operations, SDR leaders, sales leaders, and account executives. Each person holds a piece of the puzzle.

How to Conduct Discovery Interviews

Watch people work. Ask them: "Show me your screen. What do you do when a lead lands? Why did you click that?"

The workarounds reveal where the truth lives. Talk to:

1. Revenue operations for technical debt and makeshift automations
2. SDRs for which data they trust versus which fields they ignore
3. Sales leadership for missing reports and shaky forecasts
4. Account executives for real qualification versus what the playbook says

Example Gaps You'll Find

Example 1: Everyone says "We don't contact customers." In reality, half the AEs are working upsell opportunities but those contacts aren't flagged.

• Fix: Auto-flag customers, route them correctly, enrich with product usage data or revenue risk scores for prioritisation.

Example 2: The process says "All web leads go to SDRs." In reality, SDRs ignore most of them because they're low quality.

• Fix: Add automatic qualification at entry using ICP scoring, so only good leads reach the SDR queue.

Before any enrichment happens, build suppression infrastructure. This is your "do not mess this up" system.

Simple CRM Setup

If you have a basic CRM, start with two or three suppression tables:

1. Do Not Contact (GDPR requests, opt-outs, competitors)
2. Current Customers
3. Active Opportunities

Mature CRM Setup

If you have a more complex operation, you might need twenty or more suppression categories. Use a checklist approach so nothing gets missed:

1. Legal/Compliance: opt-outs, bounce lists, companies in litigation
2. Customer: current customers, product-specific customers, recent churn, VIP accounts
3. Sales Process: active opportunities, late-stage deals, named account assignments
4. Strategic: partners, investors, competitors, advisors
5. Engagement: trial users, event attendees, "not interested" cooldowns
6. Operational: employees, test data, suspected duplicates

Don't ask "What should we suppress?" Ask "Which of these categories apply to us?"

We've worked in CRMs for decades. How many native CRM workflows do you trust? Not many.

Clay gives you visibility, control, speed, and reliability. Most importantly, it keeps working without constant maintenance.

How to Structure Suppression in Clay

Build modular suppression tables in Clay, each with its own refresh schedule:

• Current customers: refresh daily
• Active opportunities stage 4 and above: refresh hourly
• Partners: refresh weekly
• GDPR opt-outs: refresh daily

Then aggregate everything into a master suppression list. Match at the right level: email address for individual people, domain for companies.

When Suppression Isn't Simple Yes/No

Suppression isn't always binary. Context matters.

If acme.com is a partner:

1. Suppress them for outbound prospecting
2. Allow contact for customer success or upsell conversations

Build your logic to express intent, so the system can apply human judgement automatically.

Use Formulas For...

Binary checks where the answer is yes or no:

1. Is this person a current customer?
2. Does this domain match our partner list?
3. Is this opportunity in stage 4 or higher?

Formulas are fast and deterministic.

Use AI For...

Grey areas that require reasoning:

1. Does this person have budget authority?
2. Is this company showing signs of the pain we solve?
3. Does this buyer profile fit our ideal customer based on multiple signals?

How to Test AI Enrichment

1. Small batch: 20-30 rows. Review every result manually

2. Medium batch: 100 rows. Spot-check 25 results

3. Large batch: Spot-check again at scale

Target accuracy: 95%+ at scale, 99-100% for narrow, high-value lists. If you're below 90%, fix your prompt before proceeding.

Total timeline: three months.

• Weeks 1-3: Discovery and mapping
• Weeks 4-5: Design and sign-off
• Weeks 6-8: Build in Clay
• Weeks 9-10: Test and refine
• Weeks 11-12: Launch and monitor

Get alignment between the system owners (usually revenue operations) and heavy users (sales leadership). Present a prioritised list of gaps you found and the trade-offs of addressing them in different orders.

"We need volume for the sales team." 

Those five words have wrecked more CRMs than any software bug.

The pattern plays out like this: loose criteria, bloated database, 70% of contacts aren't real fits, SDRs stop trusting the data, sales bypasses the CRM, the system gets reduced to a compliance checkbox, and marketing can't prove ROI. Six months later: 

"We can't trust any of this."

The core mistake is equating "companies that match a vague filter" with "people who could buy from us."

Title alone tells you almost nothing. "Head of Sales" at a 20-person company is different from "Head of Sales" at a 5,000-person company. Different authority, different problems, different buying power.

Use context instead: organisational position, company size, industry norms, and historical patterns from your own closed deals. Score the likelihood of budget authority based on all of these factors combined.

A single role can have 50 different title variations. Six personas might yield 300+ title variants. You need AI to map titles, reason about context, and score fit probability.
Feed the AI: title, company size, reporting structure, industry, and patterns from your wins. Get back a probability score and the reasoning behind it.

Enterprise Deals

Enrich deeply. Provide intelligence that's worth a meeting. Knowing they raised funding or liked a LinkedIn post won't cut it. You need real insight.

Mid-Market Deals

Confirm ICP fit and identify decision-makers. Don't over-invest enrichment where the deal size doesn't justify the cost.

The Metrics That Matter

Ignore vanity metrics like total records enriched, database size, or coverage percentage.

Track these instead:

1. Contact-to-meeting conversion rate
2. Cost per qualified opportunity
3. Sales feedback scores on data quality

The goal is higher conversion rates, sales confidence, and lower customer acquisition cost.

With Clay set up, daily work is minimal.

Suppression tables update automatically. Daily refresh is fine. Inbound lead flows run in real-time or near real-time. High-intent leads get enriched within minutes. Maintenance is periodic, not constant.

Update when you have: new product lines, additional CRMs to integrate, new partnership relationships, sales process changes, or compliance requirement changes.

Clay's modular setup makes these updates straightforward.

Revenue Operations owns the system. They manage suppression lists, read system logs, and troubleshoot issues.

Sales and SDRs use the outputs. They should not touch Clay directly. Train them on what data exists, why it's trustworthy, and how to use it in conversations.

Use Gong data to prove the value:

1. "Calls using enriched intelligence had X% conversion"
2. "Calls without it had Y% conversion"
3. "The difference is Z%"

Use conversation transcripts to drive signal discovery and quality control. Feed what you learn back into your enrichment logic.

Double down on what moves the needle. Cut what doesn't work. Fix data quality issues at the source.

Clay scales well if your logic is clean. The risk isn't technical - it's data quality decay and team adoption.

Watch for warning signs: conversion rates slumping, teams bypassing the system, suppression misses, and enrichment credit burn accelerating.

Before you touch any data or hire anyone to help, define what success looks like six months from now. Get specific: What's your target for closed deals? Average deal size? Customer acquisition cost? Sales cycle time? What does a typical day look like for your sales team when the system works?

Once you have that vision, work backward to identify the data, logic, and operations you need to get there.

• Weeks 1-3: Discovery and system mapping. This is when you (or whoever you hire) figures out how your CRM works versus how it's supposed to work.
• Week 4: Design review and sign-off. You'll see a prioritised list of gaps and a plan to fix them. Both your system owners (RevOps) and heavy users (sales leadership) need to approve.
• Weeks 5-7: Build your Clay connections, suppression logic, enrichment waterfalls, and routing rules.
• Weeks 8-9: Test everything in batches. Tune your AI prompts until accuracy hits your targets.
• Weeks 10-12: Launch, train your teams, monitor results, and start analysing Gong transcripts to see what's working.

At Series A, you have two paths. You can learn to do this yourself using this handbook as your guide, or you can hire a GTM engineer or operator to build it for you.

If you're hiring, look for someone who:

1. Asks about your end state before proposing solutions
2. Talks about suppression and data quality before talking about enrichment volume
3. References feedback loops and continuous improvement, not one-time fixes

  Red flags:  

They show up with a standard package, skip discovery, promise fast results, or focus only on the tools they'll use rather than the problems you're solving.

At Series B, your CRM complexity usually warrants an agency that's done this multiple times across different tech stacks and industries.

Look for an agency that:

Ask them: 

"Walk me through how you'd approach our discovery phase." 

If they can't articulate a clear 2-3 week process, they're winging it.

At GirlsWhoClay.io, Marina & I (Stephanie!) work with both Series A companies who want an operator embedded in their team and Series B companies who need full agency support. We follow the exact process laid out in this handbook because we wrote it based on what works.

Whether you work with us or someone else, use the red flags and green flags in this chapter to choose a partner who will build a system you can trust.

If your prospect arrives with a fixed wishlist and no appetite for being challenged, disqualify them. They want an order-taker, not a strategic partner. The plan they have won't fit their reality, and your discovery findings will be ignored.
You can't fix a system when the client won't let you diagnose the problem.

Look for clients who are:

1. Self-aware about their CRM problems
2. Believe the CRM should be the single source of truth for their business

These characteristics predict success. They're coachable, they'll trust your process, and they understand that good enrichment is about strategy, not just buying data.

Your CRM is either a growth engine or a growth brake. There's no middle ground.

If you're treating enrichment as a one-time project, optimising for volume over relevance, or letting misaligned ownership create chaos, you're bleeding pipeline and brand equity every single day.

The fix isn't buying more data vendors. It's treating your CRM as a strategic asset. Build the foundation. Qualify for relevance before you enrich. Install feedback loops that make your system smarter over time. 

Never trade trust for the illusion of completeness.

Stop buying data dumps. Start building intelligence.

Enter your email address below to get early access!

We engineer brand-led GTM systems that transform your CRM from operational liability to competitive advantage for Series A & B Companies in UK & EMEA.

About

Info

CONTACT

© Stephanie Holland, 2025.  All Rights Reserved