Daily, weekly or monthly: how to choose your DCA frequency

The two settings readers ask me about most are "how much per buy" and "how often". The difference is that with the first, people know exactly what they're torn about; with the second, the worry is vague, just a nagging sense that picking wrong will cost them some return.

Conclusion first: the frequency field barely decides how much you make. What it decides is how much you pay each year in per-transaction costs for the act of buying, and whether you're still buying on this schedule three years from now. Choosing a frequency means choosing something you can sustain, not choosing a return. The sections below follow the order I use when I reply to readers: first what it actually affects, then the cost math, then the overrated idea of "smoothing", and only at the end, when a change is worth making.

The three things frequency actually affects

Think of frequency as "sampling density" and a lot becomes clear at once. With the same annual budget, buying monthly samples 12 prices a year, weekly samples 52 and daily samples 365. The total invested doesn't change and what you buy doesn't change; the only thing that changes is how densely you sample along the timeline.

So what it really affects comes down to three things:

• The per-transaction cost of each buy. The more buys, the more those per-buy charges get multiplied. Of the three, this is the only one you can work out in advance, so that's what the next section calculates.
• How well it matches your cash flow. When money arrives decides which schedule you can actually fund and which one will sooner or later fail to go through.
• How often you touch your account. Every fill is a reminder, a nudge to open the app and take a look. Almost nobody counts this one, but it's often the most expensive of the three.

Notice that return isn't on the list. I've seen plenty of claims along the lines of "weekly earns this much more than monthly". When I run into one, I usually ask just two questions: which period was measured, and does it still hold over a different one? Crypto price paths look different every few years, and shifting the start and end dates by a few months often flips which frequency comes out ahead. Setting a parameter you'll run for many years based on a historical comparison with an unclear method is a bad trade.

One more common misconception while we're here: a higher frequency doesn't make it "more like DCA". DCA means a fixed amount, bought on a set schedule, without looking at the price. Once a month and once a day both meet that definition, and neither is more orthodox. That being so, fill this field in based on your own situation; there's no need to copy anyone else's settings.

Per-transaction costs are the real price of buying often

This is the only section where you'll need to do some arithmetic. Start by separating two kinds of charges, because they treat frequency in completely different ways.

Fees charged as a percentage are neutral to frequency. Suppose a platform charges 0.1% of the trade amount (a made-up figure for the arithmetic, not any platform's real rate). You invest $1,200 a month: in one buy that's 1,200×0.1% = $1.20; split into 4 buys of $300 each, it's 4×300×0.1% = $1.20. Not a cent of difference. So as long as the fee is purely a percentage, how often you buy doesn't change the total you pay.

What isn't neutral is anything charged per transaction. There are three common ones: a minimum fee per trade, a minimum purchase amount per buy, and per-transaction charges on fiat deposit channels. If any one of them applies, the number of buys starts turning directly into money.

Let's run it with hypothetical figures again. Say your annual DCA budget is $14,400 ($1,200 a month), and say the platform charges a flat $1 per buy. Everything in the table below is hypothetical, just to show orders of magnitude; none of it is any platform's actual pricing:

Frequency	Amount per buy (hypothetical)	Buys per year	Total flat costs	Share of annual investment
Monthly	$1,200	12	$12	about 0.08%
Every two weeks	about $554	26	$26	about 0.18%
Weekly	about $277	52	$52	about 0.36%
Daily	about $39	365	$365	about 2.5%

The top row and the bottom row are thirty times apart. But note that this doesn't mean "daily buying always loses out". If that platform's flat per-buy cost is zero, the last two columns are all zeros, and daily buying has no cost disadvantage at all. So the order should be the other way round: first check how the platform you'll use charges right now, then decide the frequency. Minimum purchase amounts, minimum fees and free promotional periods vary a lot between platforms and they change, so don't take my hypothetical numbers as real ones.

There's another threshold that's even easier to overlook: the minimum purchase amount can rule out a high-frequency plan outright. The same $1,200 a month split into daily buys is about $39 a day, and if the platform's per-buy minimum is higher than that, the plan simply can't be set up. Anyone whose per-buy amount is already small should check this first, rather than settling on a frequency and then finding it won't go in. If you do hit this, the frequency is what should give way, not the amount; raising your per-buy amount on the spot just to clear the minimum gets things backwards.

One more cost hides in how the money gets there, and it's tied directly to frequency: whether you fund every buy from a bank card, one deposit at a time, or convert a month's money into a stablecoin first and let the plan draw from that balance. The deposit step in the first approach is often charged per transaction, so however high your frequency, that channel fee repeats that many times. The second approach squeezes deposits down to once a month, so even a very high buying frequency only pays the channel cost once. If you really want to buy at high frequency, you usually need to pair it with the second approach; otherwise whatever "smoothing" you gain gets lost at the deposit step first. Exactly what gets charged and which channels are supported also varies by platform and region, so you'll need to check the current pages yourself.

Buy more often and each buy weighs less, but it's weight that thins out, not risk

First, let's pin down the word "smoothing". Splitting into more buys gives you exactly one thing you can calculate without knowing the market: each buy's share of the year's investment gets smaller. Where your cost basis ends up, and how much less your balance swings, are decided by how prices move over those years, and can't be answered at all without looking at the market. I'll come back to that below.

So 1/n describes weight, not return and not risk: split a year into n equal buys and each one is 1/n of the annual investment. 12 buys, about 8.3% each; 52 buys, about 1.9%; 365 buys, about 0.27%. You can check all three with a calculator, but they only answer "what's the largest amount I have riding on a single day's price". They don't answer anything else.

Look at that weight in three stages. Going from "buy it all at once" to "split into 12", the share riding on any single buy drops from 100% to 8.3%, the steepest fall of the three. From 12 buys to 52, it goes from 8.3% to 1.9%. From 52 to 365, it goes from 1.9% to 0.27%, a difference of just over one percentage point, while the number of buys goes up sevenfold. The further you push the frequency, the less each step lowers the per-buy weight; and as long as your platform charges a flat per-buy fee, that cost rises in a straight line with the number of buys. Given a flat per-buy fee, the two sides will sooner or later meet at a point where going further isn't worth it. Conversely, if fees are purely a percentage of the trade amount, with nothing extra per buy, there is no such point.

While we're at it, a common expectation to correct: a higher frequency won't make you "buy cheaper". It only lowers the chance of putting a big chunk of money in at an extreme price. Averaging out and buying cheap are two different things. Averaging out means your cost sits closer to the average price over the period, and whether that average is high or low depends on how the market moves during the years you're buying, not on whether you buy once a week or once a day. Frequency controls the spread, not the direction.

Which frequency that not-worth-it point lands on depends on the flat per-buy cost from the previous section. If it's close to zero, go as high-frequency as you like; if it isn't zero, monthly or weekly is usually enough. As for going from n = 1 to 12, in other words buying all at once versus in instalments, I covered that in DCA or lump sum? How a beginner should choose; this piece is about everything beyond 12.

One more reminder: there's one thing frequency can't change, and that's the total amount of risk you're carrying. With the same annual budget, however finely you split it, the size of your paper loss in a deep drawdown won't get meaningfully smaller. That number deserves to be worked out before frequency does.

Following your payday works better than following the theoretical optimum

If I could give only one piece of advice, it would be this one: look at when money lands in your account first, then set the buy date.

The reasoning is simple. The failure I see most often isn't picking the wrong frequency; it's a buy date arriving with no money in the account. Miss one buy and you have to go back and top it up by hand; have that go badly twice and you start assuming the whole thing "doesn't work", and then the plan quietly dies. A plan that can run on autopilot for years is always one that lines up with your cash flow.

So the order goes like this. If you're on a fixed monthly salary, monthly buying is the least hassle; set the buy date a day or two after payday, when the money has just arrived and this month's spending hasn't eaten into it yet. If you're paid every two weeks, buy every two weeks. If you're a freelancer working on projects with irregular payments, I'd actually suggest a low frequency and a small per-buy amount, with a buffer left in the account, rather than a high-frequency plan that only goes through when a payment happens to have come in.

The every-two-weeks option has a small trap worth knowing in advance: 26 buys a year don't line up with 12 months, so two months of the year get three buys. That's not a problem in itself, but if you budget monthly, those two months will suddenly feel tight, so be ready for it.

If you buy monthly, there's a detail worth an extra ten seconds too: don't set the buy date at the end of the month, and don't set it on the 1st. The last few days of the month are when a lot of people are paying off credit cards and paying rent, so their balance is at its thinnest; the 1st often comes before the paycheck lands. Some platforms also have their own handling for dates like the 29th to the 31st that don't exist in every month, so check whether they bring the buy forward or push it back. Move the date to a day or two after payday and you sidestep all of this at once.

The other layer is mental accounting. Money taken right after payday feels closer to "never really reached my hands"; money taken mid-month is pulled out of the balance you're actively spending, which makes it easier to borrow for something else, or to just switch the plan off. Same amount, different timing, and sticking with it really does get harder or easier. For how much each buy should be, start with How much should you put into crypto without losing your head?, and while you're at it you can run the worst case through the DCA calculator; whether this money really counts as spare money is covered in What "spare money only" really means.

Daily DCA: one hidden benefit and one hidden cost

The benefit first, and it isn't about return but about execution: buying daily kills the "which day this month should I buy" thought for good.

Most monthly buyers have done this at some point: the buy date is coming up, they glance at the price, think it looks a bit high right now, and wonder about pushing it back a few days. At that moment you're already timing the market, just wearing a DCA costume. Daily buying leaves no room for that. Today's buy is 0.27% of the year, pushing it to tomorrow means nothing, so you simply stop hesitating. It swaps your second-guessing for "no single buy matters". There's a side effect too: if a low balance breaks one day's buy, the effect on the outcome is small enough to ignore, and you won't feel the plan is ruined because one buy was missed.

The cost sits at the other end, and many people only discover it after setting things up: fill notifications will call you back into your account once a day. The whole point of automatic DCA is "set it and leave it", yet daily buying gives you more than three hundred chances a year to glance at a paper loss. For someone who has just started, seeing the same paper loss three hundred-odd times a year is not the same as seeing it twelve times. You don't understand the market any better for it; you're just reminded more often that you're losing money, and people make dumb moves more easily under that kind of reminder. That's the old topic of How not to scare yourself into selling during a crash.

A secondary cost is record-keeping. With thirty times as many transactions as monthly buying, the workload is very real when you later need to reconcile, work out your cost basis or report under local rules (the rules for reporting crypto assets differ from place to place, so go by the rules where you live; I can't give a universal answer on that).

So the conclusion isn't "don't buy daily". It's this: if you're going to buy daily, turn off fill notifications first, then confirm your platform has no flat per-buy fee. Do those two things and daily buying is a pretty good choice; skip them and both of its costs will turn up right on schedule.

When you should change it, and when you just feel like changing it

Once a parameter is set, don't keep fiddling with it; that has always been my position. But "don't keep fiddling" isn't the same as "never touch it". In the following situations, I think a change is warranted:

• Your income rhythm has changed. A new job, a different payday, a switch from salary to project payments: changing the buy schedule to match is how you keep the plan running, not market timing.
• Your per-buy amount is stuck against the purchase minimum. Either lower the frequency and make each buy bigger, or switch to a way of buying with a lower threshold. Don't force it.
• Flat per-buy costs take too big a share. Run your own numbers through the method in the table above, and if they go past the line you can accept, lower the frequency. You set that line yourself; my own habit is to keep it so small I can't be bothered to think about it.
• You just can't keep it going. You keep forgetting to transfer, the balance keeps coming up short, every buy needs a manual rescue: that means this rhythm doesn't fit your life, and it should change.

On the other hand, when any of these reasons comes up, stop before you touch anything:

• Price shows up in your reason for changing. "It's been falling hard lately, so I'll switch to daily and buy a few extra times": that's adding to your position, buying the dip, not changing frequency. Handle it separately, under your rules for adding.
• You actually just want to look at your account less. Turning off notifications solves that; the frequency doesn't need to change.
• You've just read an article saying some frequency is better. Go back to the two questions from the first section: which period was measured, and does it still hold over a different one?

Before you change the frequency, be clear with yourself: am I changing the rhythm, or am I using a rhythm change as cover for timing the market?

If you really are changing it, stick to two rules. Change only one parameter at a time; don't move the amount and the frequency together, or later you won't be able to tell which one made things better or worse. After the change, run at least one full cycle before judging it, which for monthly buying means at least six months. Constantly tweaking parameters is often emotion looking for an outlet in itself: it gives you a reassuring sense of "doing something about it", when DCA works precisely because you don't do anything. As for the reason that genuinely justifies a pause, it has nothing to do with frequency, and I covered it in Should you pause DCA in a bear market?

One more thing for anyone stuck in front of this field right now: this choice matters far less than it feels like it does. Get the amount wrong and you may be forced to give up in a deep drawdown; get the coin wrong and you may not recover for years; but pick a frequency that isn't quite "optimal" and the worst outcome is paying a bit more in per-buy costs, or a rhythm that doesn't match your pay, which you can change back once you notice. It's the easiest parameter in the whole plan to fix, so don't let it hold you up for two weeks before you start. What really costs people time has never been choosing weekly or monthly; it's never getting the first buy in.

One last thing. Once the frequency is set, there are only three things you really need to watch: whether each buy actually went through, whether this money still counts as spare money, and whether your total position has gone past the cap you set yourself. Price isn't on that list, and neither is frequency. The places where the third one tends to go wrong are all in The 5 position-sizing mistakes beginners make most; if you're just getting started, you can pace yourself straight from A beginner's first year: how to build your position up slowly.