Gary Halbert's A-pile: the mail-sorting idea that still decides open rates

The Gary Halbert A-pile/B-pile idea starts before anybody reads the letter. People glance at the mail and decide what looks personal. Those pieces make the A-pile and get opened. The obvious promotions land in the B-pile, where the trash is an immediate possibility. When I build a B2B mailer, I want the outer to look like one person sat down to contact another.

Halbert used a normal white envelope with a live First-Class stamp. He left off the teaser copy, mailing label, and company name in the return-address block. Handwriting helps in the right circumstances. It is not the main event. Halbert never published a reliable open-rate benchmark either, so I use this as a sorting model, not as proof that a particular envelope always wins.

The Gary Halbert A-pile B-pile framework

Halbert laid out the idea in a 1984 letter to his son Bond, which was later collected in The Boron Letters. His drawing of the B-pile envelope was hard to misread. It had teaser copy, a company corner card, a bulk-rate mark, and a pasted address label. Commercial on sight.

For the A-pile version, he stripped those signals away. The envelope carried the sender's return address, a live stamp, and a recipient address written or typed directly on the outer. I keep coming back to the word "typed" because summaries often reduce the whole lesson to "handwrite everything." Halbert did not. His target was the visible machinery of mass mail.

The original Boron chapter in the Gary Halbert Letter archive is worth reading without the folklore layered on top. Halbert said everyone opens personal-looking mail while only some commercial mail gets opened. That is salesmanship, not a number I would put in a forecast. I have not found an independent test supporting "everyone," and a physical envelope cannot report an open the way email software claims to.

I still use the idea because the envelope is judged before the offer. If the outer announces a promotion, the strong letter inside may never get a chance.

What actually made Halbert's envelope look personal

The original treatment was mostly restraint. I would leave off the slogan and the product promise. The fake-check window has no place here. Bright artwork made to impress an envelope-design judge works against the assignment.

Halbert expanded on the lecture later in his newsletter. This time he specified a plain No. 10 envelope, a typewritten address, a First-Class stamp, and a corner card carrying the sender's name and address. He warned readers about the contents too. Open a quiet envelope and find a loud brochure package, and the personal impression collapses immediately.

So I do not brief a B2B team with "make it look handwritten." I ask whether a real person would send this to another real person. Put a glossy logo and promotional teaser on a pen-addressed outer, then add a generic job-title salutation, and it still looks manufactured.

Teams regularly muddle permits and indicia. USPS calls the printed postage block a permit imprint indicia. Saying "no indicia" does not remove the need for valid postage. If I want the Halbert look, I choose a stamp. I do not send an unmarked envelope.

How to translate the A-pile into modern B2B mail

I start with the account. Always. A beautiful personal-looking envelope addressed to somebody who left the company is expensive undeliverable mail. First verify the person, the company, and the exact delivery point. Central receiving may require a floor, suite, or department. Our guide to address verification for direct mail covers the dull work required to reach the right desk.

Only then do I decide how much visible human effort the account deserves.

With a short list of high-value accounts, the rep can write the address and the note. Legibility wins over charm. I also put an actual sender name in the return address instead of manufacturing mystery. A mystery outer may earn an open, then feel like a trick as soon as the pitch appears.

Volume changes the production choice. Services such as Handwrytten use robots that hold real pens and write cards and envelopes. Useful, but narrow. The robot fixes a production problem; it cannot supply personal relevance. I would fund the machine-written note about the recipient's real situation before a hand-penned paragraph full of generic praise.

Typing is a perfectly sound option. USPS recommends typewritten or machine-printed addresses for automation readability. When handwriting is the choice, I insist on dark ink against a light background, complete delivery information, and a clean address block. The postal system gets the first reading opportunity, not the buyer.

The stamp is a handling decision with a creative consequence. A live stamp reinforces the personal look. A permit imprint is easier to produce. I settle that argument with a controlled test, keeping the letter, audience, offer, and follow-up identical.

Lumpy mail is a different bet

A bulky envelope gets noticed because it feels odd in the hand. I do not call that classic A-pile camouflage. In Halbert's Boron example, the "grabber" was a small bag of dirt inside the envelope. The recipient saw it after opening. The outer itself stayed plain.

B2B teams use dimensional items because an executive assistant or mailroom may give a box more consideration than routine advertising. That can earn the package a handoff. It can just as easily raise a security question or run into a gift policy. Worst of all, it can become a mystery package that nobody has a reason to deliver to the named person.

Postage is where some clever concepts come apart. The USPS nonmachinable criteria say pens, keys, coins, and similar objects can create uneven letter thickness. Rigid pieces may become nonmachinable, as can some unusual closures. A flat that fails the USPS flexibility or uniform-thickness standards counts as a parcel. Since rules and prices change, I have the mail house or a USPS Mailpiece Design Analyst inspect the finished sample before production.

My rule for lumpy mail is severe: the object must explain the message. For the broader format decision, I compare a postcard, letter, and dimensional mail piece against the amount of explanation required and the account value involved. A small physical sample for a materials buyer has a job. A random plastic toy sent to a finance executive is just texture.

The open is only the first gate

Halbert's later lecture makes a point that short summaries often drop. The piece has to remain in the A-pile after the envelope opens.

That means the first thing inside should make sense. A short letter from a named person works. So does a note attached to a useful, relevant enclosure. A stack of glossy sheets, a giant QR code, and a fake handwritten font tell the recipient that the envelope performed a costume change.

The copy still needs a reason to exist. Name a problem specific to the role or account. Give the recipient something worth keeping, such as a compact benchmark, a marked-up observation, or an invitation with real context. Then make the next step easy. The anatomy of a high-response B2B mailer explains how the outer, letter, proof, and reply path carry separate parts of that job.

Don't counterfeit privacy. Marking a cold sales letter "confidential" or "personal" does not make the relationship personal. I would rather let the handwriting and named sender imply human attention, then identify the company plainly inside. The test is simple: will the recipient feel interested after learning who sent it, or fooled?

How to test the idea without inventing an open rate

You usually cannot observe a physical open. A QR scan proves a scan. A reply proves a reply. Neither tells you how many envelopes were opened and read without action.

Run the useful test instead. Randomly split comparable recipients between your current outer and an A-pile treatment. Keep the contents and sales follow-up fixed. Track returned mail, site visits tied to the piece, replies, meetings, qualified opportunities, and revenue. A holdout group is even better when the account list can support it.

Do not change the list, offer, envelope, letter, postage, and sales cadence in one pass. You will get a result and learn almost nothing. Start by testing the outer. If the A-pile version produces more business response, test handwriting against clean typing or a live stamp against the postage method you already use.

I would not declare a winner from a handful of executive accounts. Small B2B lists wobble. Repeat the treatment across later cohorts and judge it on the business outcome you wanted, not compliments about the envelope. Our guide to measuring direct mail ROI lays out the rest of the attribution work.

Frequently asked questions

What is Gary Halbert's A-pile and B-pile theory?

Gary Halbert's A-pile and B-pile theory says recipients separate personal-looking mail from obvious commercial mail before opening it. The A-pile gets priority. The B-pile has to survive immediate judgment beside the trash.

It is a useful creative model, not a measured law of behavior. That caveat gets lost when people repeat the lecture as an open-rate statistic.

Did Gary Halbert recommend handwritten envelopes?

Yes, but not exclusively. In The Boron Letters, Halbert showed a delivery address that could be handwritten or typed directly on the envelope. His later lecture described a typewritten address.

The bigger rule was no pasted label and no obvious sales treatment. I would take a clean typed address over fake-looking script every time.

Do real stamps increase direct mail open rates?

Real stamps can support a personal-mail signal, which is why Halbert favored them. I have not seen a credible universal lift that you can plug into a forecast. Audience, sender, offer, and the rest of the outer all interfere.

Test the stamp against your normal postage on comparable names. And remember the boring part: postage methods, eligibility, and prices move, so confirm the current setup before the full run.

Is lumpy mail part of the A-pile strategy?

Not in the strict sense. A lumpy envelope announces that something unusual is inside, while Halbert's A-pile outer tried to pass as ordinary personal correspondence. His bag-of-dirt example was an interior attention device.

Lumpy mail can still work for B2B. Give the object a reason to be there, check the recipient's gift and security rules, and get the finished piece reviewed for USPS machinability. A bump without an idea is only expensive postage.