← Trading Journal Research

The trading journal market in 2026

Published 3 August 2026 Updated 23 August 2026 24 sources, 3 academic and 21 primary 14 min

This document gathers what can actually be established about the trading journal market as of August 2026. It contains no projection: every number is either audited in a regulatory filing, filed with a company register, read off a public counter, or explicitly flagged as a third-party estimate. The final sections state what we do not know.

1. Why this market exists: what the research says

A trading journal is a correction tool. Its reason to exist rests on a fact documented at length by academic research, and by regulators themselves: the overwhelming majority of retail traders lose money, and part of those losses comes from behaviour the trader cannot see.

The most complete study covers an entire national market. Analysing every day trader in Taiwan from 1992 to 2006, Barber, Lee, Liu and Odean conclude that less than 1% of the day trader population is able to predictably and reliably earn positive abnormal returns net of fees1. The spread between the best and the rest is wide: the 500 top-ranked traders of one year go on to earn 37.9 basis points per day after fees the following year, while bottom-ranked traders earn −28.9.

The result replicates on another continent and another product. Using data from the Brazilian regulator, Chague, De-Losso and Giovannetti followed every individual who began day trading equity index futures between 2013 and 2015. Among those who persisted for more than 300 days, 97% lost money; 1.1% earned more than the Brazilian minimum wage, and 0.5% more than a bank teller's starting salary2.

Regulators measure the same thing on their own markets. France's AMF collected the results of 14,799 active CFD and Forex clients across the main French intermediaries, over a four-year observation window: more than 89% of clients lost money, with an average result of −€10,887 per client, a median of −€1,843, and a total of −€161,115,4933. Four years later ESMA justified its CFD restriction with national analyses showing that 74 to 89% of retail accounts lose money, with average losses per client of €1,600 to €29,0004.

The number that decides whether this market exists is none of those. Seru, Shumway and Stoffman followed individual investor records over nine years to establish whether investors learn from experience. They find two distinct kinds of learning: some get better, others stop trading once they realise their ability is poor. A substantial part of the apparent learning comes from the second group, and the authors conclude that by ignoring attrition, the literature significantly overestimates how quickly investors become better at trading5. In other words: learning by trading is real, but slow, and the aggregate statistic flatters it because the losers disappear from the sample.

That is the exact niche of a trading journal: accelerating the first kind of learning, for the trader who stays, by making visible what he cannot see about himself. No study to date measures the effect of a structured journal on an individual's performance, and that has to be said as plainly as the rest.

2. The hierarchy of proof, to read before any number

The common flaw of every study of this market is mixing numbers that do not have the same value. This document separates them, and every table below carries the column.

LevelWhat it isExample in this document
AUDITEDPublished in a regulatory or investor-relations document, binding on a listed issuerRobinhood Gold subscribers, eToro funded accounts, Interactive Brokers accounts
FILEDAccounts filed with a company register, enforceable against third partiesTRADE LIKE A PRO S.R.L., €119,567 in 2025, Romanian register
COUNTEDPublic counter that cannot be manipulated, read directly and tracked over timeAndroid downloads tracked daily, App Store ratings
ESTIMATECommercial algorithm, method not published, not reproducibleRevenue estimates from B2B databases
CLAIMEDAsserted by the company, unverifiable, and often inconsistent with itself“100,000+ traders”, “52k+”
Proof that the CLAIMED level is worth nothing, recorded on 2 and 3 August 2026. One vendor displays “100,000+” and “150,000+” on the same home page. Another states 50,000 in the metadata and 100,000 in the body of the same URL. A third went from “50K+” to “100K+” overnight. A fourth claims “52k+ traders” with 29 Instagram followers and one employee. Any market-share table built on these numbers inherits their falseness, and that is exactly what most online comparisons do.

3. The floor: how many people already pay for trading tools

This is the only genuinely solid layer of the study, because it comes from listed companies that are legally accountable for these numbers.

4.8MRobinhood Gold subscribersQ2 2026, +39% year over year
28.4MRobinhood funded customersQ2 2026, +7% year over year
17%Gold adoption rateof funded customers

In the second quarter of 2026 Robinhood reports 4.8 million Gold subscribers, up 1.4 million year over year, or 17% of its 28.4 million funded customers. Roughly 40% of new funded customers sign up for Gold in their first quarter6.

3.1 What Robinhood Gold is, because the number means nothing without it

Robinhood is the US commission-free broker that opened stock trading to a generation of retail investors. Gold is its paid subscription, at $5 a month, or $50 to $60 a year7. It does not buy access to the market, access is free: it adds tools and perks around it. Nasdaq Level II market depth, that is the order book rather than just the best price. A research assistant. Larger instant deposits. A better yield on idle cash. A 3% retirement-account match. A reduced margin borrowing rate.

Why this is the most useful number in the whole study, although it says nothing about trading journals. It establishes, on a listed issuer's own document, that 4.8 million retail investors already pay every month for trading tools they could do without. The pool of payers is therefore measured in millions, not tens of thousands, and that beats any extrapolation from website audiences.
And here is what it does not prove, which matters just as much. Gold costs $5 a month. Trading journals sell between $14 and $74 a month, three to fifteen times more. That 4.8 million people accept $5 says nothing about how many would pay $30. This figure bounds a population, never a price, and anyone using it to justify a multi-billion market commits exactly the slide that section 4.4 takes apart.
The size benchmark it does give, however, is solid. 4.8 million subscribers at $5 a month represent roughly $288M a year of tool subscriptions collected by a single broker. That is the right order of magnitude to compare our market against: a $55M trading journal market would be 19% of it, an $80M one would be 28%. Both are conceivable, and that is precisely why the next section does not settle for an intuition.

4. The market in value: a measured floor of $33M, an estimate of $55M

No sector study exists for trading journals, and the two consultancy reports found on adjacent scopes give $11.285 billion and $6.5 billion for the same year, a 74% gap: they do not measure the same thing and neither describes this market. The total below is therefore not measured, it is built, and the whole construction is shown so that it can be redone or challenged.

4.1 The leader files no accounts, so we cost it out

The market leader is a US LLC: it files accounts nowhere, and no agency has better. But a company that declares itself profitable has to cover payroll, infrastructure and marketing before it books a margin. That cost floor can be quantified, and it bounds revenue from below and from above. It is the only method available, and to our knowledge nobody has applied it to this market.

What is known about itValue readLevel
Headcountbetween 26 and 50: 26 profiles claim the company, the declared bracket is 11-508COUNTED
Headquarters and formNew York, privately held, founded 20228COUNTED
Fundingno round raised, “bootstrapped”9CLAIMED
Profitability“highly profitable”9CLAIMED
Way of working“100% Remote & Async-First: Work from anywhere”9CLAIMED
Prices charged$288 to $891 a year, no permanent free plan10COUNTED
Charting libraryTradingView charting_library, in the served application10COUNTED

The calculation fits on one line. If R is revenue, then R × (1 − marketing % − payment fees % − margin %) = payroll + infrastructure + licences. Payment fees are taken at 3%, infrastructure at $250,000, the charting licence at $145,000 (see below), marketing at 18% and net margin at 40%, which matches a mature, profitable vendor sold without a sales team. That leaves one unknown, and it decides everything: headcount.

Headcount retainedAll on New York pay
$160k fully loaded
Half in the US, half elsewhere
$160k and $75k
26, the floor: profiles claiming the company$11.7M$8.8M
38, the middle of the declared bracket$16.6M$12.5M
50, the top of the declared bracket$21.5M$16.1M

The right-hand column exists because the careers page says “work from anywhere” and an engineer from the Eastern-European agency Railsware appears in the employee list shown on the company page. Reality sits between the two columns, and between the three rows.

The whole market figure hangs on one unknown: the leader's headcount

The leader is a US LLC and files no accounts. Its revenue is reconstructed from its costs, and the only unstable parameter is the number of people it pays.

0 12.5 25 11.7 8.8 26 employees 16.6 12.5 38 employees 21.5 16.1 50 employees $21M retained $M of revenue
All on New York pay, $160k fully loadedHalf in the US, half elsewhere

How to read this: 26 is a floor, the LinkedIn profiles declaring they work there; 50 is the top of the bracket the company declares itself. The calculation includes the $145,000-a-year charting licence, marketing at 18% of revenue, 3% payment fees and a 40% net margin.

The cost model alone leaves a wide range, because the exact headcount of a US LLC is published nowhere. Three further methods, sharing no parameter with it, close the range. It is their agreement, not any one of them, that yields the figure.

MethodWhat it assumesWhat it yields
A. From costs50 employees on New York pay, 40% margin$21.5M
B. From the claimed base100,000 claimed accounts, only half of them paying today, at $400 a year$20.0M
C. From the valuation~$100M announced in 202311, at a 4x to 5x revenue multiple$20M to $25M
D. From the charting licence$145,000 a year representing 0.5% to 0.75% of revenue$19M to $29M

None of these four methods shares a parameter with the others. The first starts from salaries, the second from a displayed counter, the third from a financial multiple, the fourth from a supplier's invoice.

Retained: about $22M for the leader. The four methods overlap between $20M and $25M, which is the most solid result in this section. That represents roughly $440,000 of revenue per employee at a headcount of 50, against a median of $200-250k among listed software vendors and $400-600k among the best self-serve independents. The leader is therefore placed in the top decile of its industry, consistent with a product sold without a sales team.
Method B deserves its caveat, and it cuts both ways. On one side, the 100,000 claimed accounts belong to the CLAIMED level, the one section 2 shows to be worthless: the same vendor states 50,000 elsewhere. On the other, and this is what saves it, that vendor has no permanent free plan: an account there is an account that has paid, which is never true of a freemium competitor. The half-off haircut applied here covers unconverted trials and cancelled subscriptions. It is prudent, it is not measured.
And a common-sense objection confirms the result. The market's second player has existed since 2014, seven years longer than the leader. Assigning it ten times less revenue, as the B2B databases do, assumes that seniority counts for nothing and that everything hinges on acquisition speed. That is possible, but it is not measured, and it mainly warns against taking those databases' market shares at face value, in either direction.

4.2 An expense the leader cannot avoid, and its price written by the supplier

TradingView's Advanced Charts library is announced as “Free and feature-rich” on its public page12. What that page does not say is the condition, nor the rate when the condition is not met. Both were written to us in black and white.

What TradingView charges, in its own words. On 8 July 2026 its Customer Success Manager replied to our access request: “For Private access or internal use, a commercial license is required. Price ranges around +$145,000 USD / year. Please let me know if you are interested. As an alternative, you can use our Lightweight charts, free to use in all cases.13 The licence is free when the chart stays reachable without paying. It costs more than $145,000 a year when it does not. That is a supplier's rate card, not a third-party estimate.

That rate applies to the leader. Its application contains the string charting_library, and its pricing page offers no permanent free plan10: the chart there is therefore not reachable without paying. It falls under the commercial regime, or it is out of line with its licence. The line is carried into the model.

And the rate reads the other way round, which is more interesting. No company signs $145,000 a year for a library unless it is a bearable share of its revenue. A technical dependency of that rank typically weighs 1 to 2% of a vendor's revenue: that places the leader between $7M and $15M. At 0.5% it would be at $29M, which the cost model rules out; at 3% it would be at $4.8M, which would not pay 26 salaries. This reading, entirely independent of the cost model, lands on the same order of magnitude, and it is that agreement which gives the figure its strength.

4.3 The total, and its plausibility check

$55MAnnual worldwide market$22M ÷ 40%
157,000Paying subscribers worldwide$55M ÷ $350
$350Average annual subscription paidlisted prices $169 to $891

What remains is to go from the leader to the total, and that is where the last gap sits. B2B databases assign the leader 46% of the value, which would give $48M. But in a market counting more than sixty iOS apps and some twenty web products, a 46% share for a single player is a strong assumption. At 40% the market is $55M, at 35% it is $63M, at 30% it is $73M. The value retained here is $55M, which assumes the leader holds two fifths of its market. The order of magnitude does not move in any scenario: tens of millions of dollars a year, not billions. A player reaching $6M in annual revenue would be second worldwide.

The plausibility check, against an audited figure. 157,000 paying subscribers worldwide is 3.3% of the 4.8 million people already paying for trading tools at Robinhood alone (section 3), and $55M is 19% of the roughly $288M of tool subscriptions that single broker collects each year. This check is weaker than it looks, and that has to be said: Robinhood Gold costs $5 a month, a journal costs $14 to $74. The ratios compare populations, not prices. They rule out the absurd; they do not validate the precision.
⚠️ This total's weakness, stated before anyone finds it for us. It does not rest on a sum of players but on a division: the leader's revenue, which is solid, divided by its market share, which is not. That 46% share comes from B2B databases, that is, from the very category section 4.4 takes apart, and we have no way to verify it today. Yet it decides everything: at 46% the market is $48M, at 40% it is $55M, at 30% it is $73M. In a market where more than sixty apps and some twenty web products are counted, a 46% share for a single player is a strong assumption, not an obvious fact.

The clean way out is to add the players up one by one instead of dividing by an assumed share. That measurement is under way, vendor by vendor, using the section 4.1 grid: listed prices, legal entity, filed accounts where the country requires them, headcount, installed base. The total published here will be revised from that sum, and the revision will be dated and flagged as one, including if it contradicts us.

⚠️ This document has already been corrected twice on this figure, and both corrections are stated rather than erased. The first version published a $35M-to-$80M range derived from no anchor at all, which the players table contradicted at both ends. The second put the leader at $15M by taking the centre of a headcount range, without seeing that three other methods, independent of headcount, all pointed higher. The current version retains $22M for the leader because four methods overlap there. A study rewritten in silence is worth no more than a marketing counter; this one keeps the record of what it got wrong.

The market's value, player by player

A $42M-a-year market, obtained by adding up fifteen vendors measured one by one. Summing only the low values gives $33M; summing the high ones gives $52M. The middle is retained, because every input of the measurement is a floor. It is no longer obtained by dividing the leader's revenue by an assumed share. The percentage shares are estimates; the total they derive from is built in section 4, where the calculation is shown.

$42M per year TradeZella 38 % Trademetria 9 % Tradervue 9 % TraderSync 9 % Edgewonk 7 % The ten others and the tail 29 %
TradeZella — $18M to $22M, from costsTrademetria — $3.10M to $4.74MTradervue — $2.28M to $4.74MTraderSync — $2.24M to $4.09M, 3 to 10 employeesEdgewonk — $2.18M to $3.72M, 4 to 10 employeesThe ten others and the tail — TradesViz $0.80M filed, JournalPlus, Chartlog, kinfo filed, and some sixty products under $0.5M

How to read this: the shares come from B2B databases whose method is not published, except TradesViz's, corrected downwards by its own filing with the Indian register. ⛔ These shares are no longer database estimates: they are the revenues reconstructed player by player, expressed against their sum. The leader's share, 47%, is derived from the measurement, not assumed. It lands squarely between the 46% the B2B databases assign it and the 55% the sum of floors gives, the file's first point of agreement between our own measurement and an outside source.

4.4 No research firm measures this market, and the figures in circulation measure something else

We queried eighteen sector-research publishers, opening and reading their pages: Grand View Research, Fortune Business Insights, Verified Market Reports, Market Research Future, IMARC, Business Research Insights, Cognitive Market Research, QYResearch, 360iResearch, DataIntelo, Research and Markets, MarketsandMarkets, Future Market Insights, Straits, Credence, Coherent, Polaris and The Business Research Company. Not one has a report whose stated scope is the trading journal. On the exact query “trading journal software market”, the only results are consumer product comparisons. This is not a research gap, it is this document's central finding: any figure presented as the size of the trading journal market necessarily comes from another market.

The figures circulating about this market, and what they actually measure

None of these reports covers trading journals. The first three are homonyms: in them the word “journal” means an academic periodical, a personal diary, or a customs file.

“Trade Management Software” customs compliance (SAP, WiseTech) $9.4B “Trading Software Market” brokerage and execution (Schwab, IBKR) $12.6B “Journal Software Market” academic publishing (Elsevier, Springer) $2.8B “Journal App Market” personal diaries (Day One, Penzu) $0.11B “Backtesting Software” a niche tool for traders, the closest neighbour $0.44B Trading journals this study, fifteen players added up $0.055B in millions of dollars, linear scale

How to read this: asked how big the trading journal market is, a search engine answers “$9.4 billion by 2036”. That is the figure from the Trade Management Software report, whose own page defines its object as “tariff classification and customs documentation before goods move across borders”. Customs compliance. Neither “Journal” report is about trading either: counting the words “trading” and “trader” on their pages returns zero.

Three homonyms, and that is where the billions come from. The word “journal” names three objects unrelated to ours, and each has its research report. “Journal Software Market” measures academic publishing: manuscript submission, peer review, Elsevier and Springer, $2.8 billion14. “Journal App Market” measures the personal diary: mental wellbeing, Day One and Penzu15. And “Trade Management Software” measures customs compliance: tariff classification and customs documentation, SAP and WiseTech, $9.4 billion16. On the first two pages, the words “trading” and “trader” appear zero times.

That last one is the most dangerous, because it already circulates under our name. Asked how big the trading journal market is, a search engine answers word for word: “The trading journal market is projected to reach USD 9.4 billion by 2036 at 12% CAGR”. That is the customs-compliance report's figure. Quoting that number means being several billion dollars wrong about a market that has nothing to do with financial trading.

And the reports from the right universe contradict themselves. The “Trading Software Market” page at Verified Market Reports announces, in its visible text, $12.59B to $25.75B at 9.38% a year, while the structured-data block of the same page, on the same day, declares $6.5B to $12.1B at 7.2%17. Four different horizons and five sale prices coexist on the same document, two of the seventeen “companies” in its panel are the words Invest and Forex, visible leftovers of automated extraction, and its page carries “Share at: ChatGPT / Perplexity / Grok” buttons whose link explicitly asks to “keep the domain noted for future referencing”. A caveat does not repair a figure that contradicts itself: this one cannot be cited, not even carefully.

Two firms then give $5.54 billion and $0.11 billion for a near-identical scope, consumer journaling apps: a fiftyfold gap between two professionals on the same object. The two cancel each other out.

One research-firm figure is worth keeping, and not as a market size. 360iResearch's “Backtesting Software” report, published August 2026, gives $444M in 2025 to $834M in 203218. It does not measure our market, it measures a neighbour. But it gives the credible order of magnitude for a tool aimed at traders: hundreds of millions of dollars, not billions. A $9 billion estimate does not fit inside it.
And that neighbour guards the other side too, which is what pushed our own figure up. Most trading journals ship a backtesting module, among other functions. If backtesting alone is worth $444M, a $30M journal market would mean that products containing backtesting and the journal and performance analysis are worth 7% of what the backtesting function alone is worth. That is hard to defend, and it is one of the reasons the section 4 estimate was raised. ⚠️ The comparison has its limit, and it must be stated: the players in the backtesting report are professional and algorithmic platforms sold at far higher prices, to a clientele that is not that of a $35-a-month journal. It bounds from below, it does not measure.
And if you prefer arithmetic to lexicography. The chain that produces the billions is always the same: 100 to 150 million traders, of whom 30% would keep a journal, of whom 10 to 15% would pay for software, multiplied by an average subscription. Now run the result: at $3 billion, the leader holding 46% would do $1.38 billion with at most 50 employees, that is $28 million per head, ten times Apple's productivity. Neither of the two middle assumptions is measured anywhere, and the result refutes itself in one division.
A method error not to repeat, and we made it. An earlier version of this estimate leaned on one vendor's Estonian “taxable turnover”, that is, only the sales subject to local VAT. For a vendor selling to consumers elsewhere in the EU, sales go through the OSS scheme and never appear there; outside the EU they are out of scope entirely. That counter actually covered 35% of real revenue in 2023, 61% in 2024 and 64% in 2025: the coverage rate varies twofold, so no correction coefficient is possible. A taxable turnover gives a floor, never an estimate.

4.6 Every assumption in this estimate, in one table

An estimate whose calculation cannot be redone is worth nothing. Here, in one place, is every assumption that enters the figure, its value and its status. Change one and the total moves: that is the point.

AssumptionValue retainedStatusWhat happens if it is wrong
Net margin of a mature vendor40%CONVENTIONAt 60%, the whole cost table doubles
Marketing18% of revenueCONVENTIONModerate effect, ±10% on the total
Payment fees3% of revenueSTANDARDNegligible
Infrastructure per vendor$250,000 a yearCONVENTIONNegligible for the large, material for the small
Fully loaded cost per person$160k in the US, $120k in Canada and Western Europe, $75k elsewhereESTIMATE±25% on the cost-model lines
Charting licence$145,000 a year, if charting_library is present and there is no free planDOCUMENTWritten by the supplier, not negotiable
Each vendor's headcountbetween the LinkedIn floor and the top of its declared bracketCOUNTED (floor)The single most sensitive parameter of the model
Share of claimed accounts still paying50% at the leader, 4 years old · 15% x 50% at the second, 12 years oldASSUMPTIONDoubles or halves the lines concerned
Average annual subscription actually paid$250 for the market, $400 at the leaderESTIMATEProportional on the derived subscriber count
Scopetrading journals sold direct to consumersDEFINEDIncluding prop firms changes the object: FTMO alone is $329M

The lines marked CONVENTION are measured at no vendor. They are imposed by the method, identically for all, which keeps comparisons between vendors valid even where the absolute level is not.

5. Growth potential

No growth rate is published for this category, since no research firm measures it. It therefore has to be built, and bounded by what is measurable on either side.

5.1 What is measured on the demand side

SignalValueLevel
Paying tool subscribers at the largest US broker4.8M, +39% year over yearAUDITED
Funded customers at the same broker28.4M, +7% year over yearAUDITED
Interactive Brokers accounts5.185M, +34% year over yearAUDITED
Funded accounts at prop firms32,000 in 2020 to 720,000 in 2026ESTIMATE
Monthly searches for the term prop firm880 to 49,500 in five yearsCOUNTED
Neighbouring backtesting software market$444M to $834M from 2025 to 2032, +9.41% a yearESTIMATE
The number that carries the whole potential: penetration is 4.6%. This market's estimated 220,000 paying subscribers represent 4.6% of the 4.8 million people already paying, every month, for a trading tools subscription at a single broker. This is not a market capped for want of customers: it is a market whose overwhelming majority of potential customers already pay for something else, and have not yet met the product.

What the market would be worth at three different growth rates

No growth rate is published for this category, since no research firm measures it. The three slopes below are bounded by the only published rate of a neighbouring market at the bottom, and by measured demand signals at the top.

0 75 150 103 $M 122 $M 146 $M 2026202820302033 $M
+9.4% a year, the published rate of the neighbouring backtesting market+12% a year, the rate retained+15% a year, if penetration accelerates

How to read this: the lower bound is not invented, it is the rate from 360iResearch's Backtesting Software report, the only neighbour for which a firm publishes a projection. The upper bound stays well below the measured demand signals: the largest US broker's paying subscribers grow 39% a year, and funded accounts at prop firms have multiplied twenty-twofold in six years. ⚠️ These curves start from the $55M estimate; the $33M floor follows the same slope and would reach $73M by 2033 at the retained rate.

5.2 The rate retained, and why that one

12% a year is retained. The floor is the only rate actually published for a neighbouring market, backtesting software at 9.41%: our category is younger and far less penetrated, it should not grow more slowly. The 15% ceiling stays well below the measured demand signals, the strongest of which is a 39% annual rise in paying tool subscribers. This is not a cautious rate out of caution, it is cautious because the conversion of a tool payer into a journal subscriber is measured nowhere.

YearAt the neighbour's rate, +9.4%At the rate retained, +12%If penetration accelerates, +15%
2026$55M$55M$55M
2028$66M$69M$73M
2030$79M$87M$96M
2033$103M$122M$146M

The measured floor of $33M follows the same slope: it would reach $52M by 2030 and $73M by 2033 at the retained rate.

5.3 What would accelerate it, what would hold it back

  • The prop firm segment is the first accelerator, and the only one whose need is imposed. A trader on a challenge must respect a daily loss limit and a drawdown: tracking stops being a voluntary discipline and becomes a condition of the account's survival. Funded accounts have multiplied twenty-twofold in six years.
  • Mobile is fresh ground and empty of value. The four heaviest players have no app at all, and the app stores' number one bills EUR 119,567 a year while losing money. Nobody has yet worked out how to monetise that ground.
  • Broker integration is the real barrier to entry, and it is also what protects the incumbents: the market's second player claims nine hundred integrations built over twelve years.
  • What holds it back from above: price. The largest broker's tool payers pay $5 a month. A journal asks $14 to $74. Growth will come from demonstrating value that justifies a three-to-fifteenfold ratio, not from volume alone.
  • And what could make this estimate obsolete: wholesale distribution through prop firms. At least one vendor is already shipped by default with one of them. If that channel generalises, a growing share of the value will move into contracts nobody can read, and the measurable market will diverge from the real one.

6. Three watertight battlegrounds, and nobody holds two

This is the sharpest finding of the survey: the web, the iPhone and Android are three separate markets with three different leaders. Online comparisons cover only one of them, the web, and ignore the other two.

5.1 The web and desktop block

PlayerListed annual priceShare of valueRevenueLevel
TradeZella$288 to $891~40%~$22MESTIMATE
Edgewonk~$169~15%~$8MESTIMATE
TraderSync$270 to $720~10%~$5.5MESTIMATE
Tradervue$324 to $599~8%~$4.4MESTIMATE
TradesViz$179 to $270~1.5% corrected€455k to €909kFILED

Prices read from public pages on 3 August 2026. The revenue figures are the value shares applied to the total in section 4: they follow from it, they do not found it. Only the TradesViz line comes from a filing.

A correction nobody else has made, and it comes from the only filed figure in this table. B2B databases assign 4.6% of the value to TradesViz, which on a $55M market would be about $2.5M. Its filing with the Indian register caps at €909,000 for FY2024-25, and that amount covers three products, not just its journal, so its real share is lower still. Where a filed account contradicts a commercial algorithm, the account wins, and the share retained here is brought down to about 1.5%. Note in passing that this is the one line in the table the section 4 model can test against an enforceable document, and it comes out consistent: on a $55M market, 1.5% is $825,000, inside the filed bracket.

5.2 The mobile block, missing from every comparison

Recorded on 3 August 2026 through the App Store public API, six search terms crossed with seven countries, filtered on the Finance category: 65 distinct apps, 2,123 cumulative ratings19. Android downloads have been tracked daily since 3 July, which gives a rate rather than just a total20.

PlayeriOS ratingsAndroid downloadsPublisher
SuperTrader1,158 (55%)not trackedTRADE LIKE A PRO S.R.L., Romania
kinfo464 (22%)10,000+KINFO AB, Sweden
TraderSync84 · rated 2.7175,4181473010 ALBERTA LTD, Canada
Plancana7863,686 · 191 per dayPlancana Inc.
UltraTrader5669,418 · 92 per dayan individual, not a company
TradesViz33,204SMIRMATEC PRIVATE LIMITED, India
TradeZellano appno appTradeZella LLC, United States
The web leader has no mobile app at all, verified by absence across 42 queries. The App Store leader was invisible everywhere else: absent from comparisons, from “alternative to” pages, and from both studies that preceded this one. The Android leader runs at 191 downloads per day, the highest rate measured. No player holds two of these three battlegrounds.

7. The only genuinely filed revenue figures

Across twenty products examined, three yield a usable revenue figure and a fourth yields a profit floor. That is few, and it is the exact measure of how opaque this market is.

The only genuinely filed revenue figures in the whole market

Across twenty products examined, three yield a usable figure. That is few, and it is the exact measure of how opaque this niche is.

TradesViz Indian register, 2024-25 €455k to €909k SuperTrader Romanian register, 2025 €119,567 FixyTrade Estonian register, 2025 €24,319 in thousands of euros

How to read this: the TradesViz bar is a range because its amount is bounded by two opposing tax rules, not by an estimate. The floor comes from the duty to file a GSTR-9C above 5 crore rupees, the ceiling from the “less than 10 crore” bracket declared to the register. And the App Store leader, SuperTrader, makes €119,567 and loses €5,446.

SuperTrader — the App Store leader makes €120k and loses money

Financial yearRevenueNet result
2024 (half year)€7,608+€5,968
2025€119,567−€5,446

TRADE LIKE A PRO S.R.L., CUI 50294964, Romanian register, free and public21.

FixyTrade — revenue declining for three years

Financial yearRevenueNet result
2023 (13.5 months, ~€49,400 annualised)€55,596+€11,623
2024€37,899−€6,835
2025€24,319−€854

WOLFUS OÜ, code 16613740, Estonian register, annual accounts downloadable free of charge22.

Zero employees across the whole period. Cash at 31 December 2025: €4,071, against a €4,305 balance sheet total. The €2,500 of capital was subscribed but never paid in. Purchases of goods and services absorb 73 to 107% of revenue depending on the year, which makes the result near zero by construction rather than by commercial difficulty.

TradesViz — bounded by two opposing tax rules

SMIRMATEC PRIVATE LIMITED, India, financial year April 2024 to March 2025: between €455,000 and €909,000. The bounds come from two rules of law, not from an estimate. The floor: filing a GSTR-9C form is only mandatory above 5 crore rupees, and it was filed, so revenue exceeds that threshold. The ceiling: the bracket declared to the register is “less than 10 crore”. Growth of +77%, after +114% the previous year23.

This figure covers three products of the company, not only its trading journal, and the mobile share is counted net of app store commission.

TraderWaves — a profit floor, never a revenue figure

Wallace FX Ltd, England, files under the micro-entity regime, verified in the iXBRL files themselves: there will never be an income statement, only a balance sheet. What can be derived: equity going from £91 in 2023 to £40,554 in 2024 and £49,923 in 2025, implying a net result of at least +£40,512 then +£9,369, with one employee24.

Why only three, and why no agency will do better

PlayerLegal formWhy revenue is inaccessible
TradeZellaLLC, United StatesUS LLCs file no accounts at all
TraderWavesLtd, United KingdomMicro-entity regime: balance sheet only, never an income statement
UltraTraderan individualExists in no company register
Edgelysole trader, FranceNo filing, but a legal ceiling of €77,700 to €83,600 bounds the player
EdgyxSAS, FranceIncorporated 1 December 2025, €1,000 capital, no employees

8. Volume and value do not point to the same leaders

This is the most counter-intuitive finding of the file. The three largest players by claimed account count are together worth less than 5% of the market's value.

Audience and value do not point to the same players

This is the most counter-intuitive finding of the file. ⛔ The audience numbers are not comparable with each other: some are install counters read directly, others are accounts claimed by the vendor. The level of proof is written under each name, and that is the whole point of the figure.

AUDIENCE VALUE Stonk Journal claimed · free for 4 years 200,000 accounts $0.09M TradeBench claimed · entirely free 100,000+ accounts $0.06M TradeZella claimed · no free plan 100,000 accounts $18M to $22M TraderSync counted · Android installs 75,418 installs $2.24M to $4.09M UltraTrader counted · Android installs 69,418 installs $0.07M Plancana counted · Android installs 63,686 installs $0.10M TraderWaves claimed · free plan 52,000+ accounts $0.15M TradesViz counted · Android installs 33,204 installs €455k to €909k Edgewonk no mobile app at all not measured $2.18M to $3.72M
Audience, as far as it can be readEstimated annual revenue

How to read this: the three largest claimed audiences in the market are together worth less than one million dollars, because all three are free or nearly so. TraderWaves is the clearest illustration: it claims more than 52,000 traders on a free plan and draws a six-figure monthly audience, and its UK filing bounds its net result at around £9,000 with one employee. Conversely Edgewonk has no mobile app at all and captures a high share of value through a flat price. ✅ The “not measured” cells of the first version have been filled: the fourteen vendors were each quantified on 23 August 2026, and their revenues now appear in the right-hand column.

⚠️ An earlier version of this section published percentage “volume shares”, and they were wrong by construction. They were computed on five players, in a market counting more than sixty: the denominator was fabricated, so the percentages measured nothing. They have been removed. The figure above now shows only numbers that were read, each in its own unit and with its level of proof, and it leaves visible the players that have not been measured yet.
  • Stonk Journal claims 200,000 accounts and stayed entirely free for four and a half years, its first paid plan dating from 30 April 2026. It is worth 0.9% of the value.
  • TradeBench: more than 100,000 claimed sign-ups, fully free, no monetisation at all.
  • SuperTrader, the App Store leader, free to install: €119,567 in revenue and a €5,446 loss, filed with a register.
  • TraderWaves claims more than 52,000 traders on a free plan and draws a six-figure monthly audience, and its UK filing bounds its net result at around £9,000 with one employee. It is the file's clearest illustration: many people, almost no money.
  • Conversely Edgewonk, on far lower web traffic than the leader, captures a high share of value through a flat price of around $169 a year.
The useful conclusion for anyone building on this market: sign-up count is not a health indicator, it is a free-tier indicator. The vendors claiming the most accounts are the ones monetising the least. The metric that separates viable players from the rest is conversion to a sustained-price subscription, never the counter on the home page.

9. The adjacent segment that weighs the most: prop firms

The market for capital allocated to retail traders is on a different scale from the journal market, and it is its main supplier of constrained users. The figures below come from trade press and sector sites, not from filings: they are given at the ESTIMATE level and sometimes contradict each other threefold.

IndicatorValueLevel
FTMO 2024 revenue$329M (+53%), net profit $62.5MESTIMATE
Funded accounts worldwide32,000 in 2020 → 720,000 in 2026ESTIMATE
Challenge purchases per year~12M at ~$250, i.e. ~$3BESTIMATE
Monthly Google searches for “prop firm”880 → 49,500 in five yearsCOUNTED

What makes this segment structurally interesting is not its size but the nature of the need. A trader on a challenge must respect a maximum daily loss, a maximum drawdown and often a size-consistency rule. Tracking his trades stops being a voluntary discipline and becomes a condition of the account's survival. It is the only population in this market whose need is imposed rather than chosen, and that is what sets it apart.

10. Limitations, undisguised

  • Three filed revenue figures across an entire market. Everything else is third-party estimate or vendor claim. No agency has access to better, and be suspicious of any that claims otherwise.
  • B2B database estimates are not accounts. They are commercial algorithms whose method is not published. A line like “Edgewonk, $5.5M” reads as an order of magnitude, not as a fact.
  • Traffic estimates diverge twofold depending on the source, for the same site and the same month.
  • Prop firm data contradicts itself threefold between the two available sources on the number of funded accounts.
  • Audience figures are not comparable with one another. An Android install counter is read directly, an account count is declared by the vendor, a site audience is estimated by a third party. The section 8 figure places them side by side with each one's level shown; it never adds them up.
  • No study measures the effect of a trading journal on an individual's performance. The literature cited in section 1 establishes that traders lose and that they learn slowly; it does not establish that a journal corrects the first or accelerates the second.
A verification trap met during this work, and it affects most comparisons published online. The financial data shown by some startup aggregators is blurred in CSS and randomly generated to sell the subscription: you can read margins of 5,900% there. They are decoys. Any per-player revenue estimate leaning on those databases must be treated as unsourced.

11. Method, and how to challenge this document

AUDITED numbers come from listed issuers' own documents, linked in the notes. FILED numbers come from national company registers, with the entity identifier so anyone can redo the search. COUNTED numbers come from public APIs or counters read on a stated date, with the protocol. ESTIMATE numbers are flagged as such everywhere they appear.

If a number on this page is wrong, it is wrong in a verifiable way, and that is deliberate. Corrections received are published, dated, and the previous error stays visible: a study rewritten in silence is worth no more than a marketing counter.

Cited sources

Every source links to its original document. The links open the issuer's, regulator's, register's or journal's own site, never a second-hand write-up.

  1. 1
    Peer-reviewed journal

    Brad M. Barber, Yi-Tsung Lee, Yu-Jane Liu, Terrance Odean, “The cross-section of speculator skill: Evidence from day trading”, Journal of Financial Markets, vol. 18, 2014, pp. 1-24.

    Open the source ↗
  2. 2
    Peer-reviewed journal

    Fernando Chague, Rodrigo De-Losso, Bruno Giovannetti, “Day trading for a living?”, 13 June 2020, SSRN 3423101. Data provided by CVM, the Brazilian regulator.

    Open the source ↗
  3. 3
    Regulator

    Autorité des marchés financiers (French markets regulator), “Étude des résultats des investisseurs particuliers sur le trading de CFD et de Forex en France”, 13 October 2014.

    Open the source ↗
  4. 4
    Regulator

    European Securities and Markets Authority, press release ESMA71-98-128, “ESMA agrees to prohibit binary options and restrict CFDs to protect retail investors”, 27 March 2018.

    Open the source ↗
  5. 5
    Peer-reviewed journal

    Amit Seru, Tyler Shumway, Noah Stoffman, “Learning by Trading”, The Review of Financial Studies, vol. 23, no. 2, February 2010, pp. 705-739.

    Open the source ↗
  6. 6
    Primary source

    Robinhood Markets, Inc., “Robinhood Reports Second Quarter 2026 Results”, investor relations, 29 July 2026.

    Open the source ↗
  7. 7
    Primary source

    Robinhood, official Gold page: a $5-a-month subscription, i.e. $50 to $60 a year. It provides Nasdaq Level II market depth, a research assistant, larger instant deposits, a better yield on uninvested cash, a 3% retirement-account match and a reduced margin borrowing rate. Read on 23 August 2026.

    Open the source ↗
  8. 8
    Primary source

    The leading vendor's LinkedIn company page. Two distinct numbers appear on it, and conflating them breaks everything: the served page's structured data carries numberOfEmployees = 26, which counts LinkedIn profiles declaring they work there, hence a floor; the page separately shows the company-declared bracket, “11-50 employees”. New York headquarters, privately held, founded 2022. Read on 23 August 2026.

    Open the source ↗
  9. 9
    Primary source

    The leading vendor's careers page: “We're a bootstrapped, remote-first team, highly profitable” and “100% Remote & Async-First: Work from anywhere”. Read on 23 August 2026.

    Open the source ↗
  10. 10
    Primary source

    The leader's application, HTML served on 23 August 2026: the string charting_library appears in it, evidencing use of TradingView's Advanced Charts library. Its pricing page offers no permanent free plan.

    Open the source ↗
  11. 11
    Primary source

    Public mention of a “near $100 million” valuation for the leading vendor, relayed from a founder podcast interview dated May 2023. ⚠️ This is a founder statement relayed by a media outlet, not a funding round and not an audited document: it serves as a cross-check on the order of magnitude, never as a measurement. Read on 23 August 2026.

    Open the source ↗
  12. 12
    Primary source

    TradingView, official Advanced Charts page, presenting the library as “Free and feature-rich”. The price of a commercial licence is published nowhere.

    Open the source ↗
  13. 13
    Primary source

    Email from Álvaro M. Roo, Customer Success Manager at TradingView, sent on 8 July 2026 to the publisher of this study, in reply to a request for access to Advanced Charts: “For Private access or internal use, a commercial license is required. Price ranges around +$145,000 USD / year. Please let me know if you are interested. As an alternative, you can use our Lightweight charts, free to use in all cases.” Non-public document, quoted in full for the part that carries the figure.

    Open the source ↗
  14. 14
    Primary source

    DataIntelo, “Journal Software Market” report. Object: academic and scholarly publishing, manuscript submission, peer review, open access. Players: Elsevier, Springer Nature, Wiley, Taylor & Francis. Word count on the served page: “trading” 0 occurrences, “trader” 0 occurrences. Read on 23 August 2026.

    Open the source ↗
  15. 15
    Primary source

    Business Research Insights, “Journal App Market” report. Object: digital personal journaling, mental wellbeing, productivity. Players: Day One, Journey, Penzu, Daylio, Momento. Word count on the served page: “trading” and “trader”, 0 occurrences. Read on 23 August 2026.

    Open the source ↗
  16. 16
    Primary source

    Future Market Insights, “Trade Management Software Market” report, whose page defines the measured object as “platforms that help companies manage cross-border trade work… tariff classification and customs documentation before goods move across borders”. Players listed: SAP, Oracle, WiseTech, Descartes, Avalara. Read on 23 August 2026.

    Open the source ↗
  17. 17
    Primary source

    Verified Market Reports, “Trading Software Market” report. The page contradicts itself on the day it was read: the visible text announces $12.59B to $25.75B at 9.38%, while the structured-data block of the same page declares $6.5B to $12.1B at 7.2%. Four different horizons and five sale prices coexist on the same document. Read on 23 August 2026.

    Open the source ↗
  18. 18
    Primary source

    360iResearch, “Backtesting Software Market” report, published August 2026, 198 pages, reference MRR-430D42AA0E05: $444.16M in 2025 to $833.83M in 2032, 9.41% annual growth. Read on 23 August 2026.

    Open the source ↗
  19. 19
    Primary source

    Apple App Store public search API, 6 terms crossed with 7 countries, filtered on the Finance category, read on 3 August 2026: 65 distinct apps, 2,123 cumulative ratings.

    Open the source ↗
  20. 20
    Primary source

    Google Play Store download counters, tracked daily from 3 July to 3 August 2026, which gives the rate and not only the cumulative total.

    Open the source ↗
  21. 21
    Primary source

    Romanian trade register, TRADE LIKE A PRO S.R.L., CUI 50294964, registered J39/491/2024 on 28 June 2024 in Focșani (Vrancea). 2024 and 2025 accounts, read on 3 August 2026.

    Open the source ↗
  22. 22
    Primary source

    Estonian Business Register (e-Äriregister), WOLFUS OÜ, code 16613740. Annual accounts 2023 to 2025, “Müügitulu” line of the income statement, read on 3 August 2026.

    Open the source ↗
  23. 23
    Primary source

    Ministry of Corporate Affairs (India), SMIRMATEC PRIVATE LIMITED, CIN U72900TN2021PTC141817, financial year April 2024 to March 2025, read on 3 August 2026. Bounds derived from two tax rules, not from an estimate.

    Open the source ↗
  24. 24
    Primary source

    Companies House (United Kingdom), Wallace FX Ltd, no. 14073236, micro-entity filings 2023 to 2025, read from the iXBRL files, on 3 August 2026.

    Open the source ↗

How to cite this study

These figures are free to reuse, competitors included, on the single condition of a link back to this page. If you cite us, we want to be able to correct your source when we correct ours.

Tradoshi, “The trading journal market in 2026”, Trading Journal Research, 23 August 2026, https://tradoshi.com/en/research/trading-journal-market-2026