The wallet wants $14 to make a $60 swap. That quote stops the trade cold, and it should. Paying close to a quarter of a small trade just to move it is not a fee, it is a donation. Gas fees frustrate nearly everyone who uses Ethereum, yet they follow clear rules once the parts are laid out. Gas is payment for computation, the price swings for known reasons, and most of what a typical user pays comes from timing, network choice, and a few wallet settings that take seconds to change. Seven working steps cut the cost down to size.
What gas actually is
Every action on Ethereum, sending coins, swapping tokens, minting, or approving a contract, asks the network to do work. Gas measures that work. Each kind of action uses a set number of gas units: a plain transfer uses 21,000 units, while a complex swap can use several times that. Units alone do not set the bill. Each unit is priced in gwei, a tiny slice of ether, and the total fee is the units used multiplied by the price per unit.
That split matters because both halves move. A simple action can still cost a lot when the price per unit spikes, and a complex action can cost pennies when the price is low and a cheaper network handles it. Wallet screens that show one dollar figure hide that split. Seeing the two halves separately is what makes the savings steps below make sense, and the mechanics are laid out clearly at Ethereum.org’s explainers on gas, fees and Layer-2 networks for anyone who wants the technical version.
Why the price spikes when it does
Block space is limited, and getting a transaction included works like an auction. When many people want in at once, they bid the gas price up to jump ahead, and everyone else must match the higher price or wait. Market panics, popular token launches, and busy weekday hours all pack the auction at the same time. During peak mania, simple swaps have cost $50 to $100 or more.
Quiet periods run the other way. Late nights, early mornings, and weekends often clear at a small fraction of the peak price, sometimes under a dollar for the same swap that cost $14 on a weekday afternoon. Nothing about the transaction changed. Only the crowd did. Paying less is mostly about not bidding against that crowd, and about doing the work somewhere cheaper when that option exists.
A worked example shows how the two halves combine. A swap using 150,000 units at a quiet price of 10 gwei costs 1,500,000 gwei in total, a small fraction of one ether. Push the price to 80 gwei during a busy hour and the identical swap, with the identical amount of network work, costs eight times as much. The code did not change. The demand did. That is why nearly every saving below attacks either the price paid per unit, the number of separate transactions, or the network doing the work, rather than the swap itself.
Way one: time transactions for quiet hours
Shift non-urgent transactions to off-peak windows. Overnight hours in the United States and weekend mornings regularly show the lowest prices of the week. A swap that can wait twelve hours often costs half or less for the patience. Free gas trackers show live prices and daily patterns, so checking the chart before sending takes less than a minute and quickly becomes habit. Urgent trades still go through at once. Routine ones rarely need to.
Way two: use Layer-2 networks
Layer-2 networks process transactions away from the Ethereum main chain, bundle many of them together, and settle the bundle back to Ethereum for security. The shared cost is what makes each individual transaction cheap, often a few cents against dollars on mainnet. Most major apps now offer a Layer-2 version that works the same way from the user’s side. If the token, app, and funds are available there, doing the swap on a Layer-2 is the single largest saving most users can make, and it stacks with every other step on this list.
Moving funds onto a Layer-2 does involve one bridge transaction, and that bridge pays a mainnet fee once. The math still favors the switch for anyone making more than a trade or two, because the bridge cost is paid a single time while the savings repeat on every swap after it. Funds that will sit untouched for a year can stay where they are. Funds that move weekly belong where moving them is cheap.
Way three: take the standard or slow option
Wallets default to fast settings that overpay for speed most transactions do not need. Switching to the standard or slow option before confirming often cuts the fee by 30 to 50 percent, at the cost of waiting a few extra minutes for confirmation. For a swap with no deadline, those minutes cost nothing. Advanced settings allow a custom price set just above the current base level for users comfortable watching a tracker, but the built-in slow option captures most of the saving with none of the guesswork.
Way four: batch related actions
Each separate on-chain step pays its own fee, so splitting one job across several sessions multiplies the cost. Approving a token on Monday, swapping on Wednesday, and moving the result on Friday pays three rounds of fees at three different prices. Doing the same steps back to back in one session pays once per step at one price, and planning a week’s trades into a single sitting can remove entire transactions from the list. Batching also reduces exposure to a price spike landing in the middle of a half-finished plan.
Way five: mind the approval tax
Before a contract can move a token, the wallet must approve that access, and the approval is itself a gas-paying transaction. Careless permission habits double the cost. Approving a token, swapping, revoking the permission for safety, then approving the same token again next week pays the approval fee twice for one trade. A steadier approach helps: approve tokens actually used on a regular basis, revoke permissions for apps no longer touched, and avoid approving and revoking the same contract in a loop. Some wallets now set limited approvals by default, which caps risk without paying for a fresh approval on every visit.
Way six: apply the percentage rule
Judge every fee against the size of the move, not as a bare dollar figure. If the quoted fee passes 5 to 10 percent of the amount moving, stop. Wait for a quieter window, switch to a Layer-2, or combine the move with other transfers until the percentage makes sense. The rule cuts both ways. A $15 fee on a $20,000 transfer is 0.075 percent, a rounding error that is not worth an hour of waiting to avoid. A $20 fee to move $50 is 40 percent gone before the trade starts, and no small swap justifies that. The percentage, not the headline number, tells the truth about whether a fee is high.
Way seven: track personal patterns for a month
Keep a simple note of each transaction for one month: the day, the time, the network, and the fee paid. Patterns appear fast. Trades made on weekday afternoons will usually top the list, weekend Layer-2 swaps will sit at the bottom, and one or two habits will account for most of the money spent. After a month, the cheap windows stop being luck and become the default schedule. That record also pairs naturally with the cost side of other purchases, such as how to buy crypto with a credit card without high fees, since total cost is what reaches the bank account.
Two warnings worth the space
A failed transaction still burns its gas. The network did the computation, reached a failing result, and charges for the work either way. Checking balances, allowances, and network choice before sending prevents the most common failures, and sending a small test first is cheap insurance before a large move. Guessing and resending can burn the same fee two or three times.
Fees are also sometimes worth paying in full. A transaction that is urgent, large, or locked to a deadline, such as a loan repayment that avoids a penalty or a time-sensitive purchase, should go at the fast setting without regret. Saving $8 while risking a $200 penalty is not thrift. The goal is paying the right price for each job, low for routine moves and full fare when minutes genuinely matter.
Put together, the seven steps rarely need to run at full strength on the same transaction. Timing alone fixes most weekday overpayments. A Layer-2 alone fixes most small-trade fees. The percentage rule alone stops the worst quotes from being accepted at all. The rest is polish that adds up over a year of regular use, turning a fee that once ate a quarter of a small swap into a minor, predictable cost of doing business on the network.
One last habit closes the loop. Gas paid around trades belongs in the same records as the trades themselves, with dates, amounts, and the purpose of each transaction. Fees tied to buying and selling affect the cost basis of a position, and clean records turn tax season from an archeology dig into a short filing task. Small notes kept at the time are worth far more than a wallet history reconstructed months later from memory.
