Yagnesh RamalingamTopic
Decision latency
Start here if a simple question takes weeks to become a number anyone will act on.
The only analytics metric that has ever mattered is time from question to trusted, actioned answer — and almost nobody measures it. Dashboards, warehouses and AI copilots are inputs to that path. They are not the path.
The explainer
Decision latency. In practice.
A warehouse can return a query in seconds and the decision can still take two weeks. The missing time is not the report. It is the argument about which report to trust, and the wait for someone who will sign the number. That interval is the one a leadership team actually feels, and almost nobody writes it down.
It is not reporting lag
Reporting lag is how long a pipeline takes to land. Decision latency includes the rest of the path: who interpreted the question, which definition they used, who owns that definition, and when someone with authority acted. A two-minute dashboard can still sit on an eleven-day decision. The longer piece is you do not have a data problem.
| Reporting lag | Decision latency | |
|---|---|---|
| What it measures | How long a pipeline takes to land a number | How long until someone acts on a number they trust |
| Starts | When the source data changes | When the question is asked out loud |
| Stops | When the dashboard refreshes | When someone with authority acts |
| Who owns it | The data team | The person who owns the meeting |
| Usual fix | Faster pipelines | A written definition, a named owner and a sign-off |
How to measure it
Pick five questions the leadership team already asks. Not hypotheticals. Timestamp the moment each one is asked out loud, and the moment someone acts on an answer they will defend. The interval is the metric. One question is an anecdote. Five is a baseline. The method is written out in time five questions, and the 90-minute version is the decision latency audit. The full guide, with nine metrics for the stages in between, is how to diagnose and measure decision latency.
Where the weeks actually go
- Interpreting the ask. The sentence in the meeting and the query that gets written are rarely the same question.
- The definition. Two teams can each defend a number because nobody wrote down what it means.
- Ownership. Without a named person who can refuse a change, every version stays in play.
- Sign-off. The number exists. Nobody will put their name on it, so the meeting decides on instinct.
Query speed is the step people instrument, and it is rarely the long one. The weeks hide in the other four.
What the unmeasured time costs
Questions that never resolve do not disappear. They come back next month, slightly different, and someone rebuilds the argument from scratch. That backlog is answer debt. It compounds the way technical debt does: every unresolved definition makes the next question slower.
There is no published benchmark for a good decision latency, and inventing one would miss the point. The comparison that matters is your own five questions, against each other and against last quarter. The slow one tells you whether the constraint is the data, the definition, or who owns it.
Articles
Articles on decision latency
12 min read
Decision latencyDecision Latency at 200 People vs 20,000: What Actually Changes
Small companies wait for missing evidence and scarce experts. Large enterprises wait through competing definitions, handoffs, and unclear authority.
Read full article
12 min read
Decision latencyDecision Latency vs Decision Intelligence: Where the Categories Overlap
Decision intelligence designs and improves the choice process. Decision latency shows where that process waits before action.
Read full article
13 min read
Decision latencyHow to Measure Decision Latency: 9 Metrics That Actually Move
One governing measure and eight diagnostic metrics that show whether slow decisions come from access, trust, ownership or execution.
Read full article
12 min read
Decision latencyThe Decision Latency Scorecard: A One-Page Diagnostic
Five questions, four timestamps, one primary delay, and one repair owner. The scorecard stays useful because it fits on one page.
Read full article
7 min read
Agentic analyticsEmbedded Intelligence: Why Businesses Don’t Need Another AI Tool
Most businesses don’t need another AI tool. They need trusted answers delivered where work already happens. Learn how embedded intelligence reduces context switching, accelerates decision-making, and helps organizations turn data into action without adding more software.
Read full article
10 min read
Decision latencyThe Hidden Cost of Delayed Answers
Most organizations have plenty of data but still struggle to answer critical business questions quickly. Learn why delayed answers slow decision-making, how answer gaps emerge, and what businesses can do to move from questions to trusted answers and action faster.
Read full article
16 min read
Decision latencyDecision Latency: How to Diagnose and Measure
What decision latency is, where the delay actually comes from, and nine metrics that measure it, from question-to-answer time to decision-to-action time.
Read full article
11 min read
Decision latencyDecision Latency vs Data Latency: Why Fixing the Pipeline Did Not Help
A faster pipeline shortens one segment of the path. It does not settle definitions, assign authority, or turn an answer into action.
Read full article
12 min read
Decision latencyHow to Run a Decision Latency Audit in Five Questions
A practical five-question audit that shows whether slow decisions come from missing data, competing definitions, unclear ownership or delayed action.
Read full article
12 min read
Decision latency7 Signs You Have a Resolution Problem, Not a Data Problem
The data often exists. The delay sits between evidence and an answer the organisation will trust, defend, and act on.
Read full article
Keep going
Related topics. Where to go next.
- Narrower topicAnswer debtStart here if the same questions keep getting asked and never answered.Explore the topic
- TopicGoverned dataStart here if nobody trusts what comes out of the warehouse.Explore the topic
- TopicExtended Planning & Analysis (XP&A)Start here if every planning cycle begins by reconciling last month.Explore the topic
- TopicAgentic analyticsStart here if you are being asked to put an AI assistant on your data.Explore the topic
Case studies
Case studies. What changed, and how.
- Before ZerentroMulti-brand consumer enterprise
11 days→4 hours
Decision latency on the weekly pipeline question
Three answers, one Monday question
A weekly pipeline question that used to take 11 days now returns a sourced answer the same afternoon.
Read the case study - Before ZerentroEnterprise operations
18→6 days
Monthly forecast cycle
When the forecast became a political artefact
Forecast cycle time dropped from 18 days to 6, because the argument about last month’s actuals stopped happening twice.
Read the case study
Questions about decision latency
What is decision latency?
Decision latency is the elapsed time between a business question being asked and a trusted, sourced answer being acted on. It is not report runtime. It is the full path: question, definition, ownership, trust, and action.
What does Yagnesh Ramalingam say about decision latency?
The only analytics metric that has ever mattered is time from question to trusted, actioned answer — and almost nobody measures it. Dashboards, warehouses and AI copilots are inputs to that path. They are not the path.
Is decision latency the same as reporting lag?
No. Reporting lag is how long the pipeline takes to land a number. Decision latency is the full path from the question being asked to a trusted answer being acted on. A fast dashboard can still have a slow decision.
How do you measure decision latency?
Timestamp five recurring questions at the moment they are asked, and again when someone with authority acts on an answer. The interval is the metric. Do not start with query logs. They measure the short step.
What is a good decision latency?
Your own baseline. There is no industry number worth copying. Time five real questions, see which one is slowest, and find out whether the wait is the data, the definition, or the sign-off.
Who should own decision latency?
The person who owns the meeting where the question is asked, not the person who owns the warehouse. The warehouse cannot shorten the time spent arguing about a definition.
Work with me
Bring the number your team can’t agree on.
A free 30-minute call with Yagnesh. You'll find out if the problem is the data, the definition or who owns it.