Sometimes, you just have to pick a lane. If the discrepancies are due to conflicting sources, cross compare the variance to see if they’re consistent. If they are, then just pick one and flow with it. For example, a consistent -5% discrepancy will still give you the same view of a trend. Of course it’s better to be able to audit, verify and cleanse the datasets that you have. But in the case of discrepancies due to data loss which cannot be recovered, then understand that you can still learn a thing or two by simply analyzing trends.

There could be several sources of discrepancies. Importing correct data: Ensure all imported data aligns. For example, correct this during collation if one source provides time in GMT and another in local time. Loading correct data: Verify the accuracy of operations. Tools often perform actions like counting or averaging columns by default without informing the user. Ensure these make sense. For instance, counting time spent on a page is irrelevant unless you count visitors. With tools like Power BI and multiple tables, ensure correct key usage for data collation. Finally, use multiple tools to verify data receipt and import. Continuously monitor data veracity.

Check for Data Gaps or Missing Data: Review your traffic data to ensure that no tracking scripts are missing or broken. Discrepancies may result from improperly installed tags (e.g., Google Analytics or other tools). Ensure Consistent Tagging: Verify that tracking pixels, UTM parameters, and analytics tags are consistent across all marketing campaigns, channels, and pages. Inconsistent or broken tagging can lead to inaccurate traffic attribution. Eliminate Bot Traffic: Filter out bot traffic, spam referrals, or irrelevant traffic. Bots can inflate traffic numbers, skewing ROI calculations. Focus on Quality Metrics: look at traffic numbers, that influence conversions, such as bounce rate, session duration, and conversion rate.

Refine Your ROI Formula Define Clear Inputs for ROI: Ensure your ROI formula is clear and considers the total cost of marketing efforts, including ad spend, labor, tools, and other resources. The formula typically is: ROI = (Revenue from Marketing – Marketing Costs) / Marketing Costs Use Revenue Attribution: Attribute revenue back to the appropriate marketing channels based on the data available. Discrepancies in traffic data may obscure which channel is truly driving value. Consider Assisted Conversions: Evaluate assisted conversions to understand how multiple marketing channels work together to drive results. This will give a more comprehensive view of marketing ROI rather than focusing on the last-click model alone.