Published Fees vs What You Actually Pay: Why Two Comparison Sites Disagree
Same platform, two websites, different costs. Usually one of them read a pricing page and the other opened an account.
If you have compared UK trading platforms across more than one website, you will have noticed the numbers do not always match. It is rarely carelessness. There are two entirely different methods for working out what a platform costs, and they produce different answers.
One method is free, the other costs real money
Read the provider’s published fee schedule and tabulate it. Fast, free, and accurate about everything it covers. The catch sits in those last three words. A schedule is a document a firm writes about itself, so it draws the boundary of its own subject: which charges get a line and which get a clause, which are quoted as percentages and which in pounds, and what conditions have to hold before the number printed is the number you pay.
Dishonesty does not come into it. The document is a description written by an interested party in that party’s preferred units, and a comparison assembled only from schedules inherits every one of those choices without ever noticing it has.
Open an account, deposit real money, trade with it, and record what is actually charged. Slow, expensive, and it picks up the things a schedule has no reason to mention. That is the approach behind independent platform testing that opens and funds live accounts rather than compiling rankings from published fee schedules.
The output reads differently from the first line. Costs arrive as cash amounts rather than percentages. Timings arrive as elapsed days rather than as target ranges. Conditions arrive as things that happened on a particular Tuesday afternoon rather than as clauses that may apply.
What does testing find that a fee schedule misses?
| Cost | In the published schedule? | Found by funded testing |
| Headline commission | Yes | Same |
| Spread at the moment you trade | Sometimes, as ‘from X’ | What you actually got |
| FX conversion on a real order | Stated as a percentage | The rounded cash amount |
| Withdrawal timing | Vaguely, if at all | In days, dated |
| Inactivity fees and when they start | Buried in terms | From a statement |
| Execution during volatile spells | Never | Observable |
Six cost lines, set out as a published schedule presents them and as a funded account records them. Nothing here is a price list: the point is which of the two documents has anything to say at all, and about what.
Does this matter if you are investing small amounts?
Less than for an active trader, but more than nothing. If you buy a few hundred pounds of US shares each month, currency conversion is probably your single largest cost and the one most likely to be understated by a percentage figure you never convert into money.
Why do providers not just publish the real numbers?
Partly because some genuinely vary. A spread is not a fixed price; it moves with the market, so a provider quoting ‘from 0.6 points’ is being accurate rather than evasive. What you get depends on when you trade.
Partly, though, because vagueness is commercially useful. Conversion charges expressed as a percentage sound smaller than the same charge expressed in pounds on a typical order. Inactivity fees described in clause 14.3 attract fewer questions than inactivity fees on the pricing page. None of it is concealment exactly. It is emphasis.
What is the biggest single discrepancy to watch?
Currency conversion, almost always. It is the cost most often understated by percentage-only presentation, it applies to the markets British investors trade most, and it is applied inside the exchange rate rather than itemised, so it never appears as a fee anywhere on a statement.
If you reconcile only one number between a published schedule and what actually happened, make it that one. There is a worked version of the arithmetic at the end of this piece.
How fast does a funded observation stop being true?
Faster than the article reporting it, which is the standing weakness of the expensive method and belongs in the middle of a piece that otherwise recommends it. A fee schedule can change on thirty days’ notice. A spread changes in a second. The moment a tester writes down what a withdrawal cost, the observation starts ageing, and having paid for it does nothing to slow that down.
What survives is narrower than the headline suggests. A dated observation proves that a gap can open on that particular line, and it tells you which lines are worth pricing yourself before you deposit. The figure has a use-by date. The demonstration does not.
Which should you trust?
Both, for different jobs. A fee-schedule comparison is a perfectly good shortlisting tool. A funded-account comparison is what you want in front of you at the point of deciding where the money actually goes.
How do you check the conversion charge yourself?
One order is enough and the arithmetic takes two minutes. Find a filled order in a non-sterling market on your own statement. Note the exchange rate the platform applied and the pound amount that actually left your account. Then look up the mid-market rate for that pair at the time stamp on the fill.
Divide the rate you were given by the mid-market rate at that moment, subtract one, and you have the conversion charge as a percentage of the trade, measured rather than quoted. Multiply that by the pound value of the order and you have it in money, which is the only version comparable with everything else on the statement.
Then run the same sum on the sale, because the charge usually applies in both directions and a percentage quoted once describes a cost you pay twice. If your measured figure matches the published one, the schedule was telling the truth in a form you could not use. If it does not, you have found the reason two comparison sites disagree about the same platform.