Why Custom Greeting Cards Still Matter for Businesses

Everybody went digital at once. Email, ads, social, notifications — all of it competing for the same few inches of attention. And while that was happening, something quieter went on in the background that almost nobody planned for.

The mailbox emptied out.

That is not a figure of speech and it is not nostalgia. The Postal Service measures it. The average American household received 637.3 pieces of mail in 2025, down from 1,040.6 in 2015 — a 39% decline in ten years (USPS Household Mail Survey FY2025, administered by NORC at the University of Chicago). Over the same decade the number of greeting, holiday and birthday cards households sent fell from 17.8 a year to 11.0. Down 38%.

So here is the situation. The inbox got louder every year. The mailbox got quieter every year. And the thing most businesses stopped doing is the thing that now has the least competition.

I have been in the personalized card and printing trade since 1984. Best Holiday Cards has been publishing since 2003. I have watched three or four supposed deaths of print from close enough to see they were exaggerated, and I have also watched plenty of businesses waste money on cards sent badly. Both things are true, and I am going to try to be straight with you about where the line is.

First, about the numbers in this article

This is the part I want to get out of the way before anything else, because it changes how you should read the rest.

I had every statistic in this article checked against its original source. Not against an article citing the source — against the study itself, the actual filing, the actual PDF.

About two-thirds of what I started with did not survive.

The industry runs on numbers that have been passed hand to hand for so long that nobody can find where they came from. A response-rate comparison everybody quotes as a 2025 figure turns out to be from a 2012 report, and the 2012 report said something different anyway. A famous “$42 back for every dollar” return figure is email’s number, from the UK, in pounds, from 2019 — quoted constantly by people arguing that mail beats email. A retention statistic attributed to a major accounting body appears in no document that body has ever published. One insurance figure I chased for an hour turned out to trace back to an uncredited reprint of a February 1998 magazine article.

I cut all of it. What is left is smaller, less exciting, and true. Every number below names its source and its year in the sentence. Where the organization that produced the number sells something that benefits from the answer, I say so.

You are welcome to check any of it. I would rather you did.

The ten reasons, at a glance

1. Keeping a client costs less than chasing a new one — with the honest version of a famous statistic
2. Your mail gets read. Your email gets skimmed in under a minute.
3. The people who already know you read twice as much of what you send — the single best number in this article
4. When the job ends, you disappear
5. Reaching out lands better than you expect it to — and that gap has been measured
6. Referrals come from the people who remember you
7. Almost nobody sends cards anymore
8. A thing you can hold is valued more than the same thing on a screen
9. Business communication is measurably getting less personal
10. What it actually costs — the whole bill, nothing left off

Then two sections most articles like this leave out: what the evidence does not say, and what I would do if I were starting from nothing.

1. Keeping a client costs less than chasing a new one

You already know this one. Everybody does. It is worth going through anyway, because the version of it in circulation is wrong in a way that matters.

The famous claim is that a 5% lift in retention grows profit by 25% to 95%, and that Bain & Company found it. Two corrections. The real source is Reichheld and Sasser, “Zero Defections,” Harvard Business Review, September 1990, and the range they reported was 25% to 85%. The 95% upper bound exists in no primary source I could find. It appears to have been rounded up by somebody, once, and then copied forever.

Second, Frederick Reichheld is a Bain fellow and Bain sells loyalty consulting. That does not make the finding wrong. It does mean you should know it, and it means the paper is 36 years old.

Now the part almost nobody quotes. Twelve years later the same journal published Reinartz and Kumar, “The Mismanagement of Customer Loyalty” (HBR, July 2002), which went looking for the loyalty–profitability link in four companies’ transaction data and found it weak. Their conclusion was blunt: about half of loyal customers barely generate a profit.

I am including that because leaving it out is how you end up with an article that falls apart the first time somebody who has read the literature gets hold of it. And because the narrower claim survives it intact:

Here is what acquisition actually costs, from the best-sourced number I could find for a relationship business. The Kitces Research Marketing Survey 2022 (457 usable responses, fielded April–May 2022) put the median cost of acquiring one financial-advisory client at $2,167. It ranged from $719 at the smallest practices to $7,037 at the largest. Growth gets more expensive as you scale. Retention does not.

Set a card against $2,167. At the volumes a small firm sends, a card runs about $2.51 including the envelope, and about $3.33 with the stamp on it. I will show you exactly where those come from in reason ten, and I will show you the whole bill rather than the flattering part of it.

What I will not do is tell you the card is why they stayed. Nobody can tell you that. There is no study that isolates a greeting card from everything else a business does, and any vendor who claims otherwise is selling you something. What I can tell you is that the attempt costs about three dollars and the replacement costs four figures, and that in forty years I have never once seen a business regret the three dollars.

2. Your mail gets read. Your email gets skimmed in under a minute.

I need to start this one by refusing to give you the statistic you came for.

You have seen the comparison: physical mail gets a 90% open rate, email gets 20%. It is on a hundred vendor websites. I am not going to publish it, because the email half of it has been unmeasurable since 2021.

That is the year Apple shipped Mail Privacy Protection, which pre-loads tracking pixels whether or not a human opens the message. Open rates for any list with a meaningful share of Apple Mail users have been inflated and unreliable ever since. Every email open rate you have been shown in the last five years is measuring something other than what it claims to measure. Anyone still running that comparison either does not know this or is counting on you not to.

There is also no such thing as an open rate for paper. Nothing tracks it. Nobody knows.

So here is what can actually be said, and it is enough.

What the Postal Service measures

The USPS Household Mail Survey FY2025 is the good one. It is prepared by RCF Economic & Financial Consulting, administered by NORC at the University of Chicago, built on address-based sampling weighted to the Census: roughly 12,000 recruited households and about 5,000 completed one-week mail diaries, fielded October 2024 through September 2025 and filed with the Postal Regulatory Commission. USPS funds it, which you should know. NORC’s involvement is why I trust the method anyway.

Households reported that one or more members read 51% of the advertising mail they received in 2025, up from 46% in 2020 and 47% in 2015. Some 61% of advertising pieces got at least some attention.

Read that trend again, because it is the opposite of what everybody assumes. Attention to mail is going up. USPS’s own explanation is the obvious one: the decline in volume has allowed each individual piece to gain more attention.

Less mail arriving means more of it gets read. That is the whole thesis of this article in one sentence, and it comes from the government’s own data rather than from me.

One honest caveat: those are self-reported diary entries. Write “households report reading,” not “51% of mail is read.” I have seen that distinction blurred so often it is practically the house style of the direct-mail industry.

What is happening on the other side

The Microsoft Work Trend Index 2025 is unusual and useful because most of it is anonymized Microsoft 365 telemetry — observed behavior, not people describing their own habits — alongside a survey of 31,000 knowledge workers in 31 markets fielded in early 2025.

Their finding: the average worker receives 117 emails a day, most of them skimmed in under 60 seconds.

Microsoft sells Copilot as the answer to exactly this problem, so disclose that. They have no stake whatsoever in paper.

Put the two side by side and you do not need a contested open-rate comparison. 117 messages a day, skimmed in under a minute each. Against a mailbox that holds less than half of what it held ten years ago, where attention per piece is rising.

You are not choosing between two channels of equal difficulty. You are choosing between a crowded one and an empty one.

3. The people who already know you read twice as much of what you send

If you read one section of this article, make it this one. This is the best number I have, and I have never seen another card company use it.

It is from the same USPS Household Mail Survey FY2025 described above. Here is the finding in the report’s own words:

“Advertising mail sent to a recipient with whom the business had a prior relationship (e.g., a previous or existing customer) was around twice as likely to be read, 64% of pieces vs. 36%, than mail sent when there was no past business relationship.”

Two caveats, stated plainly. Reading is self-reported in a mail diary. And that second column is stated likelihood to respond on a 1–10 scale, not an actual response rate. Blurring those two is exactly what I said I would not do. Even read strictly, the gap is enormous: a roughly eightfold difference in stated intent, from a federal survey with disclosed methodology.

Now think about what that means for the mail you send.

Every dollar most businesses spend on mail goes to strangers. Prospect lists, bought lists, saturation routes, every door direct. The 36% column. Meanwhile the 64% column — the people who have already paid you money, who know your name, whose address you already have — usually gets nothing at all.

The most responsive audience you will ever mail to is the one sitting in your accounting software right now. It costs you nothing to acquire. You already own the list. And in most businesses it is the list nobody mails.

That is the whole case for a client card program, and it is a federal statistic rather than my opinion.

4. When the job ends, you disappear

Here is the pattern I have watched for four decades. You do good work. The job wraps up. Everybody is happy. And then nothing happens for two years, and by the time you think about that client again they are somebody else’s.

They did not leave angry. That is the part that catches people out. They just stopped thinking about you, and nobody told you it was happening.

The intellectual version of this is old and good: Jones and Sasser, “Why Satisfied Customers Defect,” HBR, November 1995. Their line was that the gulf between satisfied customers and completely satisfied customers can swallow a business, and the Xerox data behind it showed totally satisfied customers were about six times likelier to repurchase than merely satisfied ones. It is a 1995 argument and should be labeled as one. It has held up.

What is newer is that three different professions have now measured their own version of it, and the numbers are striking.

Financial advice

Morningstar, “Why Do Investors Fire Their Financial Advisor?” surveyed 3,003 participants in 2021–22, with a firing subsample of 185 and reported inter-rater reliability. Quality of the relationship was the second most common reason for firing an advisor, at 21%, behind quality of advice at 32%.

The sentence from that report I keep coming back to:

“individuals who cited Communication as their firing reason indicated that since they haven’t heard from their advisor, the advisor must not be doing anything.”

Silence is not neutral. The client fills it in, and they do not fill it in generously.

Real estate

Both halves of this come from the same survey — the NAR 2025 Profile of Home Buyers and Sellers, mailed to 173,250 recent buyers in July 2025, 6,103 responses, margin of error ±1.25%.

91% of home buyers say they would use their agent again or recommend them. Among repeat buyers — the only people in the survey who actually had the chance — 18% did.

That gap is not a service problem. Ninety-one percent were happy. It is a memory problem, and it is the clearest measurement of “satisfied but forgotten” I have ever seen published.

The same survey adds a detail that should worry anybody in a referral business: 38% of buyers have not recommended their agent even once since the purchase, and the median number of recommendations is one.

Law

The Clio 2025 Legal Trends Report asked 1,000 US adults through a market panel about their experience hiring a lawyer. 24% of clients who recently hired one say they would be unlikely to rehire them. A quarter of the clients a firm just served are gone, in the profession’s own research.

And in B2B generally, McKinsey’s B2B Pulse 2024 — about 4,000 decision makers across 34 sectors and 13 countries — found 51% would find an alternative supplier if they were not connected to the right person for help. Half your customer base is one unanswered relationship away from shopping around.

Four professions, four different research houses, one finding. The work was fine. The relationship went quiet. A card is the cheapest thing in existence that makes a relationship not quiet.

5. Reaching out lands better than you expect it to

This is the section that replaced the one I threw away.

The draft of this article had a passage about how roughly 70% of buying decisions are emotional, attributed to Gallup. I went looking for the study. It is not a study — it is a house principle Gallup calls its 70/30 Principle, with no published sample, no method, and no fielding dates. Out it went.

What replaced it is better, and it is the most interesting research in this entire article, because it is not about buying at all. It is about why you do not send the card.

The gap between what you expect and what actually happens

Kumar and Epley, “Undervaluing Gratitude,” Psychological Science 29(9), 2018. They had people write genuine gratitude letters to real people in their lives — 107 writers, 80 recipients, plus two further experiments. Before sending, writers predicted how the recipient would feel. Then the researchers asked the recipients.

People systematically underestimate how good it feels to receive one of these, and systematically overestimate how awkward it will be. Both errors point the same direction, and both of them are reasons not to send.

Then there is Liu, Rim, Min and Min, “The Surprise of Reaching Out,” Journal of Personality and Social Psychology 124(4), 2023 — pre-registered, and independently replicated at n=742, which puts it in better standing than most of what gets quoted in business writing. Their conclusion:

“we document a robust underestimation of how much other people appreciate being reached out to” … and the effect is magnified “when it occurs between more socially distant (vs. socially close) others.”

Read that second half twice. The effect is strongest between people who are not close.

That is not your spouse. That is an advisor and a client. An agent and the couple whose house closed three years ago. A CPA and the business owner they see for six weeks a year. The exact relationships this whole trade exists to serve are the ones where the surprise lands hardest.

And surprise is the mechanism the researchers identify. Which is the honest case for a card that arrives for no reason in March over a scheduled email nobody registers as a decision anybody made.

What this actually means for you

Every reason you have not sent a card is a forecasting error, and it has now been measured twice with clean methodology and no commercial sponsor.

You think it will be awkward. It will not be; the recipients scored it at 1.95 out of 7.

You think it will land as a small thing. It lands bigger than you expect.

You think they will read it as a sales move. The research on gratitude expression says otherwise — and a card that asks for nothing is not a sales move.

I have watched people talk themselves out of this for forty years. Now I can tell them there are two peer-reviewed papers saying the thing they are worried about does not happen.

6. Referrals come from the people who remember you

Referred business is the cheapest business there is. No lead cost, no cold call, and the prospect arrives already trusting you because somebody they trust vouched.

How much of a relationship business actually runs on it:

• 54.2% of financial advisors’ new clients come from referrals, and a further 13.9% from centers of influence — CPAs, attorneys and the like (Cerulli Associates, U.S. Advisor Metrics 2025, released February 2026). One new client in seven arrives through a professional relationship rather than a client one.
• 67% of new clients and new client assets at RIA firms come from referrals from clients and centers of influence combined (Schwab 2024 RIA Benchmarking Study, 1,304 firms, $2 trillion in assets, fielded January–March 2024). Note the “and” — that figure gets misquoted as client referrals alone constantly.
• In real estate, the median agent gets 28% of business from repeat past clients and 22% from past-client referrals (NAR 2026 Member Profile). For agents with 16 or more years in, it is 49% and 32%. For agents with two years or less, referrals are 0%.

That last line is the whole argument in one number. Half a veteran agent’s business comes from people they already sold to. The new agent has none of it, not because they are worse, but because they have not had time to accumulate anybody to be remembered by.

The thing referrals need is not satisfaction. Satisfaction is already there — remember the 91%. What they need is recall at the exact moment somebody asks the question. And that moment is not scheduled. It happens in a kitchen, at a soccer game, in a hallway, and it lasts about four seconds.

You either come to mind in those four seconds or you do not.

Two more figures worth having. 86% of consumers say they would recommend a brand they feel loyal to (KPMG International, The Truth About Customer Loyalty, 2019 — 18,520 consumers across more than 20 countries; seven years old, so date it). And Bain’s work on Net Promoter found that promoters make roughly seven times the referrals of detractors. You may have seen “loyal customers are 4x more likely to refer” somewhere. That one is a mashup of two unrelated claims and I could not source it. The real figure is larger.

And when somebody does send you a referral, send them a card about it. Not an email. That is the highest-return card in this entire article and it costs three dollars.

7. Almost nobody sends cards anymore

I could tell you the corporate gifting market is worth some enormous number. Every article in this category does. I went looking for a market-sizing figure I could stand behind and could not find one — every estimate traces back to a gifting vendor or a commissioned study paid for by one.

So here is a better argument anyway, because scarcity beats market size.

From the USPS Household Mail Survey FY2025, the same NORC-administered survey as before:

Two-fifths of the mail is gone. Two-fifths of the cards are gone. And attention per piece is going up, which USPS attributes directly to the volume decline.

When I started in 1984, a card from a vendor was table stakes. Everybody sent one, which meant a card bought you almost nothing except not being the firm that forgot. A December desk had thirty of them on it and yours was the eleventh.

That is not the situation now. A business card display on a desk in December today has maybe four cards on it. Some years, one.

Being the one is a completely different proposition than being the eleventh. Nobody planned this for you. Your competitors quit, and they took the crowding with them.

That is the honest version of “stand out.” Not that a card is magic. That the field emptied while nobody was looking.

8. A thing you can hold is valued more than the same thing on a screen

The draft of this article claimed the average piece of mail stays in the home for 17 days, and that more than 70% of people say paper stock shapes how they see a brand. I cut both. Neither has a source. The 17 days appears in dozens of vendor pages and in no study. The 70% has no origin at all.

I am sorry to lose them, because I believe the underlying point. Here is what can actually be supported.

Atasoy and Morewedge, “Digital Goods Are Valued Less Than Physical Goods,” Journal of Consumer Research 44(6), 2018. Five experiments, no commercial sponsor. The field test is the one I like: tourists were offered a photograph of themselves and paid a median of $3.00 for the physical print versus $1.00 for the identical digital file. Same image. Three times the price for the version you can hold.

Then Venkatraman et al., “Relative Effectiveness of Print and Digital Advertising: A Memory Perspective,” Journal of Marketing Research 58(5), 2021 — fMRI plus eye-tracking, finding greater activation in hippocampal and parahippocampal regions for print advertising than for digital. Those are memory structures.

Disclosure, because it matters: that study was part-funded by a grant from the USPS Office of Inspector General. I am telling you because the first thing a skeptical reader should ask about a print-beats-digital brain study is who paid for it. Peer review at JMR is a real firewall and I think the paper stands, but you should weigh it knowing that.

Put those together and the claim is modest and defensible: physical things are valued more and remembered better than their identical digital versions. Not 17 days on a mantelpiece. Not 70% of people. Just that.

What I can add is forty years of watching where cards end up, which is not research and I will not dress it up as research. They go on the credenza. They go on the corkboard next to the desk. They stay up past the season, sometimes well past it, in the room where somebody decides who to call. I cannot give you a number for that. Nobody can. But it is the reason I have spent my working life on this and not on something easier to sell.

9. Business communication is measurably getting less personal

This is the finding I did not expect to have and would not have thought to look for.

From that same Microsoft Work Trend Index 2025 telemetry — observed data, not opinions:

“Mass emails with 20+ recipients are up 7% in the past year, while one-on-one threads are on the decline (-5%).”

Business communication is shifting away from one person writing to one person, and it is doing so at a measurable rate, right now. Not as a trend piece. As telemetry.

Every one of those 20-plus-recipient emails teaches the recipient a little more firmly that nothing arriving in their inbox was written to them specifically. That is a lesson your genuinely personal email pays for too.

And then there is what has happened to trust in the channel itself. Pew Research Center, July 2025, “Online Scams and Attacks in America Today” — 9,397 US adults on the American Trends Panel, fielded 14–20 April 2025, margin of error ±1.3 points. No commercial interest of any kind:

• 63% receive scam emails at least weekly. 28% get them daily.
• 73% have experienced an online scam or attack.

Nearly three-quarters of the country has been burned by something that arrived digitally. Most of them are getting fresh reminders every week.

The inbox is now a channel people have been actively trained to distrust. Nobody has ever been trained to distrust an envelope. An unexpected email from a company is a thing you inspect for danger. An unexpected card from a company is a thing you open.

That asymmetry did not exist in 2005 and it is not going away.

What personalization actually buys you

Two more pieces, both flagged for what they are.

The NAR 2025 survey asked buyers what forms of contact they value from an agent. Personal phone calls 72%, text messages 72%, social media 15%, email newsletters 7% (Exhibit 4-9). Personal beats broadcast by roughly ten to one. NAR does not test physical mail, so I am arguing by category here rather than claiming NAR measured cards — but the split between one-to-one and one-to-many is about as clean as survey data gets.

Lob’s 2025 consumer survey (two thousand US adults, method disclosed) found 52% more likely to engage with mail tailored to them and 67% who acted on a personalized detail. Lob sells direct-mail software, so that is a vendor number and you should treat it as one. I am including it because the method is published, which is more than most.

You may have seen that adding a name to a mail piece lifts response 135%, and that 64% of customers will spend more with a brand that remembers them. I cut both. The 135% has no primary source. And the 64% figure says something different in the original: 64% prefer a brand that remembers them; only 34% said they would spend more. If you want a spending figure, Medallia’s 2024 research puts it at 61% willing to spend more for a customized experience — use that one.

10. What it actually costs — the whole bill, nothing left off

Every article like this ends with a return-on-investment number. The one I inherited said direct mail returns 161%, or about $42 back for every dollar spent.

That $42 figure is email’s. It is from the UK Data & Marketing Association in 2019, and it is denominated in pounds. It has been quoted for six years by people arguing that mail beats email, which should tell you how carefully any of this gets checked.

So I am not going to give you an ROI number. There is no credible study of the return on a business greeting card program, and anybody who hands you one is making it up. Too many variables, no control group, and every figure in circulation comes from a company that sells mail.

What I can do instead is something no vendor in this category will do. I can show you the entire bill.

250 cards, priced off our own cart

A mid-tier design, 250 cards, with envelopes and your firm name imprinted. This is what the cart returns, at the early-order discount that was in effect when I priced it in September 2026:

That is $2.51 a card, delivered to your door.

It is not what it costs to send them. Two things are still missing, and this is where most comparisons quietly stop.

The two costs nobody puts in the table

Postage. At First-Class rates, 250 cards is $205.00, or 82¢ a card. Postage is a constant. It is the same whether you do everything yourself or hand us the whole job. Any comparison that leaves it out of one column and includes it in another is not a comparison.

Your time. Here is what 250 cards costs in hours, from watching people do it for forty years:

Four and a half hours. Call it two evenings, or one Saturday you were not planning to spend that way. Everybody underestimates the stamps.

The three routes, side by side

Postage sits in its own column in every row, because it is in every row:

Our full mailing service is quoted at $350 per 100 — $875.00 at 250 — and that price includes the postage. I have split it out above so you can see the labor separately: $670 of service and $205 of stamps. If you have seen our mailing service priced against a do-it-yourself column with no postage line in it, that comparison was flattering us by $205.

Now the only question that actually matters:

Read that last row before you buy the top tier. The first 2.8 hours are cheap to buy back at $67 an hour. The last 1.7 hours cost $284 an hour, because that is the stretch where somebody has to sign, stuff, seal and stamp 250 individual pieces, and there is no machine that does the signing.

My honest advice, and it costs me money to give it: most firms should buy the addressing and keep the signing. You get 62% of the hours back for 28% of the price, and you keep the part of the job that is the entire point — a real signature and two lines in your own handwriting. The full mailing service is the right answer when nobody in the office can be trusted with a deadline in the second week of December, which is a real situation and not a rare one.

What none of this tells you is whether it is worth it. That depends on what a client is worth to you, and you know that number better than I ever will. Set $831.66 against it and decide. I would rather you made that decision with the whole bill in front of you than with a statistic I could not source.

What the evidence does not say

I would rather you heard this from me than found it out later.

Nothing here proves a card caused anything. Every study above measures something adjacent — how mail is read, how gratitude lands, why clients leave, where referrals come from. None of it isolates a greeting card from the twenty other things a business does. There is no randomized trial of holiday cards and there never will be, because nobody will fund it and no firm will withhold cards from half its client list for a year to find out.

The loyalty literature argues with itself. I gave you Reichheld and Sasser in reason one and I also gave you Reinartz and Kumar, who went looking for the profit in loyal customers and found much less of it than the consulting industry claims. If you read only the first one you would come away more confident than the evidence supports.

The USPS reading figures are self-reported. People writing in a mail diary at the end of a week are not a camera. The figures are consistent across a decade and the sampling is sound, which is why I use them, but they are what people say they did.

Several of the best sources have a stake. USPS funds the household survey. Microsoft sells the cure for email overload. Bain sells loyalty consulting. Lob sells direct mail. I have flagged each one in the sentence where it appears, which is the most I can do about it. The cleanest sources in this article — Pew, the two psychology papers, the Journal of Consumer Research study, NAR, Morningstar, Cerulli, Clio — are the ones I would lean on hardest, and none of them is about greeting cards.

And the thing I believe most is the thing I can prove least. I think a card stays on a desk for weeks and does quiet work the entire time. I have watched it for forty years. There is no study. I am not going to invent a number for it, and you should be suspicious of anyone who hands you one.

What survives all of that is still a reasonable case. The mailbox is emptier than it has been in decades and attention per piece is rising. Mail to people who already know you is read at nearly twice the rate of mail to strangers. Clients who are perfectly satisfied leave anyway when the relationship goes quiet, in four different professions’ own research. And reaching out lands better than the person reaching out expects it to, especially at a distance.

That is not a guarantee. It is a good bet at three dollars.

What I would do if I were starting from nothing

Practical, in order.

Send to the people who already paid you

Not a bought list. Not a prospect list. The 64% column from reason three is your own client file, and in most businesses it is the only list that never gets mailed. Start there, even if it is forty people.

Order more than you think you need

Ordering short is the expensive mistake in this business. A reorder costs you a second setup, a second shipping charge and a week you do not have in December. The gap between 200 and 250 cards is small money. The gap between running out on December 12 and not running out is a bad afternoon.

Sign them yourself, even if you buy everything else

This is the part of the job that cannot be bought, and the arithmetic in reason ten says it is also the most expensive part to pay someone else for. Two handwritten lines and a real signature is the whole difference between a card and a mailing. If the number of cards makes that impossible, you have either too many cards or the wrong list.

Say one specific thing

“Thank you for your business” is not a sentence anybody remembers. “Thank you for sending the Hendersons our way in March” is. One concrete detail per card. It takes eleven seconds and it is the difference between a card that gets read and a card that gets counted.

Do not sell anything in it

No offer, no discount code, no “book your spring review.” The moment there is an ask in it, it becomes a mailer and the recipient files it accordingly. The card works because it asks for nothing. Let it ask for nothing.

Do it twice a year, not once

One holiday card a year is a ritual, and rituals get read as obligations. A card in December and one that arrives in April or June for no reason at all is a pattern, and the second one is the one they will mention to you. The research in reason five says the unscheduled one lands hardest.

Start in September, not November

Not because we want the order early — though we do, and we discount for it, because an early order is easier for us to produce properly. Because November lead times are not September lead times, proofs take longer when the plant is full, and every year I watch somebody good end up sending a card that arrives on December 28.

Common questions