A robo-advisor and a do-it-yourself investor can end up owning almost the same handful of index funds. What differs is who builds the mix, who rebalances it, who watches for tax losses and what that service costs each year. A fee of a quarter of a percent sounds trivial, so this comparison turns it into dollars over 20 years and then weighs it against what automation can do for you: rebalance without being asked, harvest losses in a taxable account and keep you invested when markets fall. It also covers the costs that sit outside the headline fee, such as large cash holdings.
How do robo-advisors and DIY investing compare?
The table sets the two routes side by side on cost, the work involved, taxes and control.
| Feature | Robo-advisor | Do it yourself |
|---|---|---|
| Ongoing cost | An advisory fee, a median 0.25% of assets a year in 2025, plus the funds' expense ratios; some charge a flat monthly fee instead | The funds' expense ratios only; index equity and index bond mutual funds averaged 0.05% in 2025, weighted by assets |
| Building the mix | An online questionnaire maps you to one of a set of model portfolios | You choose the stock and bond split and the funds |
| Rebalancing | Automatic, on a schedule or when the mix drifts past a set band | By hand, on a calendar, at a drift threshold or with new contributions |
| Tax-loss harvesting | Automated at many services, in taxable accounts only | Possible by hand, with the wash-sale window to track yourself |
| Cash | Usually up to about 3%, but 6% to 30% in a few programs | Whatever you decide to hold |
| Control | Limited to the firm's models and funds | Full control over funds, timing, tax lots and which account holds what |
| Behavior | Scheduled deposits and rebalancing run without you; human help at some firms | Depends on your own discipline |
| Oversight | A registered investment adviser that files a Form ADV | Your broker is regulated, but nobody is advising you |
Human access varies more than any other feature. Some services offer a planner by phone or video, sometimes only above a minimum balance, a hybrid the SEC's robo-adviser bulletin calls "bionic" advice; others offer technical support and nothing more.
What does a robo-advisor actually do?
A robo-advisor is a registered investment adviser that uses software to choose a portfolio from your questionnaire answers, then invests, rebalances and sometimes harvests losses for you. FINRA's report on digital advice breaks the service into six jobs: profiling the client, setting the asset allocation, choosing the funds, placing trades, rebalancing and tax-loss harvesting. A do-it-yourself investor handles each of those jobs alone, or skips some of them.
The questionnaire is the weakest link. The SEC warns that a recommendation is only as good as the information collected, and that some services ask little about debts, other accounts or a home. When its examiners reviewed robo-advisors, they found some firms building portfolios from just a few questionnaire answers and many not checking periodically whether accounts still suited their clients.
Rebalancing is the job robo-advisors do most reliably. FINRA describes two triggers: a calendar schedule, such as quarterly, or a drift band, such as rebalancing when an asset class is more than 5% away from its target, and notes that each rebalance can create trading costs and, in a taxable account, taxes. You can copy either trigger yourself with a reminder and a few trades, and asset allocation by age works through a rebalancing example. The difference is that software does not hesitate after a crash, which is exactly when rebalancing means buying the asset that just fell.
How much more does a robo-advisor cost?
The typical robo-advisor charges an advisory fee of about 0.25% of assets a year, plus whatever the portfolio's funds charge in expense ratios. Morningstar's 2025 survey of 16 US services put the median advisory fee at 0.25% of assets a year, unchanged from its previous study and well below the roughly 1% that traditional advisers charge.
The range has widened. The most expensive service in that survey cost 1.87% a year on a $15,000 account once a flat membership fee was added, up from a top fee of 0.89% in the previous study. Flat fees fall hardest on small balances: $5 a month is $60 a year, which is 1.2% of a $5,000 account, 0.24% of $25,000 and 0.12% of $50,000.
Then add the fund layer. The SEC's bulletin warns that a robo using high-cost funds can be expensive overall even when its own fee looks low. For comparison, the Investment Company Institute reports asset-weighted average expense ratios in 2025 of 0.05% for index equity and index bond mutual funds, 0.14% for index equity ETFs and 0.09% for index bond ETFs. A do-it-yourself portfolio of broad index funds can therefore cost about 0.05% a year in total, while a robo using similar funds costs 0.25% plus that.
Some revenue never appears as a fee. Morningstar lists interest earned on cash held at an affiliated bank, expenses of in-house funds that are not waived, and payment for order flow as ways firms can earn more from an account. The SEC suggests asking how the firm is paid, whether it uses affiliated funds and whether it pays for referrals.
What does a 0.25% fee cost over 20 years?
On $50,000 plus $500 a month for 20 years, total costs of 0.25% instead of 0.05% a year leave you about $11,750 poorer, under the assumptions below.
Worked example: 20 years of fee drag on $50,000 plus $500 a month
Assumptions: $50,000 invested on day one and $500 added at the end of every month for 240 months, $170,000 in all. The portfolio earns 6% a year before costs, credited monthly at 0.5% a month, and costs are charged monthly at one-twelfth of the yearly rate. Taxes are ignored. The 6% is an illustration, not a forecast. The 0.05% row stands for broad index funds held directly; the 0.25% row for a robo whose total cost, advisory fee included, comes to 0.25%; the 0.30% row for a 0.25% advisory fee on top of 0.05% funds.
| Total yearly cost | Balance after 20 years | Costs paid over 20 years | Shortfall against 0.05% |
|---|---|---|---|
| 0.05% | $393,528 | $1,878 | None |
| 0.25% | $381,775 | $9,208 | $11,753 |
| 0.30% | $378,900 | $10,995 | $14,628 |
The $11,753 shortfall is $7,329 of extra costs plus $4,424 of growth those costs would have earned, and it equals about 23.5 months of $500 contributions. It starts small, $111 after one year and $845 after five, then compounds to $2,664 after 10 years and $6,022 after 15. At an assumed 4% return the 20-year gap is $8,337; at 8% it is $16,663.
The break-even test follows directly. Compared with doing it yourself in similar funds, the robo has to add more than 0.20 percentage points a year, through tax savings, behavior it prevents or features you would otherwise pay for, before it comes out ahead. The compound interest calculator reproduces the table if you enter 5.95% and 5.75% as the rate with monthly compounding.
Is automated tax-loss harvesting worth the fee?
It can be in a taxable account with high tax rates, but harvesting mostly defers tax rather than erasing it, and it does nothing inside an IRA. Harvesting means selling a fund that has fallen below what you paid to realize a capital loss, then buying a similar but not substantially identical fund so the mix barely changes. Realized losses offset realized gains, and net losses can then offset up to $3,000 a year against ordinary income ($1,500 if married filing separately), with any excess carried forward.
The catch is the wash-sale rule. Buy the same or a substantially identical security within 30 days before or after the sale (a 61-day window) and the loss is disallowed for now. IRS Publication 550 also counts purchases by your spouse and purchases inside your own IRA or Roth IRA, so software harvesting in one account cannot see a trade you place in another. The SEC settled a case in 2018 with a robo-advisor that falsely told clients it monitored their accounts to avoid wash sales. Tax-loss harvesting stocks and the wash-sale rule cover the mechanics.
The best long-run evidence is a 2020 Financial Analysts Journal study by Chaudhuri, Burnham and Lo. Simulating harvesting on the 500 largest US stocks from 1926 to 2018 at a 35% short-term and 15% long-term tax rate, it found a benefit of 1.08% a year before trading costs, falling to 0.82% once the wash-sale rule was applied, with the largest gains in volatile, low-return periods. Read that as a ceiling for most people. The simulation harvested among 500 individual stocks, which offers far more chances to book losses than a portfolio of a few index ETFs, and the benefit shrinks at lower tax rates or when you have few gains to offset. Each harvest also lowers the cost basis of the replacement fund, so much of the saving comes back as tax when you eventually sell.
How do cash allocations change the math?
Cash inside a robo portfolio earns a cash rate rather than a market return, so a large cash slice can cost more than the advisory fee it replaces. Morningstar found most robo portfolios hold up to about 3% in cash, but a few programs with no advisory fee hold far more: one holds 6% to 30% depending on the risk level, and another places customers in a 30% cash portfolio by default. The firm can earn a spread on that cash, which is one way a program with no advisory fee gets paid.
Worked example: What a cash slice costs each year
Assume the invested part of the portfolio earns 6% a year and the cash earns either 4% or 1%. The yearly cost is the cash share times the gap between the two returns.
| Cash share of the portfolio | Cost if cash earns 4% | Cost if cash earns 1% |
|---|---|---|
| 6% | 0.12 point a year | 0.30 point a year |
| 10% | 0.20 point a year | 0.50 point a year |
| 30% | 0.60 point a year | 1.50 points a year |
A 30% cash slice earning 1% costs six times the median advisory fee it replaces.
Cash also lowers risk, so the fair comparison is with the mix you would otherwise hold. If you would keep that money in cash anyway, as part of an emergency fund for example, the cost is smaller. With the federal funds target range at 3.75% to 4.00%, about 4% is near the top of what cash pays, and what a given program pays depends on where the cash is swept; cash sweep accounts explained shows how to check.
Which approach helps you stay invested?
The one you will stick with through a bad year, because poor timing has cost the average fund investor more than a typical advisory fee. In Morningstar's Mind the Gap 2026, the average dollar invested in US mutual funds and ETFs earned 8.7% a year over the 10 years to December 31, 2025, against a 9.9% total return for the funds, a gap of about 1.2 points a year. In US stock funds the gap was 12.8% against 13.3%.
Morningstar's suggested fixes read like a robo-advisor's feature list: fewer ad hoc trades, regular automatic investing, fewer and more diversified funds, and automated rebalancing. A robo builds those habits in, and that is its strongest case for a cautious or busy investor. It cannot stop you from switching to a conservative model or withdrawing after a crash, though, and the SEC's bulletin notes that some services may even be able to freeze sales for a period during volatile markets.
Doing it yourself gives you control a robo cannot. You can pick each fund, decide which account holds bonds and which holds stocks, choose the tax lots you sell and add holdings outside a firm's models. You can also set up the same automation: payroll contributions to a 401(k), automatic monthly purchases in an IRA and a written rule for when to rebalance. Lump sum vs dollar-cost averaging and bear markets and corrections cover the decisions that test discipline most.
What are the risks of each approach?
Both routes carry the same market risk; with a robo the added risks are errors, conflicts and costs, and with DIY they are neglect and impulse.
Robo-advisor risks
- A thin profile. A few questionnaire answers can put you in a mix that ignores debts, other accounts or a pension.
- Conflicts. Affiliated funds, cash programs and payment for order flow can pay the firm more than the fee shows.
- Errors. SEC examiners found rebalancing and trading mistakes at some firms.
- Misleading claims. More than half of the firms examined had advertising problems, including suggestions that SIPC covers market losses. SIPC covers $500,000 per customer, including up to $250,000 in cash of missing assets if a member firm fails, not falling prices.
- Restrictions. Some programs limited withdrawals, for example to cash only.
DIY risks
- Behavior. Selling after a drop or piling into last year's winners, with no process to stop you.
- Drift. A mix that is never rebalanced slowly takes on more stock risk than you chose.
- Tax slips. Wash sales across accounts, or rebalancing in a taxable account by selling winners.
- Cost creep. In 2025 the simple average expense ratio of index equity ETFs was 0.45%, against 0.14% weighted by assets, so plenty of index funds cost far more than the popular ones.
- Overlap. Several funds that hold the same large companies add complexity without adding diversification.
Types of investment risk covers the market risks both approaches share.
How do you choose between a robo-advisor and DIY?
Price what you would actually use: if the automation, tax work and help are worth more to you than the extra fee, the robo earns it.
Questions that settle it
- Is the money in a taxable account? Inside an IRA or 401(k), tax-loss harvesting adds nothing.
- How big is the balance? Flat fees weigh most on small accounts, and percentage fees cost the most dollars on large ones.
- Did you keep contributing and rebalancing through the last market drop without a prompt? Your record is the best evidence.
- Do you want a person to call? Check whether that costs extra or needs a minimum balance.
- What does the Form ADV brochure say about fees, conflicts, cash and exit charges, and what does the SEC's adviser database show about the firm?
There is a middle route. A single target-date fund in a retirement account automates rebalancing and the shift toward bonds for one expense ratio, and target-date funds compares how the designs and costs differ. For the do-it-yourself route, index funds explained and how to choose a brokerage account cover the building blocks, and the stocks hub shows where each piece fits.