Adam Robinson wants to grade your email list A to F

You've run this play before: export the list, feed it into whatever validator your team is paying for this quarter, and get back a spreadsheet with three or four buckets of confidence and a pile of "unknown" rows that somehow make you feel worse than before you checked. Every vendor in this category has spent years arguing over how to sort emails into those buckets. Adam Robinson skipped the sorting argument and went straight for the report card.

Robinson runs RB2B and Retention.com, and he's been publishing his newsletter at a steady weekly clip all summer: a July 15 post arguing signal orchestration is dead, a July 23 reaction to HubSpot reversing its Contact Discovery feature, a July 31 update that his new venture MoltSets had crossed 5,000 waitlist signups. On August 8, he published the announcement that matters here: a new email validation tool that ditches the standard pass/fail result for a letter grade, or what he's calling a confidence score.

Inside the new email validation tool from RB2B's Adam Robinson

The mechanics are simple to describe. Instead of handing you "valid" or "invalid" and leaving you to guess what to do with the addresses that don't cleanly fit either bucket, MoltSets appends a grade to every email it delivers: A or B for sendable, C for catch-all domains, D for addresses with a known history of negative activity, and F for anything that flat-out doesn't resolve. Robinson frames it as a deliverability fix, not a list-hygiene nicety, and he's pricing access aggressively to get it adopted fast.

“We're going to validate every email we deliver...append an 'email grade' or 'confidence score.'”

He also puts a name on why this matters right now: he attributes a chunk of the industry-wide bounce-rate spike this year to authentication changes at Microsoft and Google. That part isn't controversial. Anyone running cold email in 2026 has felt the tightening from both mailbox providers, and a validation layer that catches more of the risky middle ground before you send is a reasonable response to that pressure, whatever letter it's wrapped in.

The grade tiers, and what to do with each one

Here's Robinson's framework laid out as an action table, because a grade is only useful if it changes what you actually do with an address. This is his own tiering, not an invented one:

As a triage system, that's genuinely useful. It tells a sender where to spend their best-warmed domains and where to hold back, which is a more actionable output than a flat valid/invalid split ever was. The question is what that grade is actually measuring, and how long it stays true after you get it.

Both numbers come straight from Robinson's own newsletter, not a third-party audit, and we're flagging that plainly rather than laundering vendor-reported figures as verified adoption data. A waitlist count tells you people are curious. It doesn't tell you retention, real send volume run through the tool, or how the grade holds up against a live campaign at scale. Treat both figures as marketing signal, not proof of performance.

The genuine upgrade this email validation tool makes

Give credit where it's earned. Most validation tools have historically handled catch-all domains one of two lazy ways: lump them in with "valid" and let senders find out the hard way when bounce rates climb, or lump them in with "unknown"/"risky" and let senders throw away addresses that might have been perfectly fine. Neither is honest about what a catch-all server actually tells you, which is nothing, by design. We've written before about how to handle catch-all domains without wrecking cold email deliverability, and the core problem has always been that catch-all isn't a verification failure. It's a structural property of how the receiving server is configured, and it deserves its own tier rather than getting forced into a binary it doesn't belong in.

Giving catch-all its own grade, C, instead of quietly folding it into A/B or dumping it with the invalid pile, is the one piece of this announcement that's a real methodological improvement and not just a rebrand. It won't tell you whether a specific catch-all address is a live inbox. Nothing can, short of sending to it. But it stops the tool from lying to you by omission, and that's worth something.

What this email validation tool still doesn't solve

Here's the part the letter grade doesn't change, no matter how it's packaged. We covered this exact mechanism in an earlier piece on why 'valid' doesn't mean deliverable, pulling from Jesse Ouellette's research at LeadMagic: a verification result of any kind, whether it's a word or a letter, only proves the mailbox accepted mail at the specific moment the check ran. It is a timestamp wearing a verdict's clothing. Mailboxes get closed. Domains get reconfigured. Catch-all behavior flips on and off. None of that shows up retroactively on the grade you were handed last month.

A grade doesn't fix that. It relabels it. An "A" reads as more authoritative than a "Valid" tag ever did, which is precisely the risk: a letter grade feels graded on a curve, like a durable assessment rather than a single ping. It's still a single ping. The confidence score decays at exactly the same rate the old binary check did, because it's built on the same underlying signal, sampled once. What's changed is the label's psychological weight, not the shelf life of what's underneath it.

Does a letter grade fix cold email deliverability, or just relabel it?

This is the question a practitioner should actually ask before switching tools: does the A-through-F system change your re-verification cadence, or does it just make the same stale-data risk feel more reassuring because it arrives dressed as a grade instead of a coin flip? Our own answer, laid out in building a re-verification cadence by industry decay rate, is that the cadence is the actual fix, and it was always going to be the fix regardless of which vendor's letter or word sits next to the address. B2B lists decay somewhere between 2% and 3% a month on average. That number doesn't care what grading system checked the list last. It keeps ticking whether your validator hands you a binary or a report card.

The practical upshot: use a confidence score, MoltSets or otherwise, to decide where to send today and where to slow down. Don't use it as a reason to stop monitoring sender score between validation passes, and don't let a clean A/B grade talk you out of the discipline you'd apply to any list, which is running a deliverability audit on your sending infrastructure on a real schedule, not a one-time gate you clear and forget. The grade is an input to that process. It was never going to be a replacement for it, and to be fair to Robinson, he's not claiming otherwise; his framing is squarely about routing sends by risk tier, not about promising permanence.

That's a fair thing to build, and a fair thing to charge $27 a month for if it works as described. The failure mode isn't the product. It's the buyer who reads a letter grade as a guarantee instead of what it actually is: better-labeled uncertainty, current as of the moment it was issued.

Do this next

If you're evaluating MoltSets or any confidence-score validator, don't just check whether the grading logic makes sense. Check whether it changes how your team behaves after the report comes back. Route your A/B addresses into your primary cold email sequences immediately. Slow-roll the C tier onto separate, lower-reputation infrastructure with tighter bounce monitoring, since catch-all is a real-time unknown, not a resolved one. Suppress D and F outright rather than arguing with the grade; that's what the tier is for.

Then do the part the grade can't do for you: put a date next to every validated address and a re-check trigger tied to your actual send volume, not a calendar reminder six months out. A list you're mailing weekly needs a tighter re-verification window than one you touch quarterly. Build that cadence once, and it doesn't matter whether the tool in front of it hands you a letter or a word. The grade is a snapshot. Your process is what keeps the snapshot honest.

See where you are cited today

A free snapshot audit of your rankings and AI citations before we ever talk.