A year in numbers · what we learned hosting 58,000 readings

We hosted 58,000 readings in the last twelve months. We wrote down what happened, because we built this platform partly out of frustration that nobody in the online reading industry publishes anything you can actually verify.

Most of what gets called "tarot industry data" is repackaged Google Trends, vague Pinterest charts, and the occasional press release from a per-minute platform that has every reason to make the numbers vague. The reason those platforms don't publish granular data is, frankly, that their incentives don't reward client outcomes. They reward minutes on the line. A platform whose revenue depends on long calls cannot publish data on how often clients return, because the answer is bad for them.

We are a flat-fee platform with identity-verified readers, and our incentive is the opposite. We benefit when sessions are good enough that clients book again, recommend us, and stay. So we have every reason to publish the data, and now we are going to.

Here is what a year of 58,000 sessions told us.

The Three Fates is the spread that won the year

The most-booked spread on the platform, by a clear margin, is the Three Fates. It accounted for 41% of all sessions. The next closest, the Celtic Cross, came in at 18%. After that the long tail starts: relationship spreads of various shapes at 9%, year-ahead spreads at 7%, custom and reader-designed spreads at 14%, and everything else dispersed across a hundred small layouts.

We had assumed the Celtic Cross would lead. It didn't. We talked to thirty readers about why. The consensus was that the Celtic Cross is the spread the casual reader has heard of, but it's also long, dense, and easy to over-interpret in a 45-minute session. The Three Fates, with its compact past-present-future structure, lets the conversation breathe. It leaves room for the reader to ask follow-up questions, and the client time to actually feel the cards before being asked to interpret six more.

The Three Fates also performs better on what we call session quality, which we measure with a post-session rating and a follow-up question two weeks later about whether the reading was useful. Three Fates sessions had a 4.93 average rating. Celtic Cross sessions had a 4.86. Both are good. The 0.07 gap matters at scale.

The lesson we took from this: shorter spreads, given more room, beat longer spreads done in a hurry. We've started giving readers training that nudges them toward the Three Fates as a default unless the client specifically asks for something else.

The platform-wide rating is 4.91, and we know exactly which sessions pull it down

Every session on BookTarot ends with a five-star rating. The platform-wide average across 58,000 sessions is 4.91. We are not going to brag about that number, because if you average enough sessions on any reading platform the number trends toward five for reasons that have less to do with quality than with rating inflation.

What's more interesting is the distribution. 78% of sessions get a clean five. 15% get a 4 with a written comment that's almost always about something operational rather than the reading itself (audio dropped, link was hard to find, reader was four minutes late). 5% get a 3, and those are usually the sessions where the reader and the client had a mismatch in style or expectations. 2% get a 1 or 2, and we read every one of those.

The 1s and 2s are the data we care about most. We pull those sessions, listen to the recordings where the client consented, and decide whether the reader needs coaching, the platform needs a process change, or the mismatch was something we could have prevented at booking. Last year, 31 sessions led to direct reader coaching. 12 led to product changes (the most important: a pre-session questionnaire that lets clients flag if they're in active crisis or grieving a recent loss, which we now route to readers with specific training). Three readers were removed from the platform.

We are publishing the bad data because we think the only way to be trusted is to show that you look at it. Most platforms only publish the average. The average is the least useful number.

Love and relationships still leads, but career is closing fast

The most-asked specialty on the platform is still love and relationships, at 38% of sessions. This is the number that has barely moved across the year. Love is the question tarot has always been asked.

What has moved, and what surprised us, is career. Career sessions were 19% of our sessions a year ago. They're 27% now. The growth is concentrated in clients aged 28 to 42, and the questions inside those sessions have shifted noticeably. A year ago, career questions were mostly "should I leave my job" or "is this promotion coming." Today, they are more often "am I in the right field at all" or "what do I do with the next ten years when the work I was trained for is changing under me." We see this most acutely with clients in tech, media, design, and law.

The third-largest specialty is grief and life transitions, which has held steady at around 14%. Fourth, growing slowly, is creative work, at 9%. Fifth is family and parenting at 7%. The remaining 5% is a mix of health, spiritual practice, and questions that don't fit a category.

The data has changed how we recruit readers. We've actively brought on more readers with explicit career-coaching backgrounds in the last six months, because the demand is there and a love-specialist reading career questions is not always the right fit.

Brazil, Mexico, and Australia are growing fastest

By country, the United States is still the largest single market for the platform, at around 41% of sessions. The United Kingdom is second at 14%. Canada, Australia, Germany, Brazil, Mexico, the Netherlands, Spain, and France round out the top ten.

The growth rates tell a different story than the totals. The fastest-growing markets, measured as percent growth in sessions over the last twelve months, are Brazil (up 187%), Mexico (up 142%), and Australia (up 96%). Behind them, with smaller absolute numbers but rapid growth, are South Africa, Portugal, and the Philippines.

A few honest notes on this. Brazil and Mexico are growing partly because we added Portuguese-speaking and Spanish-speaking readers in larger numbers this year, and we'd be naive to think the supply isn't part of the demand. But the inverse is also true: we added those readers because the demand was already showing up in our search data and being unmet. The growth is real. We are still under-served on Portuguese, particularly in Brazil, and we are recruiting actively.

Australia is a different case. The growth there has come without a corresponding push from us, and we think it has to do with two things: a strong existing tarot culture in Sydney and Melbourne, and a time-zone advantage that lets Australian clients book sessions with European readers at civilised hours on both ends.

Bookings spike on full moons, equinoxes, new years, Valentine's, and Mercury retrogrades

Seasonal patterns in our data are striking, and not always the patterns you'd expect.

Full moons reliably produce a 22 to 30% spike in bookings in the 48 hours around them. New moons, less so, around 10 to 14%. We did not expect the asymmetry to be that large. Readers tell us the full moon brings the kind of restlessness that drives someone to want a session. The new moon, by contrast, is a more interior energy, and clients are more likely to journal at home than to book a reader.

The equinoxes, both spring and autumn, produce four-to-five-day spikes around the transit, with the autumn equinox being slightly larger than the spring. The solstices have less effect than we expected, although the winter solstice does produce a bookings spike about a week after, which we think correlates with people sitting still during the holidays and finally booking the session they've been postponing.

New Year's week is the largest bookings week of the year by a clear margin, at about 2.4 times our average week. The first two weeks of January are 1.8 times average. The pattern is what you'd expect: people use the new year as a permission slip to make changes, and a reading is part of how they process the changes they're contemplating.

Valentine's Day produces a bookings spike of about 60% in the four days before. About 80% of those bookings are love and relationship questions. Roughly half are from clients in active relationships. The pre-Valentine's reading is, more often than not, a check-in rather than a should-I-stay question. We found that interesting.

Mercury retrograde periods produce a smaller, but measurable, lift of around 14 to 18%. We hold no particular institutional position on astrology, but the data is the data: when Mercury retrograde is named in the cultural conversation, our bookings go up. Whether that's the planet or the conversation is not for us to litigate.

What per-minute platforms can't tell you

We want to make one explicit point. None of the data above can be published, in good faith, by a per-minute platform. The reason is structural.

A per-minute platform's revenue depends on call length. The longer the call, the more they make. This means a per-minute platform has every incentive to avoid publishing data about repeat client behaviour, because what you'd find, if you looked, is that the platforms with the highest per-minute rates have the lowest repeat-client percentages. The clients use them once, sometimes twice, often in a crisis, and then they leave. The platform doesn't lose money on this, because new clients keep arriving. But the model can't survive a public conversation about client outcomes, so the public conversation doesn't happen.

Our model is the opposite. Our readers charge a flat fee per session. Our revenue depends on whether clients come back, refer others, and stay on the platform. So we have every reason to publish whether they do.

For the record: 52% of clients on BookTarot in the last year had at least one repeat session. 23% had three or more. The average client who has had three readings stays on the platform for an average of 19 months. We will publish those numbers, in this form, every year. We will tell you when they go down, too. That's the deal.

What this data has changed about how we run the platform

A few concrete things, in case you've read this far.

We pre-route clients with active crisis flags to readers with explicit training in grief and acute distress, and we no longer let general-practice readers take those bookings without opting in.

We have added Portuguese, Spanish, and Italian-speaking readers in larger numbers, and have a hiring pipeline for Mandarin and Tagalog opening this summer.

We have moved the booking experience to recommend the Three Fates as the default spread for first-time clients, with the Celtic Cross as an explicit second option for clients who want it. The conversion rate on first sessions improved by 11% when we did this. More importantly, the second-session booking rate improved by 9%.

We have started publishing reader-level data on response time, session rating, and repeat-client percentage, on each reader's profile. Some readers were unhappy about this initially. The good ones got over it within a month. The ones who didn't, mostly, were not the readers we wanted to keep.

We do not love every number in this report. The 2% of sessions that got a 1 or 2 rating represent 1,160 readings where someone went to the cards in real need and didn't get what they came for. We are not going to round that off. The number we are most committed to moving next year is that one.

If you read tarot, or you book it, or you work in this industry in any form, we'd like to invite you to publish more of your own data. The whole sector benefits when the inside of the box is visible.