OptionMetrics has launched IvyDB TradeFlow, an extension of its IvyDB historical options database. The vendor's announcement, as carried by AOL.ca, says the product is "giving institutions insights on buying and selling order flow, retail trading behavior, [and] participation in 0DTE options trading."

The release positions TradeFlow as a response to structural shifts in US equity options, where zero-days-to-expiration contracts now account for a measurable share of listed volume. OptionMetrics frames the product as a tool for desks that need to separate institutional from retail flow before sizing, routing, or hedging an options order.

What does the database actually contain?

According to the announcement, TradeFlow layers three fields onto the existing IvyDB option chain:

• Order-flow direction, indicating whether each print initiated as a buy or a sell
• Retail trading behavior, a segmentation of contracts tied to retail participation patterns
• 0DTE participation, covering same-day expiration series that IvyDB users have historically reconstructed manually from end-of-day OPRA files

The public materials do not disclose the data source, refresh cadence, or the methodology for separating retail from institutional flow. The release does not specify whether the directional flag comes from trade-side inference, exchange-provided signs, or a proprietary combination of both.

How would a desk use it?

For a buy-side execution desk, the workflow sits upstream of the order router. A portfolio manager running an intraday options overlay can query TradeFlow for prior 0DTE volume, the buy-sell balance, and the retail share before setting the spread, single-leg, or roll order. The desk times the execution against that baseline rather than against a vendor's stated statistics.

Sell-side market makers face a different question. A retail-heavy strike typically carries a different adverse-selection profile than an institutional one, because the flow that follows a customer fill may be more likely informed at the underlying level. Knowing which series carry retail skew on a given session informs delta-hedge ratios and quote adjustments in the listed book.

The 0DTE element matters because intraday expiration options have reshaped the listed-options volume mix over the past three years. Intraday gamma now factors into index rebalancing and end-of-day dealer positioning. A dataset that isolates 0DTE volume and order-flow direction gives a desk a measurement tool it can run against its own market impact.

What the announcement asserts versus what the desk can verify

OptionMetrics markets TradeFlow as an institutional tool. The release does not include back-test results, classification hit rates, or correlation between the directional flag and subsequent price movement. A buy-side user evaluating the product will need to ask the vendor for that performance data before any execution policy is tied to it.

New datasets carry integration cost. Feeds must be ingested, validated against existing IvyDB keys, and mapped into the OMS or analytics stack. The cost of that build is part of the procurement decision and rarely appears in the vendor's announcement.

Key questions for OptionMetrics

• What is the latency between exchange print and database update — intraday or end-of-day?
• How does the vendor classify a retail order: by account type, ticket size, or a behavioral model?
• Are 0DTE fields kept as a separate series, or merged into the standard IvyDB chain?
• What is the licensing model: per-seat, per-asset, or per-query?
• Does the dataset cover US options only, or extend to international listings?

Forward outlook

OptionMetrics has not yet released methodology or validation statistics. If the vendor publishes them, TradeFlow would address a documented gap: institutional desks have used end-of-day OPRA files for years, but few commercial vendors have packaged directional order-flow flags, retail segmentation, and 0DTE series into a single historical database. Until those validation numbers appear, the launch is a product announcement, not yet a measurable edge.