ENTRY #004

​

The day I built the cleanest list of my career started like every other…

Just me, my laptop, and a goal: build a prospect list so dialed in that the messaging would write itself.

I’d done this over 500 times. Sales Nav, Clay, Apollo, Ocean.io, Google Maps, PhantomBuster, you name it.

I’ve scraped, enriched, and exported more lists than I can count.

But this one? It was different.

It started with deep crawling the web to find a hard-to-reach niche.

But one thing kept ringing in my ear

This list is only as strong as you clean it.

Yes, I had done a lot to find data

But I also have to set up layers in AI prompting to qualify the data and ensure it’s worth my outreach

It’s so easy to fall into the trap of rushing to launch.

But when your data is messy or misaligned with your offer, your campaign is already set up to fail, no matter how good the copy is.

So I thought I’d open up my process.

​

Here’s how I clean and qualify my lists now before ever sending a single message:

Step 1: Get the Data

Each tool gives you a different kind of mess. So I’ve learned to expect what needs cleaning depending on where I pull the data from:

🔹 Apollo

• With Instant Data Scraper, it’s mostly clean but missing key enrichment fields like company domains or LinkedIn URLs.
• LeadMagic is my favorite, more cleaned up, and I get a lot of company information like company domain and company LinkedIn URL in one go.

🔹 Sales Navigator

• PhantomBuster pulls messy data (especially names and titles).
• Prospeo is a game changer; it somehow exports much cleaner data than anything else I’ve used on LinkedIn.

🔹 Clay

• Very clean on the surface. But when using “Find People,” it gives you basic data; you still need to enrich both the contact and company profiles to get anything actionable.

​

Step 2: Qualify the Data (Because Not Every Lead Is Worth the Effort)

Sometimes a list is built on vague filters or scraped, in which case you most times cannot filter. So this is where I get selective:

• For light qualifying, I search by keyword and use checkboxes to qualify or score with Clay’s score row feature.
• For deeper qualifying, I read variables through variables already on the table or visit the site, extract info, and infer relevance, both through AI.

​

Favorite AI Models here:

• Claygent Helium: For visiting sites and pulling details (e.g., “Does this company serve ecommerce brands?”).
• GPT-4: Best for scanning variables and summarizing or qualifying based on conditions.
• GPT-4.1 Mini: Great for rewriting or generating short sentences for personalization.

​

Step 3: Clean the Data (Because Humans Read These Emails)

This is where I make everything human-readable and usable in my outreach copy:

🔹 Names:

• Split full names, remove emojis, and fix capitalizations using formulas (or GPT if too messy).

🔹 Company Names:

• Normalize names by removing “Inc.,” “Ltd.,” or “International.” Clay’s “Normalize Company Name” feature is super helpful here.

🔹 Job Titles:

• Rewrite job titles into clean, conversational phrases using AI (e.g., “Head of Operations Team ” → “Head of Ops”).

​

Rule of thumb:

- Try Clay’s free cleaning features first.

- Then, use formulas.

- Only use AI or paid enrichments when absolutely necessary. If it costs more than 2 credits per row, I skip it unless it’s critical and the list is relatively small

​

Step 4: Avoid Wasting Time (and Credits) on Weak Leads

For every column I create, I set conditions to ensure it only runs on qualified rows.

This tiny habit saves me from wasting credits on contacts who aren’t a fit.

​

Step 5: Validate Emails (or Risk It All)

I put this last, but honestly?

It should be the first thing you do.

A beautifully cleaned and qualified list still fails if the email bounces or lands in spam.

So I always verify every email, including catch-all domains, to keep bounce rates low and sender reputation intact.

​

Final Thoughts:

I’ve seen campaigns fail not because the message was bad, but because the data was dirty.

So if you’re building lists for outbound, don’t just build fast.

No matter what you build, always clean that data.

That’s how you get better replies, more meetings, and bigger wins.

​

P.S. Want me to break down how I write copy after the list is clean? Let me know.

​

Until the next entry, stay curious and keep building.

Natasha | GTME in the wild