We read 12,282 gift requests written by real people, in their own words, with nobody asking them to. The most striking thing is what almost none of them contain.
Every gift guide ever published is organised by price. "Gifts under $50." "Best presents under £25." It is the default structure of the genre, used by every major publisher, because it seems obvious that the first thing a gift buyer knows is how much they want to spend.
It turns out that is not how people describe the problem at all.
Only 5.6% of 12,282 requests included any price signal: a currency amount, the word budget, "cheap", "affordable", or "under" followed by a number.
This is the clearest finding in the data and the one that contradicts the industry most directly. People arrive knowing who the gift is for and, sometimes, what that person likes. The price is apparently something they work out later, once they have seen options, rather than a filter they bring with them.
If that is right, then sorting gift content by price is sorting by the one attribute the reader has not yet decided.
We looked for any usable signal in each request: a named hobby or interest, an age, or a price. 66.1% contained none of the three.
An age appears in 12.3% of requests. A named interest appears in about a third. The single most common shape of a gift request is a relationship and nothing else: "a gift for my mum", "something for my brother".
People do not write briefs. They write fragments. This matters for anyone building search or recommendation tools, because seven words is not enough to distinguish one person from another, and the systems being built to answer these questions generally assume far more detail than users supply.
The longer requests, when they come, are strikingly specific:
"My boyfriend 30th birthday, he likes to draw, music, biking, hiking, padel"
"Something for my 26 year old brother who's a lawyer, likes cooking, and more"
"Something for my 18 year old niece. She is studying arabic and likes reading"
"Her fitness goal is to lose fat and grow muscles. She's an intermediate weightlifter with no allergies or dietary restrictions"
Mothers are the most asked-about recipient by a wide margin, appearing in 6.9% of requests against 4.1% for fathers: a ratio of roughly 1.7 to 1.
| Recipient | Share of requests |
|---|---|
| Mum or mother | 6.9% |
| Dad or father | 4.1% |
| Boyfriend | 3.9% |
| Friend | 3.9% |
| Brother | 3.2% |
| Husband | 2.9% |
| Girlfriend | 2.8% |
| Sister | 2.7% |
| Daughter | 2.6% |
| Wife | 2.3% |
| Son | 1.7% |
| Grandparent | 1.0% |
| Colleague or coworker | 0.5% |
| Teacher | 0.5% |
Among partners and siblings, the man is asked about more often: boyfriend over girlfriend, husband over wife, brother over sister. Among parents and children it reverses: mum over dad, and daughter over son. So this is not simply "men are harder to buy for". People appear to find partners and brothers harder than wives and sisters, while mothers and daughters attract more searching than fathers and sons.
Office gifting is a staple of seasonal content, yet colleagues and teachers each appear in half a percent of requests. "Secret Santa" appears more often than "colleague" does, which suggests people search by the ritual rather than by the relationship.
Eight occasions between them account for under a fifth of all requests, and that figure counts a request naming two of them twice. The large majority of people do not say what the gift is for. They expect it to be understood, or it does not occur to them that it matters.
| Occasion | Share of requests |
|---|---|
| Christmas | 10.3% |
| Birthday | 4.7% |
| Anniversary | 1.5% |
| Secret Santa | 0.7% |
| Wedding | 0.4% |
| Graduation | 0.4% |
| Retirement | 0.3% |
| Valentine's Day | 0.1% |
Birthdays are the one everyday occasion that registers at all, and they appear more often than every non-Christmas occasion combined. Weddings, graduations and retirements, which between them support an enormous amount of published gift content, barely feature in how people describe their problem.
Where an interest is named, technology leads, but no single category comes close to dominating. The distribution is flat, which is its own finding: there is no small set of interests that most gift buyers are working with.
| Interest named | Share of requests |
|---|---|
| Tech and gadgets | 3.0% |
| Books and reading | 2.8% |
| Gaming | 2.3% |
| Pets | 2.3% |
| Music | 2.1% |
| Travel | 2.0% |
| Sport | 2.0% |
| Cooking and baking | 1.7% |
| Fashion | 1.6% |
| Outdoors | 1.5% |
| Fitness | 1.2% |
| Art and drawing | 1.2% |
| Gardening | 1.0% |
| Coffee | 0.7% |
Only 1.5% of requests describe the recipient as hard to buy for, picky, or as someone who "has everything". The phrase is a staple of gift writing. It is not how most people describe their problem.
This data is unusual, and it is worth saying why. Search keywords tell you what someone typed into a box. These are whole sentences: people describing a person they love, in their own words, with no categories to pick and no filters to set. Nobody was surveyed and nobody was prompted. They just wrote what they needed.
The corpus is drawn from two conversational gift recommendation tools we operate. From 30,388 stored strings we removed the sites' own suggested prompt texts, identified as any string appearing 20 or more times identically. Leaving them in materially changes the results, and in two cases reverses them, so the distinction matters. We also removed one-word greetings and any string carrying assistant phrasing, including rows mislabelled as user messages. That leaves 12,282 requests written by people.
One of the two tools is conversational rather than one-shot, averaging 1.7 user messages per conversation. For those, the unit of analysis is the whole conversation, since that is the complete brief a person gave and the closest equivalent to a single written request.
Categories were matched with case-insensitive word-boundary patterns, applied identically across the whole corpus. A request mentioning two interests counts in both, so interest percentages do not sum to 100. Percentages are of all 12,282 requests unless stated.
The limitations are real and worth stating plainly. This is first-party data from two small sites rather than a representative sample of gift buyers. Traffic skews to the United States and heavily to the fourth quarter, which lifts Christmas relative to year-round occasions and understates spring ones such as Mother's Day and Valentine's Day: the occasion figures in section 5 should be read as shares of our traffic, not of gift buying in general. Keyword matching undercounts anyone who describes an interest without using the obvious word for it. And people may phrase things differently to a chat interface than they would to a search engine or a shop assistant. These numbers describe how people talk to a gift recommendation tool. We think that is interesting, but it is not the same as how people shop.