In 2024, MIT's Centre for Information Systems Research surveyed 259 companies, each averaging $10 billion in revenue, and asked a simple question: does being a real-time business actually matter?
The answer: yes, 62% higher revenue growth and 97% higher profit margins for top-quartile companies. They were also 20% better at innovation and 22% better at operational efficiency. The researchers defined a "real-time business" as one combining digitised operations, empowered employees, and live data access within appropriate governance guardrails. Not just fast dashboards. A fundamentally different relationship between information and action. (MIT CISR, 2024)
Thomas Davenport made the underlying argument back in 2006 in his foundational work on competing through analytics: the companies that win aren't necessarily the biggest or the best-resourced. They're the ones whose decision-making is closest to what's actually happening. Analytics, in Davenport's framing, isn't a reporting function. It's an operational posture.
The problem is that most business infrastructure was built for a world where data was slow. Reports were scheduled because pipelines took time. Weekly cadences became structural. And then those cadences became habits, long after the technical constraint that created them had gone.
This is what makes the lag problem so persistent. It isn't really a data problem. It's an organisational one. Companies fix their data pipelines, invest in modern BI tooling, and then continue holding the same weekly review meetings they always did. The infrastructure changes. The decision rhythm doesn't. And so the gap between what's happening and what leadership can act on stays roughly the same, just better formatted.
There's a secondary effect that's under-appreciated. When data arrives on a schedule, decisions accumulate to match it. Issues that could have been caught on Tuesday get batched to Friday. Interventions that would have taken five minutes become damage-limitation exercises. The cost isn't just the lag. It's the compounding of every decision that was made late, or not made at all, because the signal wasn't visible at the moment it mattered.
Monica Rogati's Data Science Hierarchy of Needs made a related point about foundations: you can't build useful intelligence on top of unreliable or untimely data. Real-time capability isn't a feature you add. It's a property of the underlying architecture.
Geckoboard was built on exactly this premise. The gap between what's happening and what leadership can see is not a reporting problem. It's an infrastructure problem. We built genuinely real-time data connections, not scheduled syncs dressed up as live feeds. The distinction matters more than it sounds.
The MIT research wasn't surprising to us. It confirmed something our customers had already figured out. The competitive advantage isn't the dashboard. It's the time you recover between the moment something changes and the moment someone can act on it.
That gap, it turns out, is worth 97 points of margin.
