You want more than rebalancing—now prove it
The pitch usually lands the same way: “We do more than rebalance.” It sounds like an upgrade until the first statement arrives and nothing obvious looks different—except the advisory fee. In a taxable account, the real question isn’t whether the platform can trade more often; it’s whether those trades show up as a higher after-tax ending value after you account for bid-ask spread, cash drag, and the times harvesting simply isn’t available. If the provider can’t tell you what has to be true in your account for the “more” to appear, you’re being asked to pay for potential, not outcomes.
So the burden of proof shifts to mechanics: what specific levers will change your realized taxes this year, what frictions will offset them, and what you must give up in tracking and simplicity to get there. Before any feature list, it helps to anchor on the portfolio you already have.
Start with the baseline ETF portfolio reality
The baseline is annoyingly effective: a small set of broad, low-cost ETFs that rarely trade, throw off mostly qualified dividends, and keep capital gains distributions modest. In a six-figure taxable brokerage account, that “boring” mix already has two advantages that are easy to forget when a dashboard promises optimization. First, the internal rebalancing mechanics are mostly invisible to you—ETFs can meet redemptions in kind, which often reduces the need to realize gains inside the fund. Second, you control when you realize the big tax event: selling.
That baseline also comes with constraints that don’t show up in marketing screenshots. If you’re contributing regularly, rebalancing can be handled with cash flows instead of selling appreciated positions. If you’re sitting on embedded gains, the ETF portfolio can become a one-way door: you can change the future allocation with new money, but “fixing” the past means paying tax. This is the reference point. Any direct-indexing fee, extra trading, or customization only matters relative to how tax-efficient your current ETF behavior already is.
Tax-loss harvesting: the promise meets your constraints
Once you accept the ETF baseline as the benchmark, tax-loss harvesting becomes the first place direct indexing has to earn its keep. The promise is clean: realize losses opportunistically, bank them, and use them to offset current or future gains. The constraint is messier: harvesting only happens when there are losses to take, and the account has to be allowed to trade around them. In long, steady markets, there may be fewer harvest windows than the marketing implies, especially after a big run-up where most lots are green.
The mechanics also collide with real life. Frequent contributions can accidentally trigger wash sales if you’re buying “substantially identical” exposure in another account or through automatic purchases, and some platforms respond by dialing back trades rather than risking disallowed losses. Short-term losses can be valuable, but harvesting them means you were recently down, and the replacement basket may introduce slight drift while staying out of wash-sale territory. Meanwhile, if you don’t have realized gains for years, the benefit may sit as a carryforward, not cash in hand.
Customization feels empowering until tracking error appears
After the harvesting discussion, the next temptation is “make it yours”: drop a few stocks you dislike, apply an ESG screen, cap a sector, or keep your employer’s industry lighter. It feels low-risk because the portfolio still shows hundreds of names and still uses index language. The first constraint is that every exclusion forces a substitution, and substitutions aren’t free. The platform either concentrates the remaining names, leans harder on correlated stocks, or holds a small cash sleeve while it searches for compliant replacements that won’t trigger wash sales.
That’s where tracking error stops being an abstract chart. A tight screen can shift dividend yield, sector weights, and factor exposure, which changes both pretax returns and the tax profile you actually receive. There’s also timing friction: if you’re bringing in a legacy position with embedded gains, the “custom” target may be implemented slowly to avoid realizing gains, so you live in a half-transition state for months or years. At that point, the right question isn’t “can I customize,” but “how much drift am I paying for, and when does the platform force me to realize gains to get back?”
Factor tilts and active bets: know what you bought
Once you’re already tolerating some drift for screens and transition pacing, factor tilts can slide in almost unnoticed. “Quality” or “value” reads like a style preference, but in practice it’s an active bet layered on top of the index: systematic overweights and underweights that will show up as periods of underperformance you have to sit through. The constraint is behavioral and timing-based—if you abandon the tilt after a bad stretch, you’ve paid extra fees and taken extra tracking risk without getting the long-horizon payoff the tilt was designed to deliver.
Mechanically, the tilt also changes your tax profile. Some tilts lean toward higher dividend yield, which can raise annual taxable income even if total return is similar. Others require more turnover to maintain exposures, which increases trading friction and can create more short-term gains if the platform is aggressive. Before accepting a “smart” overlay, pin down the benchmark (plain index or tilted index), expected tracking error range, and what turnover looks like in a normal quarter versus a volatile one.
Operational risk is where outcomes quietly leak
After screens and tilts, the next leak isn’t philosophical—it’s operational. The platform is now running hundreds of tax lots, replacement baskets, and constraints that can collide with market micro-frictions. A few basis points of bid-ask spread across frequent small trades, a persistent cash buffer to handle withdrawals or “no-trade” lists, or delayed reinvestment around corporate actions can quietly eat the harvest you thought you were buying. If the system is conservative about wash-sale risk, it may simply skip trades in the very moments volatility creates the biggest opportunities.
Then there’s the reporting surface area. Wrong or inconsistent cost-basis method, lots that don’t transfer cleanly in-kind, or mismatched 1099 detail can turn “tax management” into a time cost at filing. At this point, the question shifts from “does it harvest?” to “how often does the machinery fail, and who notices first?”
Your revised scorecard for choosing a provider
By now the scorecard looks less like a feature checklist and more like a set of failure tests. Ask for a plain-English estimate of expected annual tax alpha in your bracket, but also the conditions: contribution cadence, embedded gains, and what happens in a flat year when losses are scarce. Get their typical cash drag range, trading frequency, and how they measure and report tracking error after your exact screens and tilts.
Then price the hidden work. What triggers wash-sale “do not trade” behavior across accounts, who monitors it, and how often trades are skipped? How clean are in-kind transfers and lot-level reporting at tax time? Finally, demand a comparison against a simple ETF benchmark net of all fees and frictions—if they can’t show that, treat the “more than rebalancing” as a cost center.
