Market Calendar Infrastructure: Inside TradingHours.com with Alan Reed

By: Mel Sutton, TradingHours.com Interview with ALTALLO - Founder of ALTALLO

Published: 2026-02-05 · Read time: 8 min · Category: Founder Spotlight

Why trading hours data is harder than it looks, and what happens when it goes wrong.

[HOST]: At face value, trading hours data feels like something people could just Google. Why doesn't that work in practice?

[GUEST]: You absolutely can collect trading hours yourself if you only need a handful of markets. You run into challenges when you scale to more markets and need the data to stay current throughout the year. Schedules change ad hoc, sometimes with very little notice, and keeping everything up to date becomes a real ongoing effort.

The other issue is structure. Our data is already machine readable and designed to be easy to integrate into existing systems.

[HOST]: What do people most underestimate about this problem when they first hear what you do?

[GUEST]: It feels trivial at first until you realize there are hundreds of markets and trading venues with thousands of unique schedules. Each one can have half days, irregular sessions, seasonal hours, non trading days, non settlement days, working weekends, and special cases layered on top of each other. In some markets there are non trading days that still settle, and others where trading happens but settlement does not.

Holidays and schedules can also change at any time at the discretion of the exchange.

Overnight sessions add a layer of complexity. When a market has an overnight pre-open session (which is common on derivatives markets) you have to be very precise about how holidays on the prior or following day affect that session.

Those are just a few examples. There is much more complexity than most people expect.

[HOST]: What's the messiest edge case you've had to support that outsiders would never think about?

[GUEST]: One of the trickiest edge cases is Islamic holidays, which depend on moon visibility and are often confirmed only the evening before. Different countries or religious authorities may announce the holiday on different days, depending on whether they rely on local sightings, global sightings, or astronomical calculations.

[HOST]: What breaks first when trading hours data is wrong?

[GUEST]: A lot of different things, depending on the use case. Inaccuracies can lead to missed trades, risk not being rebalanced on time, settlement date miscalculations, backtesting inconsistencies, and a whole range of operational headaches.

[HOST]: What do companies usually try first before they become a customer?

[GUEST]: Most start by collecting the data in-house. That works at first, but quickly becomes unmanageable once they need to cover more than a handful of markets or require deep historical data for backtesting. Firms also tend to underestimate how difficult it is to keep up with ad hoc changes to holidays and trading schedules throughout the year.

[HOST]: At what point does maintaining this data in-house stop making sense?

[QUOTE]: Our clients employ a lot of really smart, highly compensated people. It quickly stops being cost-effective to have that high-value time spent collecting and maintaining data that can be affordably and reliably sourced.

[HOST]: Who inside an organisation feels the pain first when this data is wrong?

[GUEST]: Trading desks are close enough to their markets to catch changes to trading hours quickly. The real burden falls on back-office operations and engineering teams, who are responsible for maintaining trading hours data in-house and deal with the fallout when errors inevitably occur.

[HOST]: What's the most expensive or impactful failure you've seen caused by bad market calendar data?

[GUEST]: The most consistent cost is wasted high-value time. Engineers, quants, and ops teams get pulled into fire drills to fix something that should be boring and stable. That mental overhead adds up quickly. For a relatively small subscription cost, that entire class of distraction goes away.

[HOST]: How do you think about correctness when markets themselves publish conflicting or changing information?

[GUEST]: We always rely on primary sources. All of our data comes directly from exchanges. When information is ambiguous or conflicting, we reach out to the exchanges directly to clarify before making an update.

[HOST]: Where do you draw the line between automation and human verification?

[GUEST]: All of our data is collected manually by our research team. We have found that scrapers and AI-based approaches are too error-prone for this kind of work. Where we use automation heavily is in validation. We run consistency checks, anomaly detection, and cross-reference data to catch issues before anything reaches customers.

[HOST]: What are customers really paying for when they subscribe?

[QUOTE]: They are paying to not think about this problem anymore. Accuracy, coverage, history, and confidence that someone is watching the edge cases. The price is low compared to the cost of internal maintenance, missed issues, or constant verification work.

[HOST]: If TradingHours disappeared tomorrow, what would your customers have to rebuild first?

[GUEST]: The API and integration layer would be the hardest part. Collecting raw market hours is one thing. Turning that into clean, versioned, machine-readable data that stays consistent over time is another. On top of that, they would also need to rebuild the research process and the institutional knowledge around how markets actually behave.

[HOST]: What's a belief you had about building a data business that turned out to be wrong?

[GUEST]: I did not even realize I was building a data business at first. I assumed the data would mostly be used as a reference by traders. Something you look up occasionally and move on. The original version of TradingHours was built purely for my own personal use. In practice, the data is used in far more ways than I ever expected.

[HOST]: Why do you think "boring but critical" infrastructure businesses tend to endure?

[QUOTE]: Boring solutions are often the best. Once something works reliably, there is very little motivation to replace it. Stability and data confidence matter more than novelty in this category.

[HOST]: What gets more complicated as global markets become more interconnected?

[GUEST]: As markets become more interconnected, the number of venues and trading schedules increases quickly. Each market has its own rules, holidays, and exceptions, and serving more regions means supporting more of these unique schedules.

[HOST]: What's one problem in financial market data that still isn't solved well?

[GUEST]: One ongoing pain point is the lack of a symbology that maps cleanly to all unique trading schedules we track. MIC codes are useful, but they are not granular enough to represent the full range of trading hours and behaviors that exist in practice.