Define metrics, missing values, duplicates, refresh needs and ownership before commissioning BI or analytics, and turn that review into a clear data preparation scope.
Before commissioning analytics, align the definitions of the numbers you need with the condition of the data that produces them.Before designing a dashboard, identify the workflows, data and owners who will verify the results. This makes the required preparation scope easier to explain.
For companies commissioning BI or an analytics platform for the first time, this article presents For f's suggested discussion checklist. It is not a quality score that guarantees business results.
Start with the decision the metric will support
Even sales can mean order value, invoiced value or revenue after returns, each requiring different aggregation rules. Write a short statement identifying who will use the number, in which workflow or meeting, and for what decision.
For example: a sales manager reviews weekly orders by product to decide next week's actions. Then define the period, tax treatment, cancellations and reporting cutoff. This is a metric-design example, not a client result.
Five checks for data preparation
| Item | What to check | What to prepare |
|---|---|---|
| Definitions | Does the same field name have the same meaning everywhere? | Metric names, calculation rules and exclusions |
| Missing values | When are required values left blank? | Anonymized samples including blanks |
| Duplicates | Are there IDs that distinguish customers and transactions? | ID rules and lists before consolidation |
| Updates | How current is the data? | Update times, closing processes and delay handling |
| Ownership | Who decides how to interpret and correct unclear values? | Operational owners and data stewards |
Google Cloud's automated data quality documentation describes checking data against defined rules and reviewing or monitoring the results. It includes rules for missing values and duplicates.Source: Google Cloud Auto data quality overview
The rules still need to reflect the business. A blank is not always an error. Treating not-yet-entered, not-applicable and integration-lost values alike hides different causes.
Define the scope you need rather than cleaning everything
For an initial engagement, define the first metrics and their required periods, fields and systems, rather than preparing all company data. Where the need for correction is unclear, flag the item for review instead of overwriting it.
- Preserve source data:Make transformations and corrections traceable.
- Record decisions:Retain the rules used to merge customers, exclude transactions and fill values.
- Assign owners:Distinguish the person verifying business correctness from the person implementing the processing.
A scope might include a data inventory, quality issue assessment, agreed transformations and reconciliation of totals. Have the estimate distinguish data preparation from dashboard production.
Verify the interface and the basis for its numbers separately at handover
Choose the period and data for verification and compare totals with existing business records. Investigate differences through cutoff times, exclusions and source updates, rather than guessing which number is correct.
Include field definitions, aggregation rules, update procedures and support contacts in the deliverables. Prioritize explainable numbers before choosing advanced tools.
Frequently asked questions
Can we discuss data spread across Excel files?
Yes. Start with file types, owners and update methods. Confirm confidential and personal information handling before sharing appropriately scoped samples.
Does installing a data quality tool complete preparation?
Running checks is different from defining business meaning. Consider required values, exception handling and who will make corrections.
Do we need to commission a complete analytics platform first?
You can first assess the data and issues to determine the required scope. For f'sData platforms and analyticsservice can be tailored to your current situation and objectives.
FOR F / NEXT STEP
Talk to us about preparing data for analysis.
For f helps define the scope, deliverables and approach around your current challenges. You can start with what you know, even if your documentation is incomplete.
References and verification date
Information checked on September 7, 2026. The public sources below inform this article; procurement guidance and examples are For f's explanations. Check current official documentation when using a product. The thumbnail is an AI-generated concept image, not an actual interface, document or measured result.