Carbon Footprint: Paper Cards vs Ecards

The headline answer, in one paragraph

A single paper greeting card, cradle-to-grave (paper production, printing, envelope, postal transport, recipient disposal), works out at somewhere between 60 and 150 grams of CO2-equivalent, depending mostly on whether the card is virgin or recycled stock, whether it travels by ground or air, and whether it ends up in landfill or recycling. A single ecard, cradle-to-grave (the few seconds of server compute to render the page, the network bytes to push it to the recipient, the device-time to read it on a phone for thirty seconds), works out at something on the order of 1 to 4 grams of CO2e. That is roughly a 30x to 100x gap in favour of the ecard in the typical case, which is a real number but smaller than the one you would get if you only counted the postage and ignored everything else. The methodology that produces those numbers is below, with sources you can argue with.

How I built the comparison (methodology)

A note on what makes this hard before the numbers. Every LCA (life-cycle assessment) of consumer goods at this size involves rounded inputs and contested assumptions; the gap between two careful LCAs of the same product can be 2-3x just on methodology choices, and that is for products where the inputs are easier to bound than this one. The input data for the digital side is harder than people pretend. Device share, network share, and data-centre share are all moving fast; the IEA's 2024 update noted data-centre electricity demand roughly doubling between 2022 and 2026 driven by AI workloads, which matters less than you would think for an HTML page but it matters at the margin. And I have a commercial interest in one side of this. I have tried to be conservative on the digital side to compensate for the obvious bias, and I have flagged below where I made that choice.

For paper cards, I used PAS 2050 as the structural framework, the Carbon Trust's published methodology notes for paper goods, and the most-cited peer-reviewed LCA paper on greeting cards I could find (Madsen 2007, which is old but the constants for paper production and postal transport have not moved much). Mike Berners-Lee's How Bad Are Bananas? (2020 edition) gave the sanity-check number of about 140g CO2e for a typical paper greeting card with envelope sent by post, which is in the middle of my range. Paper-production emissions came from the European Paper Industry Confederation's 2023 sector report, which put greeting-card stock at around 1.0-1.3 kg CO2e per kg of paper for virgin coated stock and about half that for recycled. A typical greeting card with envelope is around 20-30g of paper.

For ecards, I built a per-card budget from three components. Server-side render: a single page render and email-template fill on a modern cloud function is around 0.05-0.15 kWh per thousand requests at current data-centre efficiency (rough average from IEA Electricity 2024 and a couple of recent academic estimates), which works out to about 0.05-0.15g CO2e per card on the global average grid (about 450g CO2e/kWh). Network: pushing ~200 KB of HTML, a couple of images, and an email to one recipient is roughly 0.4-1.0g CO2e using the SWDM (Sustainable Web Design Model) v4 constant of about 0.5g CO2e per MB transferred (this is contested; some researchers put it lower). Device: 30 seconds of read time on a modern smartphone is about 0.0005 kWh, which is around 0.2g CO2e. Add a generous reply-and-look-again allowance and you get about 1-4g CO2e per card, which lines up roughly with the Carbon Trust's published estimates for a typical short web session.

I am deliberately not counting the embodied carbon of the recipient's phone in the ecard number, because they own the phone anyway and would be running it regardless. The fair counterfactual for the device is 'what is the marginal CO2 of one more thirty-second page view', not 'what is the share of one whole phone's lifetime emissions allocated to this card'. If you do the latter allocation, the ecard number balloons to something silly and the comparison becomes apples-to-something-else. The Madsen LCA I used for paper does not allocate the recipient's mailbox or the postal carrier's truck in full to one card either; it uses marginal transport on a route that was running anyway. The methodologies have to match, and I have tried to match them.

The comparison table

A single send, typical case, all numbers in grams of CO2e per card. Ranges are deliberately wide because the spread between best and worst case inside each category is real.

Sources: EPIC paper-sector report (2023) for stock and printing; Madsen (2007) for the full paper-card LCA structure; EPA WARM model (v16) for paper end-of-life methane factors in US landfill conditions; IEA Electricity 2024 for grid-average emissions and data-centre efficiency; SWDM v4 (2024) for the network constant; Berners-Lee How Bad Are Bananas? (2020) as the sanity-check number. I would not bet the farm on any single row, but the bottom-line ratio is robust across the methodology choices I tried.

Where paper actually wins (the inconvenient cases)

If you stopped reading at the headline number you would walk away with the wrong picture of when this comparison is and is not interesting. There are at least three cases where paper either wins outright or narrows the gap so far that the result becomes uncomfortable for the digital side, and I want to walk through each one with the same numbers I just used on the headline.

The first case is the kept card. A paper card that the recipient keeps on a windowsill or in a shoebox for twenty years is amortised, in a loose sense, across two decades of latent function. The card emits its 100g once at production and then sits emitting nothing for the rest of its life. The ecard, in theory, also emits nothing for the rest of its life, but in practice the ecard often gets re-opened a few times over the years (the platform's email reminder, a quick screenshot for the holiday-card slideshow, a re-share to a relative). Each re-open is another small slice of network and device time. Across twenty years and a generous 30 re-opens, the ecard's cumulative number might creep up to 10-20g while the paper card stays at its initial 100g. Paper still wins on the total, but the margin has shrunk by an order of magnitude. The keepsake case is the one where the digital-superiority case is weakest, and the digital-card industry mostly does not talk about it.

The second case is the ecard opened on the wrong device at the wrong moment. A paper card has a roughly bounded carbon cost. An ecard opened by the recipient on a 4K television (200-300W draw) for three minutes, projected onto the family-room TV during a holiday gathering, will use maybe 15g of grid energy in one viewing, which is more than the paper card would have used during the read step and a meaningful slice of the ecard's whole-life budget. If the same ecard then gets opened on a gaming-PC monitor for ten minutes during a video-call reaction, the device-side draw alone can push the ecard's per-event cost into the 30-60g range. This does not flip the headline number unless the ecard gets viewed in this way many times, but it narrows the gap and is a real failure mode I have observed in my own family. The IEA's data on residential electricity use is the source for the device-draw numbers; the rest is arithmetic.

The third case is the worst-case digital pipeline. A poorly-optimised ecard platform that ships 20MB of un-optimised video on the cover, hits a US-coast data centre on a grid that runs at 600g CO2e/kWh in some hours, and runs a real-time video-encode on every send, can push the per-card ecard number into the 10-30g range. Still much better than paper at the high end, but no longer a 100x gap. This is a quality-of-implementation issue rather than an inherent property of digital, and it is solvable, but it is real for some platforms.

Add it all up and the honest summary is: the ecard wins by something like 30-100x in the typical case (a normal HTML card, opened on a phone, read for thirty seconds, not re-opened more than a couple of times). The gap narrows to 5-10x in the keepsake case. The gap narrows further or flips for the worst-case digital implementation viewed on a 4K TV repeatedly. In aggregate across all the cards a normal person sends in a year, the ecard probably wins by closer to 50x than 100x, and I would not feel comfortable writing the headline number any higher than that.

The scaled view: what this means for a workplace that sends 200 cards a year

Per-card numbers are a footnote at the personal level. A typical adult sends maybe 15-25 cards a year, which is a total annual cost of about 1-3 kg CO2e for the paper version, or 20-60g for the ecard version. The difference at this scale is about the same as one short drive in a normal car, which is real but not a thing most people will reorganise their lives over.

Where the scaled view actually matters is the workplace. A 200-person company that sends two cards per person per year, plus team cards for every farewell, work anniversary, and recognition moment, is sending somewhere north of 500 paper cards a year if it does this on paper. That is 30-75 kg CO2e annually for the card budget alone, which is around 2-3 normal car-tank-fills of gasoline. The same workplace doing this all by ecard is at 0.3-1 kg CO2e for the same volume, two orders of magnitude lower. This is the number that does justify writing a memo about; it is just rarely the number that gets written about, because the per-card number sounds small and the corporate scale gets buried in 'employee experience' line items rather than carbon line items.

If you are running team recognition and the carbon angle matters to your sustainability report, the order-of-magnitude lift from switching to digital is one of the cleanest unit-economics moves in the building, and the implementation cost is approximately zero compared to whatever else is in the same sustainability spreadsheet. I would not bet the climate on it, but I would put it in the report.

What this comparison cannot tell you

A few things the numbers do not cover, so you do not over-read what is above. This is not a comparison of which medium is more meaningful. The Renata-on-the-windowsill case (a real paper card a real person keeps for years) is doing emotional work that an ecard often does not, and the carbon number does not capture that work. A separate piece, online vs physical greeting cards, covers the meaning side of the comparison directly. The answer to 'which card lands better' is not the same as the answer to 'which card has a lower carbon footprint'.

This is also not a comparison of which medium is better for the sender, the recipient, or the group. For workplace and group occasions specifically, the format question is mostly settled in favour of digital on logistics alone (multiple signers, scheduling, time-zone delivery), and the carbon story is a tailwind rather than the headline. Group ecard vs personal ecard covers the logistics version of the same question.

And the comparison does not capture the second-order effects. If the ecard makes it so easy to send a card that you end up sending three times as many of them, the carbon savings per card get partly eaten by volume. This is the Jevons paradox my friend Jorma was pointing at in Helsinki. I do not have a clean answer to it except that the per-card carbon delta is large enough that even a 3x volume increase still leaves the digital version meaningfully ahead. I would not push that argument too far. It is the kind of thing that makes the answer messier the longer you stare at it.

Turn the carbon story into a card the team actually sends

If the carbon number is what gets your sustainability lead across the line on the switch to digital recognition cards, the practical move is to actually use it. A group ecard with multiple signers covers the volume cases (farewell cards, work anniversaries, milestone birthdays) that drive most of the paper budget in a normal office. You can create a card online in about two minutes, share the link to the team chat with one sentence of guidance, and let the signatures fill in over the week before the occasion.

For the broader 'how do I actually build one of these' question, how to make a digital greeting card walks through the format choice (one-page vs multi-page vs board) and the failure modes I have walked into personally. And for the recognition version of the same conversation, where the carbon angle is a footnote and the artifact-vs-record question is the main event, ecards vs email for recognition is the longer piece on what actually lands.

One last thing, off-topic and probably just for me. The panel where Jorma cornered me was actually a panel I had no business being on. I had said yes to it three months earlier on the assumption that the organiser would find a better speaker and quietly drop me, and instead they had emailed in late January to confirm and I had to fly out. The room was the smallest plenary room at the venue and had a single window looking onto a parking lot with one of those Finnish heated-cable systems running along the kerb to keep the snow off, and I remember staring at the cable while a representative from a recycled-paper trade group was speaking about supply chains and thinking that I had nothing useful to say after her. Jorma was sitting in the third row in a sweater that looked like it had been knitted by someone in 1986. I did not catch his last name. He bought me an Americano in a paper cup with one of those compostable lids that you press down on and it never actually stays closed, and I drank it on the walk back to the hotel through about an inch of fresh snow, the cup leaking slightly the whole way. I came home that weekend, opened a fresh Google Sheet, and started this spreadsheet on the kitchen counter with the laptop balanced on a stack of unread New Yorkers from 2024. The Americano cup is, almost certainly, the highest-carbon thing involved in writing this whole article. I have not bought a coffee in a paper cup since.