A direct indexing product can start as a clean client-facing workflow. Over time it becomes an optimization system. Every account accumulates its own tax lots, wash-sale constraints, restricted securities, cash flows, model drift, and client-specific rules.
The constraint set compounds
Tax-loss harvesting, turnover limits, tracking-error control, sector exposure, ESG exclusions, liquidity, and minimum trade sizes interact. Treating each as a separate screen or post-processing rule creates fragile behavior. The constraints need to be solved together.
Scale changes the product
One account can tolerate manual review. Thousands of accounts require deterministic batches, replayable results, failure isolation, and clear audit records. Solver latency becomes a product constraint because operations teams work inside trading windows.
Infrastructure is the moat
The durable advantage is not merely offering customization. It is delivering customization reliably under real market, tax, and operations pressure. That requires an optimization engine built for production workflows, not a single-purpose calculator.
Practical next step: audit your current direct indexing workflow for hidden manual gates. Each manual exception is a signal that optimization infrastructure should become more explicit.