There is a number your onboarding vendor shows you after a CRM upload, and there is a number that determines how many people you can actually reach with paid media. These two numbers are not the same, and the gap between them is rarely explained in a kickoff call.
Understanding where records disappear — and why — is one of the more consequential skills a media buyer can develop, because it changes how you size campaigns, set frequency expectations, and evaluate whether a data partner is worth the cost.
What "Match" Actually Confirms
When a CRM file is onboarded through an identity graph — whether that is a managed service, a self-serve clean room, or a platform-native tool — the match process links hashed customer records to persistent identifiers in a reference graph. A match is confirmed when the incoming signal (hashed email, name-and-postal combination, phone) finds a node in that graph.
That confirmation is real. But it answers a narrow question: does this person exist in the identity infrastructure? It does not answer the operational question: can you deliver an impression to this person through a media channel you actually have access to?
Those are different questions, and the pipeline between them has several steps where records fall out.
The Four Dropout Points Nobody Talks About
1. Graph coverage versus platform inventory
An identity graph may carry 250 million resolved profiles, but the DSP or publisher you intend to activate through sees only the subset of those profiles where a live, cookieless ID — a RampID, UID2, or equivalent — has been synced and accepted. Sync rates between identity providers and activation endpoints vary significantly. A record that matches in the graph may not have a synced ID present in your buying environment at the moment your campaign runs.
2. Consent and signal suppression
Privacy regulation and platform-level signal controls have introduced suppression layers that operate downstream of match. A matched record tied to a user who has opted out of interest-based advertising in a given jurisdiction, or who uses a browser or device that blocks identifier propagation, will not receive delivery even though it counted toward your match total. These suppressions are applied at activation time, not at onboarding time, so they never reduce the number your vendor reports.
3. Recency of the identity link
Identity graphs are probabilistic and time-sensitive. A link between a hashed email and a device ID that was established eighteen months ago may no longer be valid — the device was sold, the household moved, the email was abandoned. Graphs vary in how aggressively they refresh links and deprecate stale ones. Match rate calculations generally do not distinguish between a fresh, high-confidence link and an aged, low-confidence one.
4. Audience overlap with existing exclusions
Campaigns routinely carry suppression lists: existing customers excluded from acquisition campaigns, recent converters excluded from retargeting, frequency-capped users. Matched records that fall into these exclusion pools are subtracted from the deliverable universe after onboarding. This is correct campaign hygiene, but it means the operational audience is smaller still.
A Practical Way to Estimate True Reach
Instead of accepting a match rate as a proxy for audience size, build a four-step estimate before you set budget or pacing expectations.
First, take your reported match count and apply a platform sync discount. If you are activating through a major DSP with strong identity partnerships, a reasonable starting assumption is that 60–75% of matched records have a usable synced ID in that environment. If you are activating through a mid-tier or niche publisher, that figure may be closer to 30–50%.
Second, apply a consent and signal availability discount. In campaigns with significant exposure to iOS, Safari, or regulated geographies, another 15–25% reduction in the addressable pool is a conservative assumption.
Third, account for link recency. If your CRM file contains records that are more than two years old — common in B2B files where contact data ages quickly — add an additional 10–20% discount to reflect stale identity links.
Finally, subtract your suppression audiences.
The result will be significantly smaller than your stated match rate implies. For many CRM files, the truly addressable audience at campaign launch is 30–50% of the matched record count. For older or less complete files, it can be lower.
This is not a failure of the technology. It is the honest arithmetic of identity-based media.
What to Ask Data and Platform Partners
The information you need exists; you often have to ask for it specifically because vendors default to reporting the most favorable number.
Ask for post-sync reach estimates within your specific activation environment, not graph-level match rates. A credible identity partner can tell you how many matched records have a live, synced ID in the DSP or publisher you are using.
Ask whether match rate reporting includes or excludes suppressed and opted-out records. Some vendors report gross matches; others report net of consent signals. The distinction matters enormously for regulated categories.
Ask for link age distribution. If a vendor cannot tell you what percentage of your matched records are backed by identity links refreshed within the last six months, that is information in itself.
Ask for delivery-based validation after the first flight. Impression delivery against the onboarded segment, divided by your pre-campaign addressable estimate, gives you a measured reach efficiency figure you can use to calibrate future buys.
Why This Changes Budget Sizing
If you size a campaign assuming 80% of your CRM file is reachable and the actual figure is 40%, you will either exhaust the addressable pool faster than expected — driving up frequency and diminishing returns — or you will underdeliver against your planned impressions and not understand why.
Both outcomes are common. Neither is visible in the match rate report.
Building conservative, multi-factor reachability estimates into campaign planning is not pessimism. It is the kind of precision that separates buyers who reliably hit delivery targets from those who are perpetually surprised by pacing shortfalls and frequency spikes.
Match rates are a starting point. Reachability is the number that matters.