Not all data deserves equal trust. Data quality issues often manifest through inconsistencies, unexpected outliers, or metrics that contradict established patterns. Identifying these problems early prevents flawed analysis and misguided decisions.

Key evaluation criteria include:

• Accuracy: Your data should reflect reality. For example, website visit counts should match actual user behavior rather than including bot traffic or test accounts in your production metrics.
• Completeness: You need all necessary data points without gaps. This means ensuring your analysis includes all relevant time periods, user segments, and interactions rather than partial information that might skew conclusions. A common mistake is including data from before a feature went live, so avoid that.
• Consistency: Related metrics should align logically with each other. If signups increased but new user logins didn't, or if purchase counts don't match revenue figures, something's likely wrong with your tracking.
• Timeliness: Data must be current enough to be relevant for your decision. Last quarter's customer preferences may not represent current behavior, especially in fast-changing markets. Always check when your data was collected.^[1]

Remember that even sophisticated analysis based on flawed data will lead to poor decisions. Building confidence with data starts with ensuring the data itself deserves your confidence.