9 Checks I Run Before Any Outbound Tool Goes Live (B2B Sales Prospecting Checklist)

Who this checklist is for

I manage quality and brand compliance at a B2B SaaS company. Every batch of list data and email copy that goes into our outbound motion passes through me first—roughly 200 per quarter. In 2025, I rejected about 30% of first deliveries purely on deliverability risk. Not obvious spam. Just things that looked fine and would have burned domain reputation because nobody checked the details.

This is the checklist I use to triage. It's nine steps and usually takes two or three review sessions to run through, depending on how many tools you're evaluating.

Use it if:

• You're standing up an AI SDR or outbound sequence platform
• You're comparing options like okki-go vs ZoomInfo
• You're trying to decide when a LinkedIn tool is worth paying for versus when it just adds noise
• You just received a cheap contact list and you're tempted to "just try it"

Skip it if you want inspiration. This is a gate, not a brainstorm.

The nine steps

1. Define your ICP before you open any tool

Most teams do this backwards—they sign up for a platform, then figure out who they're looking for. That's the single biggest cause of rework I see.

Before you pull a single email, write down: employee count band, industry tags, tech stack signals, and title level. Three lines max. If your team can't agree on the ICP in fifteen minutes, the problem isn't the tool. It's the positioning.

Checkpoint: Every reviewer can state the ICP in one sentence. If they can't, stop.

2. Write your verification standard before importing anything

Verification isn't just "run the tool." It's the set of rules you've agreed on: what confidence score passes, how you handle duplicates, what your catch-all domain policy is, and what makes you throw out an entire batch.

We got this wrong in 2022. We ran verification on a 40,000-contact batch, saw an OK overall number, and shipped it. The catch-all domains came back with an 18% bounce rate. Took us a month to recover domain reputation. Now every contract includes catch-all rules in writing.

Checkpoint: Verification standard signed off before the list arrives. Not after.

3. Read the email verification API documentation like it's a contract

This is the step most teams skip. They look at the demo, like the interface, and buy. Then six months later they discover the validator was guessing catch-alls instead of actually detecting them.

What to look for in an email verification API documentation review:

• Verification method: SMTP handshake, predictive model, or both? Predictive validators are faster but less accurate on newer or lower-volume domains.
• Catch-all handling: Does the API default to "valid," "invalid," or "unknown" for catch-all domains? That single word is your bounce rate's biggest hidden variable.
• Real-time vs batch: Can you validate in real time on submit, or only in bulk after import?
• Confidence scoring: Do you get a boolean, or a score from 0 to 100? A clean yes/no doesn't let you tier your outreach.
• Rate limits: Does the API degrade silently past a threshold? If it does, and your validation runs at 2am, you're paying for half a result.

We evaluated one vendor whose demo was flawless—until we read the docs and realized their "verification" was checking against their own historical bounce log. Which meant our first thousand contacts had no signal at all.

Heads-up: If the documentation doesn't mention catch-alls, treat that as a red flag, not an oversight.

4. Enrich with a waterfall, not a single source

Single-database enrichment goes stale fast. Contacts change roles, companies rename, email addresses go dormant. Waterfall enrichment—where multiple sources are chained so the next source fills in what the previous one missed—raises both hit rate and accuracy.

Checkpoint: Ask how many sources are in the enrichment layer. If it's one, push back.

5. Send small before scaling cold, especially without intent signals

For genuinely cold lists—no web visits, no content engagement, no external intent data—don't send the full batch.

Send 200. Watch three things: bounce rate, spam complaint rate, and first-touch reply rate. If bounce rate is over 3%, you know the whole batch is compromised. If complaints are over 0.1%, you stop.

Checkpoint: 72-hour hold on small sends before scaling.

6. Scope your LinkedIn tool properly

If you're asking what a LinkedIn tool is and when a B2B sales team should use it, the one-sentence version is: it's any tool that reads LinkedIn profile data, automates engagement, or pipes LinkedIn signals into a CRM sequence.

When it's worth using: when you're doing account-based plays with high ACV, and context matters more than volume. Target lists in the hundreds, not thousands. That's where a LinkedIn tool earns its keep. For high-volume cold outbound, LinkedIn automation gets throttled and your accounts get flagged—which is how we learned to stop trying to force it.

7. Run a pilot with a human in the loop

Full automation sounds great until the AI writes a personalized email referencing a conference your prospect didn't attend two years ago.

The pilot should have one human approving every send. Yes, it's slower. It's also faster at finding problems. Our rule: for the first 500 messages in a pilot, edit or reject each one. By week two, the rejection rate drops below 3% and you can open up partial automation.

8. Calculate total cost of ownership, not the per-seat price

This is the trap almost every team falls into. A platform at $99/month sounds cheaper than $500/month. Until you add the SDR hours spent cleaning bad data, the email credits used replacing invalid addresses, and the recovery period after a domain reputation hit.

Do the math roughly: if the cheap list has 15% more bounces and you send to 10,000 contacts a month, that's 1,500 wasted recipients every month. Then add the cost of re-verifying and rebuilding. The $99 platform is usually not the cheaper option.

We paid for this lesson once. A list that saved us $2,000 in quote turned into three weeks of rebuilding plus a customer relationship I'd rather not have contaminated.

Checkpoint: Compare on cost per verified contact, not cost per seat.

9. Monitor deliverability and domain health, not just open rates

Open rate is a vanity metric. Deliverability is the one that decides whether you have a pipeline next quarter.

Track: domain reputation, sender score, spam complaint rate, and bounce rate. If any of the four starts sliding, pause outbound. The dashboards inside your platform don't always reflect reality.

Things to watch out for

Here's what's tripped me up repeatedly.

• Don't buy a list as a one-off. If a vendor sells you a file of names and emails and walks away, you're the product. Data decays and you have no upstream to refresh from.
• Don't confuse "integration" with "native." Many platforms say they do email verification because they call a third-party API. That's fine, but know whose model is doing the work, because you can't escalate to anyone when it's wrong.
• Don't lock scope during the eval. I once believed a sales rep's framing on "intent data" during the demo, then found out afterwards it was title match. Spent two weeks re-running scoring.
• Don't skip review. Even one person spending half a day a week on samples catches the big stuff.

One last thing on the whole checklist: the value isn't in the tool. Tools amplify whatever you already do. Clean ICP definition, honest data, and a real verification standard will do more for you than switching platforms ever will.