How dependable is your data today?

Choose the answer closest to how things work now, not how they are supposed to work. Your chart builds as you go. Answers stay in this browser tab; nothing is stored or sent unless you choose to email us your results.

Overhead view of a spreadsheet on a laptop beside printed figures and a calculator — Fig. 1Answer for the spreadsheets you actually use, not the system on the roadmap.

01Sources and ingestionHow does data from your main systems reach the place where you analyze it? 1Someone exports files by hand, usually into spreadsheets. 2Some loads are scheduled, but failures are found by the people reading reports. 3Most loads are automated and alert someone when they fail. 4Loads are automated, monitored, and meet agreed freshness targets.

02Storage and architectureWhere does the data you analyze live? 1In spreadsheets, shared drives, and the source applications themselves. 2In a database or warehouse, but teams keep their own copies and extracts. 3In one warehouse or lakehouse with a documented structure. 4In one platform with separate raw, staging, and modeled layers, and access rules for each.

03Transformation and modelingWhere are business rules applied, such as what counts as an active customer or as revenue? 1Each report or spreadsheet applies its own logic. 2In some shared SQL views or scripts, but logic is also repeated in reports. 3Mostly in version-controlled models that go through review. 4In tested, documented models that every report uses.

04Data qualityHow do you usually find out that data is wrong? 1A stakeholder notices a number that looks off. 2Analysts run manual checks before important reports. 3Automated tests run on key tables with every load. 4Tests, freshness checks, and alerts cover critical data, and owners respond.

05Ownership and definitionsTwo reports show different revenue figures for the same month. What happens? 1Nobody is sure which one is right. 2An analyst investigates, every time it happens. 3There is a written definition, but not every report follows it. 4Each key metric has a named owner, a written definition, and one source everyone uses.

06Operations and reliabilityHow are pipelines scheduled, changed, and monitored? 1Manual runs, or scripts on one person's laptop or server. 2Scheduled jobs, but changes go straight to production. 3An orchestrator, version control, and a review step for changes. 4All of that, plus automated deployment, monitoring, and written runbooks.

07Reporting and BIHow do people get the numbers they need? 1They ask an analyst, who builds a one-off spreadsheet. 2Dashboards exist, but many are stale or duplicate each other. 3A core set of dashboards is maintained and matches agreed definitions. 4People explore governed data themselves, and unused reports are retired.

08AI and ML readinessIf a machine learning or AI project started next month, how ready would the data be? 1We would have to find and assemble the data first. 2The data exists, but history, labels, or documentation are missing. 3Documented history is available, with gaps in access rules or lineage. 4Training data is versioned, documented, access-controlled, and traceable to its sources.

8 questions left.

Hand marking printed charts with a stylus beside a tablet — Fig. 3A score is a starting point for a conversation, not a verdict.

How the score is computed

1
Points. Each answer is worth 1 to 4 points, in the order listed. The first answer is 1 point and the last is 4.
2
Total. We add your eight answers, so the total runs from 8 to 32.
3
Stage. 8 to 13 is Stage 1, Ad hoc. 14 to 19 is Stage 2, Repeatable. 20 to 25 is Stage 3, Defined. 26 to 32 is Stage 4, Managed.
4
Weakest link. If any answer scores 1, the stage is capped at 2. One broken foundation, such as manual loads or no quality checks, limits everything built on top of it.
5
First engagement. We suggest starting with your lowest-scoring area. If several areas tie, we pick the one that comes first in the list, because earlier areas feed the later ones.

This is a self-assessment for orientation, not an audit. Stage names follow the familiar idea of maturity levels; the questions and thresholds are our own. A two-to-four-week Assess phase replaces this estimate with evidence from your systems.

How dependable is your data today?

Stage 2: Repeatable

How the score is computed