Resonate Append FAQs

Resonate’s enrichment solution appends a dynamic and holistic understanding of each consumer to your file to help you reveal your most valuable audience, what moves them and what they’ll do next so you can maximize spend and drive revenue growth. By appending individual-level attributes—like intent, preference, and motivation—you can empower faster, smarter engagement across the customer lifecycle.

Below, we’ve answered the most frequently asked questions about how Resonate Append works so your team can start realizing its value quickly.

If you have additional questions, we have additional answers. Please contact your customer success manager or reach out to resonatesupport@resonate.com, and we’ll be in touch with a response shortly.

Use the links to jump to each question and response:

• What steps do you take to ensure the accuracy of your predicted attributes?
• What IDs are accepted with Resonate Append?
• Which ID types do you recommend for Resonate Append?
• What are the expected match rates by ID type?
• How does Resonate take an address and resolve back to HEM if I provide a HEM and postal address, but there is no HEM match in Resonate?
• If the household has multiple HEMs, how does Resonate determine which HEM?
• Is there a minimum number of records I can match against?
• Are we able to test the data? 
• What is the recommended refresh frequency?
• Can I use a hashed email (HEM) as my unique ID? 
• We don't create unique Customer IDs—what can we do?
• You only appended my info to the Customer ID, can you append my information to other PII or ID types?
• Can you merge our files for us?
• We have CRM Insights within Ignite Platform in addition to Append—why do the results sometimes vary? 
• Why are only a subset of Resonate's 15K+ attributes available via Resonate Append?
• Why are attributes that were previously appended to my data no longer available?
• Why does my file not return complete predictions for each individual? 
• Why do some questions where I expect a single answer provide me with multiple values (ex: Race/ethnicity)?
• Are appended attributes privacy-safe? 

What steps do you take to ensure the accuracy of your predicted attributes?

At Resonate, we’ve built a methodology designed to deliver precision at scale, giving marketers the confidence to act on behavioral intelligence across the real, addressable U.S. population.

We know that modeled attributes aren’t literal truths about individuals—and they’re not meant to be. Their power lies in uncovering behavioral similarity and population-level patterns that deterministic data simply can’t reach. Still, accuracy and confidence matter—especially when data drives dollars. Here's how we ensure quality:

Our Methodology

Our methodology is what enables us to provide what no one else can by delivering the “why” behind behavior at scale. Here’s how that works.

Resonate provides person-level predictions across the adult U.S. population (250M+ people). These are model-generated inferences, not facts collected directly from a consumer.

1. It starts with deep, verified inputs. We anchor our models in the most comprehensive and continuously fielded U.S. Consumer Study, linking self-reported motivations, values, and behaviors to real, identifiable individuals.
2. Then we validate with real-world signals. We connect those truths with 30B+ daily observed online behaviors—plus offline data sources like U.S. Census and credit bureau records.
3. We scale to US population using AI (rAI). Our proprietary AI infrastructure, rAI, connects the verified truths with the observed behaviors and scales them across the entire U.S. adult population—creating individual profiles forecasting how people are likely to think, feel, and act.
4. We weight to census and validate against credit bureau data. Our models are weighted and benchmarked to population census baselines, and we benchmark them against major credit bureau data. We use built-in quality thresholds for statistical confidence and reliability.

Also impacting your match rate and precision is the ID types you use. We maximize your addressable market by matching at both the individual and household level. That means you reach more of your target audience with precision. We use waterfall matching which prioritizes the highest quality match. Yet if an individual-level ID match or resolution can't be made, we will resolve to household-level attributes, which you can learn more about here.

So what if a predicted attribute is “wrong”? In rare cases, you may find that an individual’s predicted traits don’t perfectly align with what you know. That’s expected—and by design. We’re not predicting one person’s exact truth. We’re delivering statistically reliable, behaviorally rich insight across millions, enabling marketers to act with confidence across channels and at scale.

This is how brands unlock: Smarter segmentation beyond basic demos, personalization that performs, and targeting that finds more people like your best customers.

Bottom line: Modeled data isn’t “bad data.” It’s the only way to reach the full addressable population with meaningful precision and future-looking insight.

What IDs are accepted with Resonate Append?

Resonate accepts a variety of ID types for your flexibility. We require a Customer ID and at least one of the following identifiers for matching:

• HEM (MD5 is preferred or SHA-256, SHA-1)
• MAID
• ZIP11
• Postal Address
• IP Address

Please note, any emails should be personal emails (not business emails), and then hashed before sharing with Resonate. See our hashing script instructions for details. 

Which ID Types do you recommend for Resonate Append?

Resonate’s waterfall matching requires a Customer ID + at least one of the other accepted ID types (HEM, MAID, Postal Address, ZIP11, IP Address). Expected precision and match rates can vary depending on the data IDs you provide.

Please see our “Resonate Append fill rate, match rate, and waterfall matching” article for recommendations on expected outcomes based on your ID types. Please note that if you are providing emails, they should be personal emails and not business emails.

What are the expected match rates by ID type?

Resonate delivers market-competitive match rates, which vary based on the combination and type of identifiers you provide. We have guidance on reaching and optimizing match rates within this article. But match rates are only part of the value.

A key differentiator of Resonate Append is our 100% fill rate on all matched records. Unlike other providers who may leave fields blank—even when a match occurs—Resonate ensures every matched record is fully enriched with the licensed data. That means no guessing, no incomplete rows, and no surprises.

To further ensure data quality, we include a validation step as part of our Proof of Value process. This allows you to review match performance and attribute accuracy before moving to full-scale activation—so you're confident in every insight you're activating against. You can learn more about the validation step in this article.

How does Resonate take an address and resolve back to HEM if I provide a HEM and postal address, but there is no HEM match in Resonate?

When a HEM (Hashed Email) you provide doesn’t match directly to a known HEM in Resonate’s identity spine, we don’t stop there. Instead, we use the full postal address (including street-level detail) to construct a ZIP11—a highly precise version of a ZIP code that includes ZIP+4 and the house number.

We then match that ZIP11 against our ID graph to identify the best available HEM(s) associated with that household. From there, we assign a representative HEM that enables downstream enrichment and audience modeling.

This fallback process is part of our waterfall matching strategy, which maximizes fill rate and actionability while preserving the best resolution quality possible.

If the household has multiple HEMs, how does Resonate determine which HEM to use? 

If a Household has multiple HEMs (e.g., multiple adults using different emails), Resonate chooses a representative HEM for the household. Our goal is always to ensure you get the highest-confidence match possible for modeling and enrichment—whether that’s at the individual or household level. You can learn more about our process in our waterfall matching article.

Is there a minimum number of records I can match against?

We support a minimum of 20K records in your file. Incremental files can offer fewer records.  

Another instance where we accept fewer records, is within a proof of value phase, where we work through preparation of no more than 5K to 10K records as a sample prior to your full 20k min. file. You can learn more about this phase in this article.

Are we able to test the data? 

Yes, Resonate provides a Proof of Value engagement that enables you to test the technical feasibility and value and expected match rates for the Appended data. You can learn more in this article.

What is the recommended refresh frequency?

For Resonate Append, we update our models in-cycle with our wave releases every eight weeks. We recommend refreshing your list of customer IDs quarterly, unless you have new IDs on a more frequent basis, which you can update up to monthly (pending the quantity of records in your file do not exceed the quantity of records purchased). 

Please note, for IP Addresses, we update on a more frequent cadence to ensure you can keep pace with fast-changing online behavior. Accordingly, for this ID-type if you send us data monthly, you'll have monthly updated predictions. 

Can I use a hashed email (HEM) as my unique ID? 

Though we accept HEMs as one of the identifiers we can match on, they cannot be used as your unique customer ID. To ensure privacy compliance and avoid returning any PII, Resonate requires a persistent, unique customer ID alongside the HEM. This approach keeps the enrichment process privacy-safe and ensures enriched records are tied only to your internal identifier—not any identity-based signal like email. See the next response for information on locating existing systems-created Customer IDs within your tech stack. 

We don't create unique Customer IDs—what can we do?

Even if your team doesn’t manually assign customer IDs, most CRMs, CDPs, and audience platforms automatically generate a unique identifier for each record. These internal IDs are essential for rejoining enriched data from Resonate Append back to your systems.

We recommend checking your CRM or data platform for existing system-generated IDs, which are typically used for internal record tracking and deduplication.

Here's where to find them in some of the most common systems:

• Salesforce: Use the “Contact ID” or “Lead ID” field—available via report exports or data loader tools.
• HubSpot: Use the “Contact ID” from your contact export—it’s included by default in full contact exports.
• Adobe Experience Platform (AEP): Use the personID, ECID (Experience Cloud ID), or another primary identity namespace configured in your schema.
• LiveRamp: Use the RampID, Customer ID, or any persistent identifier you’re using within LiveRamp’s identity graph.
• Snowflake or other CDPs: Look for fields like user_id or customer_id that uniquely identify rows.
• Braze, Segment or other CX platforms: Check for fields like external_id, user_alias, or anonymous_id.

If you’re unsure which ID to use, your CRM or data team can confirm the best fit. Including a unique ID ensures Resonate can return your enriched file in a format that’s easy to match back to your original dataset.

You only appended my info to the Customer ID, can you append my information to other PII or ID types?

To maintain strict privacy compliance, Resonate only appends data to your provided persistent, non-PII customer ID—not to identity-based signals like HEMs, MAIDs, or IPs. This ensures that enriched data is never tied directly to personally identifiable information and stays privacy-safe throughout your environment.

Can you merge our files for us?

For compliance and privacy reasons, Resonate does not perform any data merging that involves PII or identity-based identifiers. We return the enriched file keyed to your unique customer ID, allowing your team to safely and securely handle any necessary merges within your internal systems. Our team can help guide your internal teams on how to merge the data if additional support is needed—please speak with your Customer Success Manager for additional support.

We have CRM Insights within Ignite Platform in addition to Append—why do the results sometimes vary? 

You may notice slight differences between the insights appended to your file via Data Append (at the individual ID level) and the insights you see within Ignite Platform (via CRM Insights). This is expected, and here's why:

• Different levels of analysis: The Ignite Platform delivers insights at the audience or aggregate level, while Data Append returns insights at the individual ID level. Translating attributes from an audience view to an individual record can naturally introduce some variation.
• Timing differences: Because Resonate’s data is continuously refreshed, you may also see differences depending on when the file was uploaded or appended. If the file was onboarded and appended at different times, you could be seeing different versions of the underlying data—especially for fast-changing attributes like intent or sentiment.

These differences are normal and reflect both the granularity of the data and the real-time nature of Resonate’s continuously updated insights.

Why are only a subset of Resonate's 15K+ attributes available via Resonate Append?

To maintain data quality and maximize performance, only a select set of Resonate’s 15K+ attributes are available via Append. Unlike the Ignite Platform—which includes built-in governance to flag statistically insignificant or underperforming insights in real time—data append requires a higher standard for modeled accuracy and consistency before delivery.

That’s why we curate a high-performing subset of more than 12K attributes specifically for Append use cases—ensuring every attribute meets strict thresholds for reliability and scale outside the platform.

Why are attributes that were previously appended to my data no longer available?

Occasionally, certain attributes that were previously appended may become unavailable. This can happen for two key reasons:

• Study evolution: Our insights are powered by our US Consumer Study, which evolves over time to stay aligned with market dynamics and consumer behavior. As part of this evolution, some questions—and their associated attributes—may be removed or replaced. When this happens, those deprecated attributes are also removed from the Resonate Append offering.
• Model performance thresholds: For each wave of the US Consumer Study, we evaluate whether an attribute meets our minimum threshold for predictive performance and data quality. If it doesn’t meet that threshold in a given wave, it may be temporarily excluded until its reliability improves.

We proactively communicate any deprecations or performance-driven exclusions and provide alternative suggestions when relevant so you can stay informed and plan accordingly.

Why does my file not return complete predictions for each individual? 

Though Resonate provides 100% of the attributes filled on matched IDs, there are occasionally  instances where you might not see specific predictive attributes returned for an individual. This is by design.

Resonate only returns predictions when the individual completed that type of action. If a person has not demonstrated relevant behaviors—we won’t return a prediction for that attribute. For example, someone who has not spent money on a cruise won’t receive a prediction for amount spent on cruises.

This ensures accuracy and avoids making assumptions based on insufficient data. Instead of applying generic or extrapolated values across everyone, we prioritize delivering precise, behaviorally grounded predictions. This is what sets Resonate apart—and why our data is trusted for activation, analytics, and personalization. If you have additional questions around this, please reach out to your Customer Success Manager. 

Why do some questions where I expect a single answer provide me with multiple values (ex: Race/ethnicity)?

Some attributes—like Race/Ethnicity, income sources, or product selection criteria—are based on multi-select survey questions or reflect real-world complexity. For example, individuals may identify with more than one race, hold multiple jobs (e.g., full-time and freelance), or consider several factors when making purchase decisions.

Because Resonate’s data is modeled, these multi-response values reflect the model’s prediction that more than one category is relevant for that individual. This approach better captures the nuance and diversity of real consumer behavior.

Are appended attributes privacy-safe? 

Resonate is privacy-first by design. We are SOC-2 certified and our data and processes are strictly developed in compliance with the ever-shifting US privacy laws. Resonate models use fully pseudonymous identifiers and ethically sourced behavioral data—so your teams can move fast without adding risk. Learn more about our privacy and compliance in our Trust Center.

If you have additional questions, we have additional answers. Please contact your Customer Success Manager or reach out to resonatesupport@resonate.com, and we’ll be in touch with a response shortly.