Emoji.Ad
248 emoji domains in .to
I own 248 emoji domains in .to. That gives me an obvious interest in these tests coming out well. So this page publishes the parts that came out badly, alongside the exact questions, the sample sizes, the controls and the coding rules — so anyone can check the arithmetic instead of taking my word for it.
Primary-domain replacement · universal comprehension · faster typing · professional bio use · standalone clickable link.
A secondary visual signature for established brands, especially in mobile-first visual campaigns.
The brand stays. The emoji signs. The QR routes. The destination stays readable.
Every wave uses an open-ended question, except three forced-choice comparisons, which are clearly marked in the table below.
These panels measure comprehension, declared behaviour, forced preference, stated hesitation and perceived memorability. They do not measure sales, click-through, real navigation, or delayed recall. In the memorability waves the stimulus stayed on screen while people answered, so what is recorded is what someone says they would remember, not what they later do remember. That distinction matters and it is not a formality: a claim like "five times more memorable" would not survive it, which is why it is not made anywhere on this site.
Single coder, no blind second pass. All free-text answers were coded by hand, by me. There is no inter-rater reliability figure. The coding rules below are published precisely so that the boundary cases are visible and arguable — several of them are genuinely arguable.
Scroll the table sideways to see every column.
| # | Market | n | Stimulus | Exact question | Control |
|---|---|---|---|---|---|
| 1 | France | 75 | 😎.to alone, no context | "Qu'est-ce que c'est, selon vous ?" | None |
| 2 | France | 75 | Bus-shelter poster, fictional brand SOLARE, www.😎.to |
"Vous voyez cette affiche dans la rue. La marque vous intéresse. Que faites-vous ensuite ?" | Wave 3 |
| 3 | France | 75 | Same poster, www.solare.fr |
Identical to wave 2 | Control for 2 |
| 4 | France | 75 | Same poster, brand name removed, www.😎.to only text |
"Vous voyez cette affiche dans la rue. Ce produit vous intéresse. Que faites-vous ensuite ?" | None |
| 5 | France | 75 | No-brand poster, www.😎.to |
"Vous voyez cette affiche dans la rue. Ce produit vous intéresse. Écrivez exactement ce que vous taperiez dans votre navigateur. Si vous ne taperiez rien, dites-le." | Wave 6 |
| 6 | France | 75 | No-brand poster, www.solare.fr |
Identical to wave 5 | Control for 5 |
| 7 | USA | 75 | No-brand poster, www.😎.to |
English translation of wave 5, word for word | Wave 5 as cross-market comparison |
| 8 | USA | 75 | Two identical X profiles; bio link 😎.to vs leomarchand.com |
"This creator's profile has a link in their bio. Which version would you be more likely to tap? Why?" (forced choice) | Wave 9 |
| 9 | USA | 75 | Same profiles; bio link 😎.to vs linktr.ee/leomrchd |
Identical to wave 8 (forced choice) | Alternate opponent |
| 10 | USA | 75 | Audio-only scenario, no image | "A radio ad ends with: 'Order at the pizza emoji, dot T-O.' Type exactly what you would type into your browser to reach that site. If you would type nothing, say so." | None |
| 11 | USA | 75 | Two brand posters; 🏀.to vs nike.com/basketball |
"Which poster's web address would you be more likely to visit? Why?" (forced choice) | Within-pair |
| 12 | USA | 75 | Subway corridor, seven posters, six carrying conventional .com addresses and one carrying www.😎.to |
"Look at this scene for a moment. What is the first thing that catches your eye? Why?" | None |
| 13 | USA | 100 | Brand poster carrying a QR code, a conventional URL and an emoji address | "Imagine you passed this poster on the street. What is the one thing about it you think you would be most likely to remember later? Why?" | Wave 14 |
| 14 | USA | 100 | Identical poster with the emoji address removed | Identical to wave 13 | Control for 13 |
| 15 | USA | 100 | Same brand poster, conventional URL plus the emoji address labelled official shortcut | "If you saw this poster in the street and wanted to visit the brand's basketball page afterward, what would you do? Please state exactly what you would type, tap, scan, or search, and explain why." | Wave 16 |
| 16 | USA | 100 | Identical poster with the emoji address removed | Identical to wave 15 | Control for 15 |
| 17 | USA | 100 | Verified brand X profile, 3.1M followers, bio link 🏀.to |
"Imagine this were the brand's official X profile and you were interested in their basketball line. Would anything about the link in the bio make you hesitate to tap it? Why or why not?" | Wave 18 |
| 18 | USA | 100 | Identical profile, bio link nike.com/basketball |
Identical to wave 17 | Control for 17 |
The answer contains the address as it appeared on the stimulus, including the glyph itself — the respondent found 😎 or 🏀 on their own keyboard and placed it in the text field. Copy-paste from the image was impossible, so this records an act, not a stated intention.
Counted separately, never merged into the headline figure: answers that describe the gesture correctly without executing it ("I would find the pizza emoji and enter that, then add .to"), and near-misses with the right characters but a missing dot or prefix ("wwwsolare.fr").
The answer keeps the .to ending but replaces the glyph with a word — solare.to, sunglasses.to, cool.to, basketball.to, www.to, or a keyboard emoticon such as www.8).to. These are failures, and they are the single most consistent pattern across all eighteen waves.
The answer points at the address — "the site indicated on the poster", "the link written at the bottom" — without writing any of it. Reported separately because it shows the address was seen, which is a different finding from the address being usable.
The answer names that element, alone or among others, in reply to what the respondent would remember. Multi-element answers are counted once for each element named, so the column totals exceed the sample size. This is why the second comparison on the home page is expressed in mentions and not in percentages.
The soft edge: answers naming "the website" or "the link" without saying which one, on a poster carrying two addresses, were not attributed to either. There were five such answers in wave 13.
Counted only when the answer names the link, the address, the glyph or the .to ending as the source of the hesitation. Three categories were kept separate and are reported separately below: distrust (scam, phishing, virus, "I wouldn't know where it goes"), non-perception (the respondent does not recognise it as a link at all), and typing rehearsal (the respondent worries about entering the glyph even though the link is clickable).
Not counted as hesitation, and reported as off-topic: dislike of the artwork, of the dark background, of the brand, of the sport, or of the brand's politics. That bucket is large — roughly a fifth of each wave — and it is the reason the two waves are compared on link-attributable hesitation only.
Only an explicit choice was counted. Answers refusing both options were recorded as neither. Display order was alternated; in wave 8 it was not, and roughly ten answers referred to "the one on the right" with no other reason given — a known weakness of that wave.
On n=75 per cell, 16% against 13% is not a measurable difference. The honest reading is not that the emoji address wins; it is that a printed address is not a traffic channel for either format. The conventional URL produced its own spelling failures — solar.fr, solaire.fr, soware — because the French word solaire is a trap of its own.
One asymmetry survives and it favours the conventional address: about fifteen respondents in wave 6 typed the brand name into a search engine and would have reached the site anyway. An emoji address offers no such second route.
basketball.toWave 15 was designed to give the emoji address every advantage available: a real and instantly recognisable brand, and an explicit label telling the reader this was the sanctioned short route. Twelve people out of a hundred took it. The rest typed the conventional URL or searched the brand name, and a recurring reason was that they did not need a shortcut for an address they could already remember.
The motive of the twelve matters more than the number. Almost all of them named curiosity rather than convenience — wanting to find out whether an emoji address works at all. That is a first-visit effect, and nothing in these panels suggests it repeats.
Two failure modes appeared here that had not appeared before. Five respondents described tapping or scanning the glyph on a paper poster, which is not a comprehension failure about emoji domains but about the medium. And four converted the glyph back into its English word, the same substitution recorded in France with solare.to — a reminder that the corresponding word domain is registrable by anyone, for the price of a normal renewal.
Respondents explained the gap themselves, unprompted: two taps on a phone keyboard, an unfamiliar shortcut on a desktop one. In wave 10, roughly a third of the panel described the gesture correctly while saying they could not perform it where they were sitting. The results point strongly toward a mobile-first proposition rather than a desktop-first one.
In wave 1, with no context whatsoever, 8 respondents out of 75 identified 😎.to as a web address and about 7 more got partway there. Sixty described a smiley and nothing else. In wave 12, inside a corridor of seven posters, not one respondent out of 75 named any address — neither the emoji one nor the six conventional ones. At that scale the address line is roughly one percent of the surface and competes with two-metre blocks of colour. It does not exist.
.com (wave 8)linktr.ee link (wave 9)Wave 9 is the harder result. The opponent was an ugly truncated link with a slash and an abbreviated username — one respondent called it weird — and it still won 72 to 3. The reason given, over and over, is familiarity rather than aesthetics, and the recurring vocabulary is scam, phishing, sketchy, could redirect anywhere. The three favourable answers were all aesthetic and none was functional: very cute, more playful, piques my curiosity.
The obvious objection to that pair was that the distrust might attach to the unknown sender rather than to the address. Waves 17 and 18 were run to settle it.
This is the single most useful comparison in the study, because only one element changed between the two screenshots. The brand does an enormous amount of work: from zero out of seventy-five under an unknown photographer to roughly forty-five percent explicit comfort under a verified account with 3.1 million followers. Respondents named the mechanism themselves — the gold check, the follower count, the age of the account.
And it still is not enough. Thirty-six against five. Verification buys permission for about half the audience; it does not make the address read as an address. If the strongest available brand cannot close that gap in a clickable, no-typing context, no smaller brand will.
This is the result that costs me the most. The clickable bio link is the use case with no typing friction at all, and it is where the emoji address performs worst across all eighteen waves. The binding constraint was never the keyboard. It is that a reader cannot tell where the link goes before committing to it.
This is the cleanest comparison in the study, because in wave 13 both addresses were on one poster, seen by the same hundred people, in the same second. No version bias, no ordering effect.
It is also third place. The brand mark drew roughly 45 mentions and the product image roughly 20 in both versions. The emoji address does not dominate a poster; it opens a pocket of salience where a conventional URL opens none.
The obvious objection to the 14 was that the emoji address simply steals attention from the QR code. Without it, the QR code rises from about 12 to about 17 mentions and the conventional URL from 1 to 3 — but the whole bottom zone collects 20 mentions instead of 27. That pattern suggests the emoji may enlarge the visible action zone rather than simply reallocating attention, although this sample is too small to treat the difference as conclusive.
The concession that belongs with that figure: on the poster carrying both addresses, the conventional URL fell from 3 mentions to 1. The emoji address does take something from it. It is a small something taken from a small number, but it is real and it is the strongest argument available against this result.
Across the poster waves, the behavioural wave and the bio-link wave, roughly four percent of every panel says some version of the same thing: that they would take the emoji route because it is unusual. It is the most reproducible positive finding here, and also the smallest. Four percent is a niche, not a channel — but it is the same four percent whether the address is printed, labelled a shortcut, or clickable under a verified profile.
The most useful thing wave 17 produced is not the headline number. It is that the thirty-six hesitations are not one objection but three, and they have different remedies.
Scam, phishing, hijack, virus, "the link is too short and generic", "I need to see words in the link so I know what I'm clicking". What is missing is a readable destination, and no styling supplies it. This is the objection that also decided waves 8 and 9.
These respondents do not distrust the link. They do not see that it is one. "I wouldn't even think that it was an actual hyperlink." "It doesn't look like something that is clickable, it is just an emoji of a basketball." "I would have difficulty knowing if the emoji was meant to be a link in bio or just something to catch the eye." "The link in bio is so short I barely noticed it."
This is an affordance problem, not a trust problem, and it is the one finding in this study that points at a fix rather than at a limit. Nobody appears to be working on it.
Faced with a link they only have to tap, some respondents still mentally rehearse entering it: "I am actually confused what to type." "If I type the emoji, I don't think it would go to the site. Sometimes I like to copy and paste URLs." The typing objection survives even where typing is not required, which suggests it is partly about control over the destination rather than about effort.
Wave 18 respondents said plainly what the emoji version lacked. The link isn't obfuscated. It is a standard URL that I'd expect to see. I know where I am going to when I click on it. And the most instructive of them all: I can tell quickly where it will take me, the "basketball" after the slash also tells me it won't be to the general site. Transparent.
Read together, those answers do not reject the glyph. They reject the absence of a destination. That reframes the proposition: the emoji address may belong alongside a readable destination rather than in place of one — an emoji signature paired with a visible target, rather than a bare emoji standing where a URL used to be. That configuration has not been tested here. It is the next wave, and until it is run this remains a hypothesis rather than a finding.
Selected across waves, translated from French where relevant. The unfavourable ones are here for the same reason as the favourable ones.
"The thing I would probably most remember about it would be that dumb useless emoji link."Wave 13 · USA
"I know what a linktree is, I'm used to it. The other one looks like it could redirect anywhere and that concerns me for security purposes."Wave 9 · USA
"The two links are just bad to me, you would think the emoji link would be easier to type but it is not, it's gonna be confusing for people."Wave 13 · USA
"It could be a hacked page and a phishing scheme for all I know. I can't trust what is currently presented."Wave 17 · USA — under a verified brand profile
"The link is too short and looks strange to have a basketball emoji as part of the URL. So my first instinct is to avoid clicking it."Wave 17 · USA
"Just be clear and simple, get rid of the basketball emoji shortcut."Wave 15 · USA
"I burst out laughing. That address is impossible."Wave 4 · France
"Why make it a puzzle? You're just making it harder for people to do business."Wave 10 · USA
"🏀.to — I would remember it because it is easy to remember and different. Less to type than a .com."Wave 13 · USA
"I would actually be more likely to tap the basketball link than a regular link just out of the novelty of using an emoji url."Wave 17 · USA
"I trust the link and think it looks very modern."Wave 17 · USA
"The domain with the emoji is interesting more than anything else. Piques my curiosity."Wave 17 · USA
"The new trend to use an icon .TO to connect. On this poster it is attractive due to the black background and the basketballs matching."Wave 13 · USA
"I typed what is written, www.😎.to, because those are the glasses I want and I'd land straight on it."Wave 5 · France
"I wouldn't even think that it was an actual hyperlink."Wave 17 · USA — the affordance failure, in seven words
"I would have difficulty knowing if the emoji was meant to be a link in bio or just something to catch the eye."Wave 17 · USA — the two functions, correctly separated by a respondent
"I can tell quickly where it will take me, the 'basketball' after the slash also tells me it won't be to the general site. Transparent."Wave 18 · USA — the control wave writing the specification
"It depends. If I'm on mobile there's a good chance I might be curious about the novelty of it. On PC, I have to remember the keyboard combination that brings up the emojis and pick it out… It's a pain."Wave 10 · USA — the device effect, stated before it was inferred
"Does it imply it is basketball.to? Or is the emoji hiding some kind of nefarious ad-ridden link?"Wave 17 · USA — substitution and distrust in one answer
Waves 2 to 7 used a bus-shelter mockup for a fictional sunglasses brand, SOLARE, with a single line of address at the bottom. Between the paired waves only that line changed; everything else — photograph, framing, type size, position — was identical, and the control was produced by editing the exported file rather than regenerating the image.
Waves 8 and 9 used two screenshots of the same fictional creator profile, identical down to the follower count, differing only in the bio link.
Wave 12 used a corridor of seven posters for seven fictional brands, six carrying conventional .com addresses.
Waves 17 and 18 used two screenshots of the same profile page, identical in banner, avatar, bio text, verification badge, follower count and first post, differing only in the bio link line.
Waves 11 and 13 to 18 are not reproduced here. They used a real, well-known brand as a familiar-brand stimulus in a research mockup, because no fictional brand can carry the recognition the hypothesis required. That is defensible as private comparative research; republishing the artwork is not. The figures stand; the images stay unpublished. The brand is not affiliated with, and does not endorse, Emoji.Ad.
Dropped: that an emoji address is universally and instantly understood; that it is faster to type; that it works as a primary corporate domain; that it belongs in a professional bio; and — after waves 17 and 18 — that brand verification is enough to make it work as a standalone clickable link.
Kept: that under an established brand, on mobile-first visual media, it can open a pocket of attention a conventional URL does not; that in this poster test the actionable zone drew more mentions when the emoji address was present; and that a small, stable minority is drawn to it precisely because it is unusual.
Open: whether pairing the glyph with a readable destination removes the distrust while keeping the salience. That is the next wave, and it is not answered here.
The brand stays. The emoji signs. The QR routes. The destination stays readable.
The .to registry no longer accepts new emoji registrations, so what exists is all there will ever be. Every figure on this page came out of testing that register against conventional addresses.