Hashed email audience onboarding is the process of turning your customer and prospect email addresses into secure, one-way codes, matching those codes to household and device identifiers inside a data partner's identity graph, and then activating the matched audience on CTV, programmatic video and display. For high-consideration brands it is the bridge between a CRM full of valuable people and a streaming campaign that can actually reach them. The catch is the match rate: depending on file quality and vendor, anywhere from roughly a fifth to more than four-fifths of your records will connect to a reachable household. This guide explains how the process works, what match rates to expect at each step, how to improve them and how to judge whether a high match rate is also an accurate one.
How does hashed email audience onboarding work?
The mechanics are simpler than vendor decks suggest. There are five steps, and each one can lose records.
- Export and clean. You pull email addresses from the CRM or customer data platform, ideally with first name, last name and postal address attached.
- Normalize. Every address is converted to a standard form: lowercase, stray spaces at either end trimmed, and invalid formats removed. Some identity frameworks apply extra rules, so follow the partner's specification exactly.
- Hash. Each normalized address is run through SHA-256, producing a fixed-length code that cannot practically be reversed. The Unified ID 2.0 normalization guide documents one widely used standard, including the exact encoding its systems expect.
- Match. The onboarding partner compares your hashes with the hashed emails in its identity graph and links each match to a persistent identifier, then to the devices and households associated with it.
- Activate. The matched audience is pushed to a DSP, a streaming publisher or a clean room, where it can be targeted, suppressed or used for measurement.
The first two steps are entirely within your control, and they are where most avoidable match-rate loss happens.
What is a good CTV match rate for hashed email audience onboarding?
There is no single industry number, and every vendor calculates match rate slightly differently. That said, the published ranges are consistent enough to plan against. LiveRamp's own match rate documentation estimates 35%-55% for files of personal email addresses and 20%-35% for business email addresses, and notes that hashed emails can match slightly lower than plaintext. Several identity vendors now claim far higher figures on email alone, typically in tests they run themselves.
Matching to an identity graph is only the first funnel. The audience then has to match again to the specific CTV environment you are buying, and a further share of matched households will not be active on that platform during your flight.
| Funnel stage | What it measures | Stillwater planning range | What drives the low end |
|---|---|---|---|
| File validity | Share of records with a deliverable email format | 85%-98% | Old records, typos, role addresses |
| Identity graph match | Valid hashes matched to a person or household | 35%-70% | Business emails, no postal address, thin consumer file |
| CTV addressable match | Matched households reachable on CTV supply | 45%-75% of graph matches | Narrow publisher list, limited ID coverage on apps |
| In-flight reach | Addressable households actually served during the campaign | 50%-85% of addressable | Low budget, short flight, tight frequency caps |
| End-to-end reach | Records reached at least once | 10%-35% of the original file | All of the above combined |
The bottom row is the number that matters, and it is the one rarely shown. A file of 100,000 customers commonly yields 15,000 to 30,000 households actually reached on CTV. These are Stillwater planning ranges, not guarantees; run your own match test before you size a budget around them.
Why are hashed email match rates for CTV lower than for social platforms?
Brands used to uploading customer lists to large social platforms are often surprised by CTV numbers. The difference is structural.
Social platforms match against their own logged-in user base, so a single email can resolve directly to an account. CTV runs across dozens of apps and device makers, each with its own identity coverage. A household can be matched in the identity graph yet still be unreachable if the streaming apps it watches do not pass a matching identifier into the bid request, or if the inventory is sold only through direct deals that do not accept your audience.
Business email addresses are another drag. Executives in wealth management, private aviation or B2B-adjacent luxury categories often appear in your CRM under a work address. Those addresses rarely tie to a home television. Where you can, capture a personal email or a postal address as well.
How do you improve match rates in hashed email audience onboarding?
The most effective improvements happen before the file ever leaves your system.
- Add postal address and full name. Multi-key matching that includes name and address routinely outperforms email-only files, because the identity graph has more ways to find the household.
- Separate personal and business emails. Upload them as distinct segments so you can see where loss happens and set expectations for each.
- Refresh the file. Records older than three years decay quickly as people change providers and addresses. Remove hard bounces and suppressions before export.
- Normalize exactly to specification. Hashing an address with a trailing space or a capital letter produces a different code that will never match. Test a handful of known records first.
- Use more than one identity route. Many DSPs can activate both a commercial identity graph and an open framework such as UID2. Layering them can widen coverage on apps that support one but not the other.
- Buy inventory that accepts addressable audiences. Ask the agency which CTV publishers and private marketplace deals will honor your segment before you commit.
We cover the strategic side of building and protecting these files in our guide to first-party data for luxury advertising, and how different vendors stitch identifiers together in our post on identity resolution.
Is a higher match rate always better?
No. A match rate measures how many records connected to something. It does not measure whether they connected to the right household.
Accuracy is the quieter problem. Industry research reported by Digiday on CTV identity matches found error rates well above 50% for some match types, particularly those that infer households from IP addresses. A vendor that matches 85% of your file through loose probabilistic links can deliver fewer real customers than one that matches 50% deterministically.
Three questions separate the two:
- What share of matches are deterministic (based on a logged-in or verified link) versus probabilistic (based on shared IP or device signals)?
- What is the denominator? Some vendors calculate match rate against valid records only, others against the full file. Ask for both.
- Has accuracy been tested against a truth set? Credible partners can show validation against known records, not just volume.
For luxury brands, where the CRM may hold a few thousand very high-value households, precision matters far more than scale. Serving a spot to the wrong household in a shared apartment building wastes a $50 CPM and, more importantly, misreports reach.
How should you use an onboarded audience on CTV?
Once matched, a CRM audience can do four distinct jobs. Each one has a different value and a different risk.
| Use case | How it works | Best for | Watch out for |
|---|---|---|---|
| Suppression | Exclude current customers from acquisition campaigns | Every brand; cuts waste quickly | Suppression only works on matched records |
| Retention and upsell | Target existing customers with new offers or renewals | Memberships, clubs, card programs | Small segments struggle to build frequency |
| Seed for lookalikes | Model new households that resemble best customers | Acquisition at scale | Poor seeds produce poor models |
| Measurement | Match exposure logs back to CRM sales | Proving CTV revenue | Requires clean, hashed sales data too |
Suppression is the most underrated of the four. For a brand acquiring customers at $1,500 to $5,000 each, every impression spent on someone who already bought is pure waste. Our post on affluent lookalike modeling explains how to build the seed audiences that make the third use case work.
The fourth use case closes the loop. By matching hashed exposure data to hashed sales records in a secure environment, you can show which households saw your CTV ads and later bought, and compare them with a holdout. We describe that method in detail in our guide to proving CTV revenue with CRM matchback, and the privacy architecture behind it in our post on data clean rooms.
Is hashed email onboarding privacy compliant?
Hashing is a security measure, not a legal exemption. Regulators in several jurisdictions treat hashed identifiers as personal data because they can still be linked to an individual through matching. That means your privacy notice, consent records and opt-out handling still apply.
In practice, a compliant program does the following:
- Discloses in the privacy policy that customer information may be shared with advertising partners for matching and measurement.
- Honors opt-outs and state-level rights requests by removing those records before every refresh, not just once.
- Uses onboarding partners with contractual limits on how matched data can be reused.
- Routes sensitive categories, such as health or financial status, through legal review before activation.
For regulated categories, particularly wealth management, financial advisory and healthcare-adjacent services, compliance teams should approve the segment definition, not just the creative. Our wealth management advertising page outlines how we structure approvals in that vertical.
How long does hashed email audience onboarding take?
Plan for one to three weeks from export to a live, addressable CTV segment. The upload and matching itself usually completes within a few days. The time goes into the steps around it: legal review of the data-sharing terms, agreeing the file specification, confirming which DSP and publisher destinations will receive the segment, and waiting for those destinations to populate the audience so it is ready to bid on.
Refresh cadence matters just as much as setup time. Stillwater plans for monthly refreshes on acquisition and suppression segments, and weekly refreshes where a brand sells memberships, cards or bookings at high volume. Each refresh should remove recent opt-outs, add new customers to suppression and report the match rate for the new file, so a sudden drop in match quality is caught before it affects a full month of delivery.
What are the most common mistakes with hashed email audience onboarding?
These errors account for most disappointing results we see when reviewing a brand's existing setup:
- Uploading email only when name and postal address are already in the CRM.
- Mixing personal and business addresses in one segment, which hides where loss happens.
- Hashing before normalizing, so capitalized or space-padded addresses never match.
- Judging success on graph match rate instead of end-to-end reach on the actual CTV buy.
- Activating a tiny segment on broad CTV and expecting it to build frequency; a 5,000-household audience needs concentrated, addressable supply.
- Refreshing once a year, so new customers keep seeing acquisition ads for months.
- Ignoring accuracy, and buying the highest match rate rather than the most verifiable one.
- No holdout, which makes it impossible to tell whether the CRM audience responded to the ads or would have converted anyway.
A practical sequence for your first onboarding test
If you have never run hashed email audience onboarding on CTV, or your first attempt underwhelmed, use this sequence:
- Audit the CRM: count records with personal email, business email and postal address.
- Build three segments: recent customers for suppression, high-value customers as a lookalike seed and lapsed customers for re-engagement.
- Run a match test with two partners, asking each for deterministic and probabilistic shares and both denominators.
- Confirm which CTV publishers and private marketplace deals will accept each segment.
- Hold out 10%-20% of each segment before launch.
- Activate for at least one full sales cycle, then match exposure and sales back to the CRM.
Want a CTV program built on your best customers?
Stillwater Media designs hashed email audience onboarding, suppression and CRM matchback into every CTV and programmatic engagement, with match-rate reporting from file to household so you see exactly what your data delivers. We accept a limited number of engagements each quarter for brands with an LTV above $5,000 and sales cycles longer than 30 days. Apply to work with Stillwater Media.
About the author: Stillwater Media is a selective performance media agency for luxury and high-consideration brands, headquartered in North Carolina and serving clients nationally and internationally. Its team plans and buys premium CTV, programmatic, streaming audio, DOOH and social media for brands in private aviation, wealth management, luxury real estate, hospitality and premium DTC. Signal. Strategy. Scale.
Frequently Asked Questions
What is hashed email audience onboarding?
It is the process of normalizing and SHA-256 hashing customer email addresses, matching the hashes to households in an identity graph and activating the matched audience on CTV and programmatic. It lets brands reach or suppress CRM records without sharing plaintext emails.
What is a good CTV match rate for a CRM file?
Identity graph match rates commonly run 35% to 70% for personal email files, and lower for business emails. After CTV addressability and in-flight delivery, Stillwater plans for 10% to 35% of the original file to be reached at least once.
How can I improve my hashed email match rate?
Add full name and postal address, separate personal from business emails, refresh old records and normalize exactly to the partner's specification before hashing. Layering more than one identity route can also widen coverage.
Is a higher match rate always better?
No. A match rate shows how many records connected, not whether they connected to the right household. Ask for the deterministic share, both denominators and validation against a truth set before choosing a partner.
Is hashed email onboarding compliant with privacy laws?
Hashing protects data in transit but many regulators still treat hashed identifiers as personal data. Brands need clear privacy disclosures, opt-out handling at every refresh and contractual limits on how partners reuse matched data.
