profit margins... from fee's.. HA, good one

(be aware you may read many LN promoters that enthusiastically promote profits, purely because in real life they are not making nor interested in profit via fee's.. but they just want to recruit you to get some adoption stat growth, usually they are an affiliate of a major hub service and will want you to channel connect to the hub service, which is where they make their profits)

now lets move onto some math to show how much effort is needed,
imagine blockchain users only paid 250sat tx onchain fee*.. meaning you paid 1000sat to lock 4 channels and would need to spend 1000sat to unlock them later
(lock=join LN: create channel..unlock=leave LN: close channel)
(*im using low number for demo..)

to make it worthy for you, you will need to process:
 atleast 2000 payments charging 1sat just to break even.
 atleast 20000 payments charging 0.1sat just to break even.

also worth noting, apart from exchange-exchange reserve swaps between partners (high volume high value) many users are not part of a popular path to get used as routes that often. yep the non service to service paths do not get much traffic. and those popular paths dont want competition/extra user route hops complicating their path.. so goodluck connecting to certain other partners to create a good popular route

also if you had users trying to use your route and were transferring just $20 of value. 00020500sat
if all routes were heading down one route(direction) your partner (your incoming balance) would need 0.41btc (you rent a inbound channel with 0.41btc allocated)
then you own channel towards destination would need atleast 0.41btc for that path to perform the 2000payments

also to note eventually you may need to pay something yourself to rebalance your channels(pay in opposite direction) which would fill balance back to your side and deplete balance to you on the other side to get you back to start position to perform more payments
(or you could just close channels and reopen them costing you a onchain fee)

some would say you can do it with less funds upfront and just rebalance more often, but then its costing you a fee per rebalance, so you will need to calculate the timing,cost, of doing such too often/not often enough

its also worth noting the more someone wants to on average route through you (more then $20) means the math of how much you need to lock up or how often you need to rebalance is a math calculation you have to work out

to note.
if its low value movement (under $20) then the fee $0.001 becomes an factor
EG
$20 value(20500sat) $0.001 fee(1sat) is fee= 0.005% of value
$2 value(20500sat) $0.001 fee(1sat) is fee= 0.05% of value

you can try to push your luck and be greedy by charging a 10sat fee to reduce number of payments needed to break even/profit
EG
$20 value(20500sat) $0.01 fee(10sat) is fee= 0.05% of value
$2 value(20500sat) $0.01 fee(10sat) is fee= 0.5% of value

but then other routes with cheaper fee win the route and no one chooses your route. you can try to be generous and lower the fee, but then you would need to process more payments to break even

........
now the crux of the issue
bitcoins blockchain fee is usually 10-20sat per byte and with a 250byte average tx. the fee to lock and unlock 4 channels wont be just 2000sat
but
at a 10sat per byte onchain fee :2500sat * 4 =10000sat to lock&unlock (join and leave LN)
at a 20sat per byte onchain fee :5000sat * 4 =20000sat
which increases the amount of other peoples payments you have to achieve via them routing through you just to break even

..
and then.. and then. finally you can maybe, probably start thinking about "profit"
by which i mean if you are charging a 1sat ln fee.. well at 1sat=$0.001 means you need to do 300,000 payments a week just to equal a $300 a week income, so you better partner with popular services and be on the best route path of very regular customers of those services which then means increasing your locked value to  enable you to route them or end up in a bottleneck/rebalance often scenario.

basically do all the math of the 2000 payment scenario first explained at the top. the multiply it all by 150x just to get a average us min wage income
enjoy