Pick the one that describes you best. If more than one fits, pick the highest, whatever your revenue. Risk comes from the jobs your data does, not from revenue.
For SaaS data leaders
SaaS Data Level Scorecard.
Find your level, and what breaks next. Ten questions, about four minutes. Your level shows as you score. Add your email to see what to fix now and what to fix before you grow.
1 = Not in place, 2 = Ad hoc, 3 = Partial, 4 = Standard practice, 5 = Tested and proven. Only give a 5 if you have the evidence shown under the question.
01
Reporting separationLevel 1
Do reports run on their own copy of the data, so they never slow down the product?
A 5 needs this evidence
A heavy report has run on the reporting copy at peak time and the product didn't slow down.
02
One source of truthLevel 2
Does every team get the same number for revenue, customers and churn, from one written definition of each?
A 5 needs this evidence
Two teams pulled those numbers on their own and got the same result.
03
Customer-facing validationLevel 3
Is every number a customer sees (invoices, in-product dashboards) checked against the source before it ships?
A 5 needs this evidence
A check has actually stopped a wrong number before a customer saw it.
04
OwnershipLevel 3
Can you name the owner of every number your customers see?
A 5 needs this evidence
There is a written list of those numbers with an owner's name next to each.
05
Safe re-runsLevel 3
If a job runs twice, is the result exactly the same as running it once?
A 5 needs this evidence
You have re-run a job and confirmed the output matched.
06
Customer data separationLevel 3
Is customer data separated at the platform level, so one missing filter can't expose another customer's data?
A 5 needs this evidence
A test query with the customer filter removed still returned only one customer's data.
07
Access by policyLevel 4
Is access decided by a written policy, not by whoever has the admin password?
A 5 needs this evidence
Your last access review found every permission matched the written policy.
08
Blast radiusLevel 4
If one pipeline or system fails, do unrelated features and customers keep working?
A 5 needs this evidence
A real or practised failure stayed contained and unrelated features kept working.
09
Cost ownershipLevel 4
Can you say who owns each part of the cloud bill?
A 5 needs this evidence
Every line of last month's bill maps to a named owner.
10
Data as a productLevel 5
Can other teams use a dataset without asking permission every time?
A 5 needs this evidence
Another team built on a dataset this quarter without asking its owner for access or an explanation.
30-minute fit call
Bring the decision that will not stay small.
A practical first conversation for data and architecture leaders. The point is to determine fit, not to force a sales process.
See what the fit call covers