Validation is a critical quality checkpoint that helps ensure your CRM file is clean, structured correctly, and ready for enrichment. By submitting a small, representative sample before delivering the full file, you reduce risk, accelerate turnaround times, and improve the overall quality of enrichment.

This step enables you to confirm that Resonate can successfully process your data, match it at expected rates, and generate meaningful appended insights—ultimately helping you get the most value from your enrichment investment. 

To make the most of this step, follow these proven best practices. 

Ensure proper file formatting.

Customer ID is required on all records. Format your sample file according to Resonate’s guidelines using UTF-8 encoding, consistent headers, clean field names with no special characters, and .CSV file type.

Save your file with the prescribed naming convention: <clientname>_<Segment>_<yyyymmdd> 

• Example: Clientname_ResonateAppend_20250201.csv 

There are additional details and best practices.

Validate ID structure and quality.

Hash emails using Resonate’s script to guarantee lowercase, trimmed, and properly formatted hashes (MD5, SHA-1, or SHA-256 only). For postal address-based matching, ensure all required fields—Address Line 1, City, State, and ZIP Code or ZIP11—are complete and validated. 

Check for data cleanliness.

Remove duplicate, inactive, or low-quality records. Suppress rows missing critical IDs (e.g. customer ID + HEM, HEMTYPE or MAID, or ZIP11/Address) or containing malformed values (e.g., improperly hashed emails, incomplete addresses).

Submit a representative sample

Predictive data—like the kind Resonate provides—delivers the most value when applied across broad, diverse audiences. Its power lies in surfacing patterns and insights that only emerge at scale, making it essential that clients’ validation samples reflect the full variability of their CRM file. We recommend that clients submit a test file with a random sample of 10% of their customer records (with a minimum of 10,000 records) that spans a mix of geographies, customer segments, and behavioral types. A well-balanced sample ensures that match rate estimates are reliable, attribute distributions are meaningful, and any issues surfaced during validation will be representative of what you can expect in the full enrichment. 

Data quality at Resonate

Our overall match rates are strong and compare favorably with those of major credit bureaus. In fact, in an audit, Resonate’s modeled predictions for key demographic attributes—such as age and gender—aligned with data from a top 3 credit bureau over 80% of the time. This level of alignment underscores the accuracy and reliability of our predictive models and their value in powering meaningful audience insights. 

We take pride in ensuring that the data our clients rely on is not only comprehensive but also validated against trusted third-party sources, giving marketers the confidence they need to activate, measure, and grow with precision.

Testing your match rate and attribute alignment

As part of our onboarding and quality assurance process, Resonate offers clients a validation step to evaluate match quality and attribute accuracy before full activation. Clients provide a randomized sample—a minimum of 5,000 and a maximum of 10,000—that includes a diverse mix of geographies, customer segments, and behavioral profiles. This test file allows Resonate to:

• Confirm proper formatting and ingestion readiness
• Evaluate match potential against our data environment
• Validate attribute quality by comparing Resonate’s age and gender predictions with a top 3 credit bureau

We review these match and attribute validation statistics with each client, ensuring full transparency and enabling early insights into data performance and audience fidelity. This collaborative step helps uncover issues early and ensures downstream activations are built on a strong foundation of trust and accuracy. 

Flag issues early

Use this process to surface formatting errors, missing fields, or redaction-related concerns before they affect the full file. Addressing these issues upfront avoids costly delays or reruns.