The Differences Between Real-Time, Delayed, and Historical Data
Market data comes in three forms depending on when it is delivered: real-time, delayed, and historical. Each serves a different purpose, and choosing the wrong type for a given use case produces either unnecessary cost or unreliable results.
Real-time data feeds
Real-time data arrives within milliseconds of a trade being executed on an exchange. Prices, volumes, and order book changes are pushed continuously to the recipient without intentional delay.
This is the standard for active trading, algorithmic systems, and any application where a stale price produces a wrong decision. Intraday traders, futures traders, and automated strategies all depend on real-time feeds. For tick-level granularity — every individual price change with its exact timestamp — see tick data.
Real-time data costs more than delayed or end-of-day data because exchanges charge licensing fees for immediate access. Those fees vary by exchange and asset class.
Delayed data feeds
Delayed data is published with an intentional lag, typically 15 minutes. Exchanges charge lower licensing fees for delayed redistribution, which is why many free or low-cost data sources use it.
The 15-minute gap is large enough to make delayed data unsuitable for intraday trading. A signal based on a price that is 15 minutes old no longer reflects the current market. For general market monitoring, educational purposes, or research that does not depend on current prices, delayed feeds are a cost-effective option.
Historical data feeds
Historical data covers past prices, volumes, and related statistics over a defined period. It is the basis for backtesting trading strategies, conducting research across market cycles, and training models on past market behaviour.
The key quality criteria are completeness, correct adjustments for corporate actions (splits, dividends), and the length of the available archive. A strategy tested only on two or three years of data misses major volatility regimes — bear markets, financial crises, sharp recoveries — that affect whether the strategy holds up in practice.
Historical and real-time data are usually separate products. Confirm the available depth before committing to a provider if backtesting is part of your workflow.
Which feed suits which use case
The right choice follows from the time horizon of the decision being made.
Active intraday and algorithmic trading requires real-time data. A 15-minute lag invalidates intraday signals, and end-of-day data provides no intraday information at all. For a direct comparison of real-time and end-of-day approaches, see real-time vs. end-of-day.
Swing trading and portfolio management based on daily bars can work with end-of-day data. The extra cost of a real-time feed adds nothing when positions are held for days or weeks and reviewed after the close.
Backtesting and quantitative research need historical data at the appropriate granularity. Strategies that trade intraday need intraday history; strategies on daily bars need daily history going back far enough to cover meaningful market conditions.
Cost is a secondary factor but not irrelevant. Real-time feeds for multiple exchanges can add up quickly. Paying for real-time data on markets you do not actively trade is unnecessary. The market data pricing page shows available packages broken down by exchange.
TAI-PAN data feeds
TAI-PAN, provided by Lenz + Partner, covers all three feed types: real-time data via push technology, end-of-day records from 1987, and intraday histories from 2002. Coverage includes stocks, futures, forex, and indices across major exchanges. Subscriptions are modular — you select the exchange packages you need. A 30-day trial is available for €1.
