You do not need a research budget to sanity-check a platform. You do need to stop reading the marketing and start doing arithmetic.
Most people choose a trading platform roughly the way they choose a takeaway: a recommendation, an advert, or whatever came up first. That works out fine for dinner. It works out expensively for an account you may hold for a decade.
What follows is a method you can run yourself in an evening. It is adapted from how proper platform testing is done, minus the part where you need a budget for funded accounts.
Step one: what are you actually going to do?
Write down your genuine intended pattern. Which markets, roughly what trade size, how often, and how long you expect to hold. Not the aspirational version, the honest one. Every cost that follows depends on this, and it is why generic comparison tables mislead. A table built on somebody else’s trading pattern is answering somebody else’s question.
Step two: is the provider properly regulated?
Check the FCA register. Then check the firm reference number displayed on the website against the register entry, because clone firms copy genuine reference numbers and substitute their own contact details. Matching both is the point.
If a firm is not authorised, stop. There is nothing further to evaluate, whatever the platform looks like.
Step three: what does your pattern cost, in pounds?
Convert every percentage into money using your own numbers from step one. Most people skip this and it is the step that actually changes decisions.
| Cost line | How to calculate it | Commonly missed? |
| Spread | Typical spread x trades per month x 2 | Yes |
| Commission | Per-trade fee x trades per month | No |
| FX conversion | Percentage x value of non-sterling trades | Yes |
| Overnight financing | Daily rate x average days held x position size | Very |
| Platform and data fees | Monthly cost, annualised | No |
| Withdrawal and inactivity | Per-event charge x likely events | Yes |
Fifteen minutes with your own numbers beats any generic comparison table.
Step four: what happens to their existing customers?
For leveraged products, read the published loss disclosure, the percentage of that firm’s own retail clients who lost money over the reporting window. Every regulated provider must display one and almost nobody reads it before opening an account.
Read it to calibrate rather than to compare. What it establishes is that at no authorised firm do most clients come out ahead.
It is a poor instrument for ranking two firms against each other, because it measures the behaviour of a customer base at least as much as anything the broker does, and a provider with a cautious clientele will post a better number than one with an aggressive clientele without being the better provider.
Step five: can you actually get your money out?
Test the exit before you need it. Deposit a small amount, then withdraw it, and time the round trip. This is precisely the sort of thing surfaced by research sites such as The Investors Centre, which funds live accounts with its own money to test UK trading platforms rather than working from published fee schedules. Of every step here, it is the one that most reliably produces an unpleasant surprise.
Worth being precise about what that evidence covers, including when a reviewer produced it. A withdrawal is one round trip on one day, so a clean result shows the process worked that time. Failures in this area are rarely constant. They cluster around a verification query, a bank holiday, or a first payment out to a newly added account, none of which a single successful test would necessarily have met.
A platform that is easy to fund and awkward to leave is a known pattern, and you would much rather discover it with fifty pounds than with five thousand.
What should you deliberately ignore?
Three things that consume disproportionate attention. App design, which is genuinely pleasant and almost irrelevant to outcomes over years. Welcome offers, which are worth having on a platform you would have chosen anyway and a terrible reason to choose one. And the number of available markets, unless you have a specific plan to use them, because breadth you never touch costs nothing and gains nothing. Attention spent on those is attention not spent on the cost arithmetic, which is the part that compounds.
How do you separate two platforms that look identical?
Run your pattern through both in pounds, then use three tie-breakers in order. First, the withdrawal round trip from step five, because it is the only one of the three you will have measured yourself. Second, whether the provider charges for anything you are likely to need later, such as a transfer out, a paper statement or dormancy. Third, which of the two makes it easier to find out what it charges, since a firm publishing its whole schedule in one place is a firm you can hold to it.
If they still look identical after that, they probably are. Pick one and stop researching. Beyond a point, further comparison is just procrastination with a spreadsheet.
How often should you redo this?
Annually is plenty for most people, and only properly if something has changed – a fee schedule update, a shift in your own trading pattern, or a provider being acquired. Platform pricing does move, and it more often moves against existing customers than new ones, because new customers are the ones being competed for.
Set a calendar reminder rather than relying on noticing. Fee changes arrive by email in language carefully designed not to alarm anybody, and they are extremely easy to miss.
What does a pass actually look like?
Four boxes. FCA-authorised, with a reference number you have matched. A total annual cost you can state in pounds for your own pattern. A loss disclosure you have actually read. And a withdrawal you have tested end to end. Any platform clearing all four is probably fine for you, regardless of where it sits on anybody’s ranking. Any platform failing one of them is worth a second look before you commit, however good the app looks.
Is this worth an evening?
Set it against the cost of being wrong. On a pot you intend to hold for years, the difference between a well-chosen and badly-chosen platform runs into hundreds of pounds a year in fees you will never see itemised. An evening is cheap by comparison, and you only have to do it once.