An eighth of an inch decides whether your mailing costs sixty-two cents a piece or a dollar ninety.

Postal prices are the one thing in prospecting that is genuinely public. They are filed with a regulator, published in a schedule anyone can download, and they change on a published date. After a series on channels where every number comes from somebody selling the channel, that is a relief.

The catch is that the published price is not a price, it is a grid, and the grid is discontinuous. The money in a direct mail campaign is not won by negotiating; it is lost by a designer choosing an envelope, or by an account manager adding one more insert.

The three cliffs

A mailed letter has to satisfy three constraints at once to stay in the cheapest category: no more than a quarter of an inch thick, no more than three and a half ounces, and a length-to-height ratio between 1.3 and 2.5. Miss any one and the piece reprices.

The three thresholds at which a mailed letter repricesTable setting out the three physical thresholds at which a mailed letter in the United States changes price category, with the effect of crossing each one quantified against published commercial automation prices in force from July two thousand twenty six. The first threshold is thickness, fixed at one quarter of an inch. A piece exceeding it is classified as a flat rather than a letter, and on the commercial automation five digit tier the price rises from sixty two point one cents to one dollar and two and a half cents, an increase of about sixty five per cent, while in the advertising mail class it rises from thirty nine point five cents to seventy eight point three cents, an increase of about ninety eight per cent. The second threshold is weight, fixed at three and a half ounces. Commercial first class letters are priced flat across the entire band from one ounce to three and a half ounces, so a mailer pays the same price all the way up and then, at the next tenth of an ounce, reprices as a flat at the four ounce breakpoint, taking the commercial automation five digit price from sixty two point one cents to one dollar and eighty nine and a half cents, an increase of about two hundred and five per cent. This is the most severe discontinuity in the grid. The third threshold is the aspect ratio, being length divided by height, which must fall between one point three and two point five inclusive. A square envelope has a ratio of one and is therefore non machinable, as is a long narrow envelope above two point five. In the advertising mail class a non machinable letter is expressly priced as a non machinable flat, so the five digit price rises from thirty nine point five cents to eighty eight point six cents, an increase of about one hundred and twenty four per cent. The table concludes that the cost of a campaign is decided by the choice of envelope and the number of inserts rather than by negotiation.Three cliffs, and none of them is negotiableCommercial automation, five-digit tier, prices in force from July 2026.A quarter inch of thicknessOver it, the piece is a flat. First-Class: $0.621 becomes $1.025.Advertising mail: $0.395 becomes $0.783.+ 65 %+ 98 %Three and a half ouncesCommercial letters cost the same all the way up, then reprice asa four-ounce flat: $0.621 becomes $1.895.+ 205 %A ratio outside 1.3 to 2.5Non-machinable. In advertising mail, a non-machinable letter ispriced as a flat: $0.395 becomes $0.886.+ 124 %
Three constraints, held simultaneously. The weight cliff is the cruellest, because commercial letters are priced flat right up to it.

The weight cliff is the one that catches people, because the approach to it is flat. Commercial First-Class letters cost the same from one ounce all the way to three and a half. A mailer adds an insert, adds another, sees no change in the quote, adds a third, and at one tenth of an ounce over the line the piece reprices as a four-ounce flat. The published tier we used goes from about 62 cents to about $1.90.

The aspect ratio produces the single best line in the price schedule. In the advertising mail class, the footnote says that non-machinable letter prices “will be same as corresponding Nonautomation Flats”. A square envelope, the one a designer reaches for when the piece should feel like an invitation, costs what a magazine costs.

There is also a rigidity test that catches anything with a stiffener: the piece must bend when subjected to forty pounds of tension around an eleven-inch turn. Plastic cards, rigid inserts and chipboard backing fail it.

What it takes to get commercial prices at all

Retail prices are for one letter. Commercial prices require volume, preparation and paperwork, and the fixed costs matter more than the per-piece saving for a small B2B mailer.

Minimum volumes: five hundred pieces for presorted First-Class letters, and two hundred pieces or fifty pounds for advertising mail. Both classes must meet their minimum separately.

Fixed costs: a one-time permit application fee, plus an annual mailing fee charged per class and per office of mailing. A company mailing both classes from one location pays the annual fee twice.

Software: automation prices require addresses coded with certified matching software, which is a certification of the software rather than of you, and a precondition for every automation price in the schedule.

Requirements for reaching commercial postal pricesTable of the requirements a business must satisfy to obtain commercial rather than retail postal prices in the United States for mailing letters to other businesses. Presorted first class letters require a minimum mailing of five hundred pieces, and advertising mail requires a minimum of two hundred pieces or fifty pounds of pieces, with each class required to meet its own minimum separately. A one time permit application fee applies, and an annual mailing fee is charged for each class and for each office of mailing, so a company mailing both classes from a single facility pays the annual fee twice in addition to the application fee. Automation prices require that addresses be matched and coded using software certified under the postal accuracy programme, which certifies the software rather than the mailer and is a precondition for every automation price in the schedule. On first class automation letters, moving from the coarsest presort tier to the finest saves a little under nine cents per piece, and the finest automation tier is roughly a quarter cheaper than a stamped retail letter. The table concludes that for a small business to business mailer the fixed annual costs and the certified software requirement, rather than the per piece saving, are usually what determine whether to hold a permit or to work through a mailing house.The price of the cheaper priceMinimum volume500 pieces First-Class, 200 pieces or 50 lb advertising mailFixed costPermit application, then an annual fee per class and per officeSoftwareCertified address coding, required for every automation priceWhat the sort is worthJust under 9 cents from coarsest to finest tierFor a small mailer, the fixed costs and the software decide the question, not the per-piece saving.
The per-piece saving is real. For a small mailer the fixed costs and the software requirement usually decide whether it is worth holding a permit at all.

What the presort tiers are worth: on First-Class automation letters, moving from the coarsest sort to the finest saves a little under nine cents a piece. Against a stamped retail letter the automation five-digit price is roughly a quarter cheaper. Those are real savings, and they are also why a mailing house exists: sorting to the finest tier requires software, volume and a deposit process most small mailers will not build.

Address quality, and the asymmetry that hides your errors

Commercial mail carries an obligation to update addresses within ninety-five days of mailing. Above a monthly error threshold, an assessment is charged per piece and the result posts to your mailer scorecard.

The licence for the national change-of-address dataset is priced for service bureaux, not for a company mailing five thousand pieces a quarter. That single fact, more than any other, is the answer to whether you need a mailing partner.

Then there is the asymmetry nobody mentions when recommending the cheaper class. An undeliverable First-Class piece is forwarded or returned with a reason, at no charge, by default. An undeliverable advertising mail piece is disposed of: no charge, no notice, no information. The cheaper class silently destroys the feedback loop you would use to clean your list.

What happens to an undeliverable piece, by mail classComparison of the default treatment of an undeliverable mail piece in the two commercial mail classes used for business prospecting in the United States, and of the endorsements that change that treatment. By default and with no endorsement applied, an undeliverable first class piece is treated as though forwarding had been requested: it is forwarded at no charge where a change of address is on file, and otherwise returned to the sender with the reason for non delivery, again at no charge. By default and with no endorsement applied, an undeliverable advertising mail piece is disposed of by the postal service, with no charge and no notice to the sender, so the sender never learns that the address was wrong. The comparison notes that advertising mail is the cheaper class per piece and that this default silently destroys the feedback loop a mailer would otherwise use to clean the list. Three endorsements change the default. Address service requested causes the piece to be forwarded with a separate notice sent to the mailer, or returned with the new address, with a fee in the advertising mail class. Return service requested causes the piece always to be returned with the new address or the reason, free in first class and at first class return postage in advertising mail. Change service requested causes a notice to be sent to the mailer while the piece itself is destroyed, which is the cheapest way to clean a list without paying to receive paper back. Fees range from a few cents for automated electronic corrections to about one dollar for a manual notice, and mailers using the full service barcode option receive automated address corrections at no charge.What a bad address tells you, by classFirst-Class, by defaultForwarded if a change is on file,otherwise returned with the reason.No charge. You learn something.Advertising mail, by defaultDisposed of. No charge,and no notice.Cheaper per piece, and silent.Return the pieceAlways returned with the new address or the reasonDestroy it, send a noticeCheapest way to clean a list without paying for paper backAutomated electronic corrections cost a few cents, a manual notice about a dollar, and full-service barcode mailers pay nothing.
The cheaper class destroys the piece without telling you. The information can be bought back, and is free to full-service mailers.

You can buy the information back. Endorsements let you choose between having the piece returned, having a notice sent while the piece is destroyed, or having the address corrected electronically. Fees run from a couple of cents for automated electronic corrections to a dollar for a manual notice, and mailers running the full-service barcode option receive automated corrections at no charge.

That is the practical route for a B2B mailer: run full-service, take the small per-piece discount, take the free address corrections, and let the update obligation take care of itself.

There is no Do Not Mail list

The question comes up in every discussion of this channel, and the answer is short: the United States has no legal equivalent of the Do Not Call Registry for postal advertising. No statute, no federal register, no enforcement agency.

A voluntary opt-out exists. It is operated by the trade association of the advertising industry, which acquired the direct marketing association and folded it in as a division. It charges the consumer a small administrative fee for a ten-year registration, it describes its scope as promotional mail from companies the registrant has no relationship with, and it carries its own disclaimer that it will not eliminate all promotional mail. It binds only mailers who choose to license and apply the suppression file.

It is also presented as a consumer service. We could not establish that it covers business entities as such, and we are not going to assert that it does.

One legally binding postal opt-out does exist in US law, and it is worth a sentence precisely because of how narrow it is. Any addressee who decides, in their sole discretion, that a mailing is sexually provocative can obtain an order requiring the sender to stop within thirty days and to delete their name from all mailing lists, enforceable in federal court. It is a real, court-backed, list-deletion right. It is also useless as a general marketing opt-out.

What does apply to a B2B postal list is state privacy law, and the relevant one reaches further than most people think. California’s regime lost its business-to-business exemption at the end of 2022. Two qualifications keep most small mailers out of scope anyway: it protects natural persons, so a list of company names and street addresses with no named human is not personal information, and it applies only to businesses above revenue or data-volume thresholds that a small B2B mailer will not meet. A list of named individuals at those companies is a different matter.

Where a business mailing address comes from

Unlike phone numbers and email addresses, postal addresses genuinely are in the public record, and in at least one state they are free, complete and unrestricted.

We queried one state’s full business entity register live: over three million rows, published under a public domain licence, including principal and mailing addresses and full registered agent details. No reuse restriction of any kind.

Two warnings before anyone builds a list from it.

This does not generalise. There is no national business register in the United States. Each state runs its own with its own terms, and several of the largest publish no open bulk file at all. “Check your state’s open data portal” is the honest advice, and it has to be checked state by state.

The register is not a list of operating businesses. It is an unfiltered record going back to the nineteenth century in the case we examined, and the first sample row we pulled was a delinquent entity. Filtering on status is the minimum, and even then you have a register of legal entities rather than of businesses at deliverable addresses.

The federal contractor registry gives physical and mailing addresses in its public tier, but the universe is limited to entities that registered to do business with the federal government, and it carries a flag letting entities opt out of public display. Honour it.

And the richest business data the federal government holds, the statistical business register, is the one source that is forbidden by statute from ever becoming a mailing list.

Measurement, and the one place paper has an advantage

The barcode that unlocks automation pricing also unlocks tracking and free address corrections. That is the measurement stack, and its entry price is compliance rather than money. A free programme also lets a mailer attach a colour image and a clickable link to the daily preview email that subscribers receive, though the postal service reserves the right to monetise parts of it later.

The more interesting point is legal, and it runs the opposite way to email.

In the European Union, measuring whether a marketing email was opened now generally requires prior consent, because the pixel accesses the recipient’s device. The United States has no equivalent: there is no federal statute requiring consent before storing or reading information on a user’s terminal equipment, and no cookie-consent law.

What does regulate tracking in the US is the wiretap family of statutes, and their scope is communications transmitted through wires, radio or electronic systems. Printed matter is outside them entirely. A code printed on paper intercepts nothing and transmits nothing, and nothing happens until the recipient picks up a phone and scans it, which is an act they initiate.

Measuring a printed piece against measuring an emailComparison of the legal basis for measuring responses to a marketing email against measuring responses to a printed mail piece in the United States. Measuring an email is regulated through the wiretap family of statutes, whose federal provision prohibits intentionally intercepting any wire, oral or electronic communication and whose scope is limited to communications transmitted through wires, radio or electronic systems. The theory applied to email tracking pixels under those statutes and their state analogues requires an interception of an electronic communication that the recipient did not initiate, and a pixel qualifies because it fires automatically when the message renders, without any act by the recipient. A printed code inverts each element: it is ink, so printing it intercepts and transmits nothing; nothing occurs unless the recipient affirmatively picks up a device and scans it, which is an act the recipient initiates and which defeats the interception premise; and tracking begins only once the recipient reaches the advertiser website. The comparison notes that the United States has no equivalent of the European privacy directive, no federal statute requiring consent before storing or accessing information on a user terminal equipment, and no cookie consent law, and that the federal commercial email statute applies to electronic mail only and does not require prior permission even there. It concludes that the consent question does not attach to the printed piece at all but to the landing page, where it attaches identically whether the visitor arrived from a postcard, a billboard or a search result.Paper is outside the statute; the page is notMeasuring an emailWiretap statutes and their state analoguesReaches communications carriedby wire or radioThe pixel fires on render, with noact by the recipientThat passivity is the whole claim.Measuring a printed pieceNo US equivalent of the privacy directiveInk intercepts nothing andtransmits nothingNothing happens until the recipientpicks up a device and scansThe question moves to your website.
The wiretap statutes reach communications carried by wire or radio. Paper is outside them, and the question moves to the landing page.

So the consent question does not attach to the printed piece. It attaches to the landing page, and it attaches there identically whether the visitor arrived from a postcard, a billboard or a search result. Design the page as you would any other, and stop worrying about the paper.

The benchmark problem, in its purest form

We looked for a neutral source of direct mail response rates. There is none, and the reasons are structural.

The postal service publishes volume, revenue and weight. It has no mechanism to observe a response, because it never sees the advertiser’s conversion. The regulator publishes financial and volume analysis. The postal inspector general has done genuine independent research, but on neuromarketing, comparing emotional engagement and recall between physical and digital advertising, not on campaign outcomes.

What the headline direct mail response rate rests onAnalysis of the provenance of the direct mail response rate figure that circulates as an industry benchmark in the United States. The figure comes from a response rate report produced by the trade association of the advertising industry together with a commercial research partner, and promoted to association members, previous respondents and the research partner community, with participants offered a summary of results as an incentive. The current edition sits behind a member login, so its methodology cannot be verified publicly. In a publicly mirrored earlier edition that could be read in full, the survey was fielded online over roughly four months, collected four hundred and thirteen responses of which ninety two were complete and three hundred and twenty one partial, and retained two hundred and fifty as qualified and complete enough for inclusion. The association states in the report that a truly representative sample would be cost prohibitive to obtain given the number of organisations practising direct marketing, and therefore that the data should not be considered benchmark data but used for informational purposes only, a statement it makes twice. The headline direct mail response rate for house files of fifteen point six per cent rests on a base of twenty six respondents, and the prospect file figure of ten point eight per cent on a base of twenty five, flagged in the report itself as a small base. A question added to that edition asked whether answers were based on actual metrics, best estimates or a mix, and the share choosing actual metrics ranged from five to fifty per cent across questions with an average of twenty one per cent. The analysis concludes that the defensible claim is not any particular response rate but the structural fact that the only benchmark for the channel is produced, priced and gated by the channel trade association with no public methodology and no neutral alternative.The benchmark, taken apartWho publishes itThe trade association of the channel, with a research partnerWho can read the methodMembers only. The current edition sits behind a loginSample for the headline rate26 respondentsReading from actual metricsabout one in five, on averageWhat the publisher says about it“should not be considered benchmark data, but used forinformational purposes only”. Printed twice in the report.
A 15.6 per cent response rate drawn from 26 self-selected respondents, most of whom were estimating, in a report whose publisher says it is not benchmark data.

Everything traces to one report, produced by the trade association of the channel. The current edition is behind a member login, so its methodology cannot be checked. An earlier edition that can be read in full is disarmingly honest about itself: the association writes that a truly representative sample “would be cost-prohibitive to obtain”, and prints twice that the data “should not be considered benchmark data”.

The numbers themselves are thinner than the reputation. The headline house-file response rate rests on twenty-six respondents. And the report added a question asking whether answers came from actual metrics or estimates: the share choosing actual metrics “ranged from 5 percent to 50 percent, with an average of 21 percent”.

So the defensible sentence is not a response rate at all. It is this: direct mail is the only major acquisition channel in the United States whose headline performance benchmark is produced, priced and gated by its own trade association, with no public methodology and no neutral alternative. That claim is verifiable. The 15.6 per cent is not.

Two figures to refuse outright if a supplier offers them: claims of 80 to 90 per cent response rates, and a 112 per cent return figure, both attributed loosely to the postal inspector general’s research. They correspond to nothing in it.

What the volumes say, and what mailers are already doing about the grid

The regulator’s own analysis records that First-Class volume fell by about a third over the last decade, and advertising mail by about thirty per cent, while noting the postal service’s own characterisation of advertising mail as generally resilient.

One detail in that analysis is worth more than the trend. Within the most recent year, the regulator notes migration of flats into letters, attributed to a change of marketing strategy or “an effort to reduce overall postage costs”.

That is mailers reshaping their pieces to stay on the cheap side of the cliffs in the first section of this article. It is the market confirming that the grid, not the headline price, is where the money is.

What to do with this

Fix the physical specification before you write a word of copy. Decide the envelope, the paper weight and the number of inserts against the three thresholds, then design inside that envelope. A campaign that respects the letter limits and a campaign that does not are different budgets, not different quotes.

Weigh a finished sample, assembled, on a scale. Not a mock-up, not the printer’s estimate. The tenth of an ounce that reprices the job is exactly the kind of thing an estimate rounds away.

Run the full-service barcode option even at small volumes, because it is what turns the address update obligation from a cost into free information.

And if you need a response rate to justify the spend, do not go looking for one. Run a small first wave with a measurable response path, and let it produce your number. You will have a sample of one campaign, which is a smaller sample than the industry benchmark and considerably more relevant to you.

The related pieces are running a cold email campaign and cold calling rules and reality.