Working out the point where fees eat the yield

A friend put a couple of hundred dollars into a product paying 5%, and a year later found the withdrawal fee exceeded everything he had earned. He asked whether he had picked the wrong product. He had not. The size and the cost structure did not match.

Which costs are fixed

Rates and thresholds vary by platform, asset and network; go by what the page shows at the time.

It is the second category that consumes a small balance's return.

Estimating the threshold

1. Estimate a year of earnings: principal × your realistic expected rate, not the page figure — see it said 8%.
2. Total the fixed costs: in once and out once, so at least two.
3. Divide. If a year of earnings is not a comfortable multiple of the fixed costs, it is not worth doing.

Why a multiple and not simply “more”? Because you are also carrying platform risk, price risk and your own attention. Earnings that merely cover the fees mean taking all the risk for nothing.

Putting numbers on it

Hypothetical figures for illustration; no real platform's rates. Fixed costs of $5 in total, a realistic 4% annual rate:

Illustration only; use your platform's real rates and thresholds.

The value is not the specific numbers but the shape: fixed costs dominate at small sizes and become irrelevant rather abruptly above a threshold.

What a small balance can still do

• Move less often. Every round trip is another fixed cost. Shuffling between products is the largest hidden expense for small balances.
• Compound rather than withdraw. Pay the withdrawal fee when you actually need the money.
• Mind minimum thresholds. Some higher-rate tiers have minimum sizes you may not reach.
• Accept that this stage is not about yield. When the balance is small, the room to optimise a rate is much narrower than the room to grow the balance. Not a pleasant sentence, but true.

Why there is no calculator here

This is exactly the sort of question that turns into a form with input boxes, and I think that would be worse than useless: fees change, network congestion can multiply on-chain costs, and withdrawal fees vary hugely across assets and networks. A calculator with baked-in parameters returns a precise answer built from stale inputs.

What this article gives is a method: separate fixed from proportional, estimate a year, do one division. The method does not expire, and you can substitute today's real numbers whenever you like.

The invisible proportional cost

Splitting costs into fixed and proportional leaves one thing out that is neither, quite: the bid-ask spread.

Any time you need to convert — stablecoins into BNB to join a campaign, a reward token back into stablecoins — you cross a spread. It is not itemised like a fee, and it is a real cost.

Indicative magnitudes; actual values change constantly. Glance at the top of the order book before trading.

Folding that in makes the earlier conclusion more conservative: every conversion is a cost, so the path “in, convert, participate, convert back, out” should be costed as five events rather than two.

It also explains something: products you can enter directly with stablecoins tend to deliver closer to their headline, because they skip two conversions. Where two products quote similar rates, the one requiring fewer conversions usually wins, even at a slightly lower nominal rate.

Practically: if the aim is just to earn something on idle stablecoins, use a stablecoin product rather than converting for a higher rate elsewhere. The conversion out and back easily consumes the difference, and adds a stretch of price exposure on top.

The cost that never appears: your time

One more category that never shows up on a statement: the attention you spend.

Moving money between products, comparing rates, watching maturity dates — each takes a quarter of an hour, and they keep taking it. If a balance earns twenty dollars over a year and you spent five hours on it, the implied hourly rate is grim.

A rough self-check:

• If an action's likely gain is less than an hour of your time is worth, skip it;
• If you are comparing platforms over fractions of a percentage point, first convert that fraction into money;
• If a setup needs a weekly check, its hidden cost is higher than you think.

For small balances, “set once and leave it” beats “optimise continuously” almost always, because the first costs nothing in time and the yield gap is minor.

Which stage you are at decides what to optimise

Where to put the attention at each stage. Not financial advice — an observation about where effort pays.

The first row is unwelcome and accurate: at a small balance, the room to improve a rate is much narrower than the room to grow the balance. Time spent researching yield products is often better spent elsewhere.

It is also part of why this site has no yield calculator. A return prediction accurate to two decimal places does not help most people's actual decisions. Knowing which row you are on, and doing that row's work, does.