What 14,346 people really asked a psychic

Every reading I write begins with a box on an order form. No checkboxes, no dropdown of topics, just an empty field and whatever the person decided to put in it. I have been reading those boxes for years, and if you had asked me what was in them I would have given you a confident answer without pausing.

Then I counted them, and a decent part of my confident answer turned out to be wrong.

What follows is 14,346 questions, submitted with readings between February and July 2026, looked at as counts instead of as a queue. Average length: 80 words. Some of it confirmed what I already believed. The biggest finding changed how I write.

What I counted, and what nobody read

The method belongs at the top, because in a piece like this the method is the credibility. Without it, a psychic telling you what her clients ask is just a psychic telling you things.

All of this comes from aggregate keyword counting. A program went through all 14,346 messages at once and handed back totals: how many contained words from a money list, how many ended with a question mark, how many referred to somebody other than the sender. Totals are all I have. No individual question was read for this study, nothing was pulled out for being interesting, and there is not one real sentence from a real client anywhere in this article.

That last part is not squeamishness. People put things in that box they have not said to their own sister. Whatever a study like this is worth, it is not worth a line of somebody's worst week turning up in a blog post, anonymized or not.

So when I describe what a question looks like, I am building the shape out of the counts and writing the sentence myself. The examples below are invented. The proportions are real.

Only 30.2% end with a question mark

This is the finding I have not stopped thinking about.

Nearly everything that arrives is a situation. Someone sets down what is happening, at an average of 80 words, and then stops. Here is an invented one, built to that average shape:

"We were fine and then his replies got shorter and now he says he needs space but he still calls on Sundays and I honestly don't know what I'm doing anymore."

Technically there is nothing there to answer. In practice I know what she wants: what the Sunday calls mean, and underneath that, how much longer she should keep her life on hold. The punctuation goes missing because the need is not grammatical. A situation that will not sit still in your head is very hard to phrase as a query, so it comes out as a paragraph, and sending the paragraph is the closest thing available to setting the weight down.

The other shapes are familiar, and now I know roughly how common they are. One is the timeline with names and dates and no question anywhere in it, sent by people who handle what they cannot control by being organized about it. Another is the short message with an enormous decision folded up inside. My favorite, if that is the right word, ends "I just want to know if I'm going mad", which is not a question about the future in any sense.

Seeing the number changed how I work. I used to answer the question. Now I answer the situation and name the question inside it, because plenty of people genuinely do not know what they are asking, and telling somebody what they asked is often the most useful line in the reading.

What people bring, by size

The buckets overlap on purpose, because real situations do. A message about following a partner to another city counts under love and under moving, and if there is an aging mother in it, it counts under family and health as well. These do not add up to anything tidy and were never meant to.

• Love and relationships: 54.7%
• Career and work: 39.2%
• Family: 31.2%
• Money: 23.2%
• Timing words, "when will" and "how long": 17.2%
• Health: 17.0%
• Moving and relocation: 15.9%
• Pregnancy and children: 14.7%
• Spiritual purpose: 7.9%
• Asking outright whether someone is coming back: 5.1%
• Grief and loss: 3.7%

Love, at 54.7%

More than half of everything. That part did not surprise me. The breadth of what sits inside it might surprise you.

The over-half is not all romance in the greeting-card sense. It holds marriages where nobody has said anything true in a long time, and couples who have never agreed what they are to each other and are both pretending that suits them. Some of it is somebody who left and cannot tell whether they were right. Some is somebody who was left and is still waiting for a real explanation. A fair amount is somebody with nobody in particular in mind, asking whether that ever changes.

Sitting inside that bucket, 5.1% of all messages ask outright whether a particular person is coming back. I handle those more carefully than anything else that arrives, because it is the corner of my trade where the most damage gets done, to the people least able to fend it off. The long version of what this kind of reading can and cannot show is at what a love reading can actually tell you, including the parts where the honest answer is unwelcome.

Career and work, at 39.2%

The second country, and larger than most people would guess for a psychic's inbox.

These are rarely about careers in the abstract. They are about a manager who has gone cold, a promotion that went to somebody else, a business that is not paying for itself yet, a decision about handing in notice with nothing else lined up. And they are very often about a person: the colleague, the boss, the partner who does not want you to take it. Hold that thought for the next finding.

Family at 31.2%, money at 23.2%

Family outranking money is what I would put in front of anybody who assumes this work is mostly romance and fortunes.

Family is the sister who stopped speaking after the funeral, the parent getting frail with nobody appointed to deal with it, the adult child who has gone quiet, the in-law who runs everything. Money turns up inside those messages at least as often as it turns up on its own: the house, the will, who has been paying for what, the thing that was promised and never written down.

Grief and loss, at 3.7%

The smallest count on the list, and the one I sit with longest.

People do not usually ask me to reach anybody. In my experience of writing these, what they ask is whether the person knew, whether the last conversation was the real one, whether the guilt they are carrying is the right size. I answer those as carefully as I know how, and I do not pretend to more than I have. When the weight of it needs somebody in the room every week, I say that in the reading.

Health at 17.0%

Higher than I expected, and the count that comes with a rule attached.

I am not a doctor. A reading cannot tell you what a scan tells you, and anybody in this trade who implies otherwise is dangerous. When a health question arrives, I say plainly inside the reading that the medical part belongs with a medical person. Then I answer what I can speak to: the fear around it, and the conversation nobody in the family is willing to start.

40.9% are about somebody else

40.9% of the messages mention a third person. He, she, him, her, they. Somebody else's feelings, or somebody else's silence.

I think this explains why people come here better than any theory about belief does.

You can think your way around most of your own life, given enough nights and enough patience. Your own feelings are available to you, including the ones you would rather not have. Your options can be written down and looked at. The thing thinking cannot open is the other person. You can replay a conversation until the words wear out and you will not get a single new fact from it, because the facts are inside somebody who is not telling you.

That closed room is what people are paying to have described. Plenty of the career questions turn out to be about it too.

More people ask what to do than what is coming

10.5% of messages contain "should I". 8.9% open with "Will".

The stereotype of my work has the customer sitting passively across a table, waiting to be told what will happen to her. The counts lean the other way. More people are asking about their own next move than about their fate, and I did not expect the gap to point in that direction.

A further 6.8% contain "why", and those look backward instead of forward. Why he did it, and why she did not see it coming. Those are the hardest to write well, because a why question wants a reason, and reasons are the thing human beings invent most freely.

I will not oversell the keyword counting here. A message can contain "should I" and still be entirely fatalistic underneath, and plenty are.

1.3% apologized

Roughly one in seventy-five messages contains an apology. Sorry. Silly question. You must think I'm ridiculous.

My guess before counting would have been far higher, because the reflex is everywhere else in life. People say sorry to a doctor for taking up his morning with a lump they have been ignoring for months. They say it to the plumber. Handing your trouble to a stranger seems to require a small payment in embarrassment first.

The counts cannot tell me what anybody was apologizing for. Only that they did it. What I can tell you is what happens to those apologies at my end, which is nothing at all. There is no category of question here that needs one attached, and I have yet to receive one I thought was silly. Being embarrassed about what you want is a tax a lot of people pay, and it buys nothing.

The busiest hour is between 8 and 9 in the evening

The folklore version of my job has people writing to me in the middle of the night, wrecked and unable to sleep. It is a good image and the counting will not support it.

The peak hour is between 8 and 9 in the evening, UK time. The quietest is between 5 and 6 in the morning. What that describes is a sofa after dinner: the day's obligations finished, the phone in hand, the first stretch of time that belongs to nobody else. People write to me in the gap where they finally get to think.

The small-hours tail is real and it is thin. Some do write in the dark, and when they do it tends to be the worst night of a bad month, which is exactly why I like that the box is open at that hour. It is simply not where the volume is, and I would have told you otherwise before I counted.

What this study cannot tell you

The limits are more interesting than several of the findings, so they get a section instead of a footnote.

The buckets are keyword matches, which makes them crude. A message reading "money isn't the issue here" counts as a money question. Sarcasm, negation and figures of speech all defeat this method comfortably, so treat every percentage above as an approximate direction.

The window is February through July 2026 and nothing else. I cannot tell you whether people ask different things in December, or in the week after New Year, or when the clocks change, because I have not counted those months yet. Anybody claiming a seasonal pattern from six months of data is drawing a curve through a straight line. I plan to run this count every year, which is the only way that question gets answered honestly.

The clock comes with a caveat as well. The timestamp is the moment the order was created, which is a decent proxy for when the message was written and not the same thing as it. My clients are all over the world, so an hour measured in UK time mixes a lot of different local hours.

Counting also knows the words and not what is under them. The buckets can see that a message contains money words. They cannot see that the money question is a marriage question with its collar turned up.

And none of this says anything about whether the readings themselves are any good. That is a separate argument, and I have made it, along with the objections I think are fair, at whether an online psychic reading is real. A study of what people ask cannot settle what they get back.

Last one. Everybody counted here chose to order a reading from me. That is a self-selected group, and I would not stretch these percentages over the general population.

What the counting says about people

Here is what I did not expect to take away from a spreadsheet.

The raw material is not very various. Love, work, family, money, health, where to live, whether to have a child, the person who died, and how it all turns out. That is close to the whole of it. A keyword program that understands nothing about any of these lives sorted most of 14,346 private messages into a handful of buckets, and not one of the people who wrote them could see anybody else's.

Which is worth knowing if you have ever sat with the cursor blinking, sure that yours is the strange one, or the too-small one. It is on the list. So is nearly everybody else's, including the ones being typed this evening while you read this. Half the value of asking, I have come to think, is finding that out.

If you want yours counted in next year's study, the box is still there on the order form for a full psychic reading. Send it exactly as it comes out of your head. Going by the counts, that is what nearly everybody else did.