Real-time, delayed, and end-of-day
Market data is commonly available at three levels of freshness. Real-time data arrives as the market moves, delayed only by the milliseconds required to transmit it. Delayed data is the same stream offset by a fixed amount, so a 15-minute delayed quote reflects the market as it stood fifteen minutes earlier. End-of-day data is a single snapshot per session, usually the closing values, and is the lightest of the three. Each level suits different applications.
- Real-time: updates as the market moves, sub-second transport.
- Delayed: a fixed offset, commonly 15 minutes behind.
- End-of-day: one snapshot per session, typically the close.
Latency versus delay
Latency and delay are distinct. Latency is the unavoidable time required for data to travel from the source to an application, measured in milliseconds and reduced by efficient engineering and nearby infrastructure. A delay is a deliberate policy of withholding information for a fixed period. A real-time feed has latency but no delay. When comparing providers, the two questions to separate are how fast the live path is and whether the data is deliberately offset.
When delayed data is sufficient
Many applications do not require live data. A long-horizon research notebook, a dashboard that refreshes every few minutes, an educational chart, or a public display can use delayed or end-of-day data without any loss of function. Because it is cheaper to serve and subject to fewer distribution constraints, delayed data is the common default for free tiers and read-only displays. If a fifteen-minute-old value would not change a decision, real-time data is generally unnecessary.
When real-time data is required
When timing affects a decision, live data becomes essential. Order entry, execution systems, live alerting, risk monitoring, and any actively watched chart depend on prices that reflect the current market. A stale price in these contexts is not merely less accurate; it can be incorrect, because a decision is made on a value the market has already moved past. For these applications the relevant measure is latency rather than delay, since the data is live by definition.
Real-time and historical on SiftingIO
SiftingIO is designed for real-time delivery. Prices are aggregated on a continuous clock and pushed over WebSocket with sub-100ms median latency from primary regions, so a live chart or alert reflects the market as it moves. The same platform provides historical bars over REST for context. Every market uses one schema and one credential, allowing live streaming and historical data to be combined without separate integrations.