The Audience Size Your Planning Tool Shows Assumes Every Record in Your CRM File Is Addressable Today

When a media planning tool asks you to upload a CRM file or connect a segment, it typically returns an estimated audience size fairly quickly. That number often becomes the foundation for reach projections, budget allocation, and conversations with stakeholders about expected campaign scale. The problem is that the number usually reflects how many records exist in your file, not how many of those records can realistically be reached through the channels you intend to buy.

Understanding the gap between file size and actual addressability is one of the more practical skills a media buyer can develop, because the gap is almost always present and almost always larger than the planning tool communicates.

Where the Estimate Starts

Most planning tools calculate an audience size estimate by running your CRM records through their connected identity graph and returning a count of matched identifiers. The match itself is real, but what that match represents depends on several factors the top-line number does not show you.

First, an identifier match is not the same as delivery eligibility. A record might match to a known device ID, a hashed email, or a probabilistic household cluster, but that identifier still needs to be present in available, purchasable inventory at the moment your campaign runs. An identifier that matched in the graph six months ago may no longer be active in the supply path your DSP sees today.

Second, different identifier types carry different delivery rates. A match to a deterministic hashed email that a publisher recognizes directly tends to produce more reliable delivery than a match to a probabilistic device ID inferred from household IP patterns. Planning tools rarely break down your estimated audience by identifier type, so a file that looks large may be mostly composed of weaker signal matches that produce thin reach when the campaign actually runs.

Third, addressability is channel-specific. A record that is reachable through email has a different identifier than a record reachable through connected TV, which differs again from one reachable through a social activation path. A planning estimate that reports a single number is collapsing those distinct populations into one figure, which can make a file look more versatile than it actually is across your planned channel mix.

The Freshness Problem Inside Your File

CRM files are rarely uniform in age. A file of two hundred thousand records might contain fifty thousand customers acquired in the past year, a hundred thousand from two to four years ago, and another fifty thousand who have not had any transaction or contact in more than five years. The planning tool treats all of them roughly the same when building its size estimate.

Older records degrade in addressability over time. Emails change, device IDs cycle, and the probabilistic signals used to maintain graph links weaken as behavioral data refreshes without finding a match. A record that was fully reachable during a prior campaign may have drifted into partial or no addressability since then, but it still counts in your file and still contributes to the estimate.

This is worth examining before a campaign rather than after. A rough internal audit that segments your CRM by recency of last known activity can give you a more realistic view of which portion of the file is likely to produce actual delivery. Hypothetically, if you assumed that records with no activity in more than three years carry significantly lower addressability, and that population represents a third of your file, your planning estimate could be overstating practical reach by a meaningful fraction.

What to Ask Your Onboarding or Identity Vendor

Rather than accepting the top-line audience size from a planning tool, consider asking your identity resolution or onboarding vendor to break the estimate down in a few specific ways before you commit to a budget based on it.

Ask for addressability by channel. If you plan to run across display, CTV, and a social activation path, request separate estimates for each. The number will be smaller per channel than the aggregate, but it will be more honest about what each portion of your budget can actually reach.

Ask for addressability by identifier type. How much of the matched population resolves to deterministic identifiers versus probabilistic ones? Platforms and publishers treat these differently, and knowing the composition helps you set realistic delivery expectations.

Ask how recently the graph links were validated. Some vendors refresh their graphs continuously; others work from periodic snapshots. A match to a graph snapshot that is several months old may not reflect the current state of those identifiers in live inventory.

None of these questions require your vendor to do something unusual. They are reasonable requests that any vendor with a mature identity infrastructure should be able to answer, at least partially. If a vendor cannot give you any breakdown beyond the total matched count, that is itself useful information about how much confidence to place in their estimate.

Adjusting Your Planning Assumptions

The goal is not to distrust planning tools but to use their estimates as an upper bound rather than a delivery guarantee. A practical habit is to build your media plan around a conservative share of the planning estimate, with the understanding that actual delivery will land somewhere in a range below the top-line number.

How conservative that assumption should be depends on the age and quality of your CRM file, the channel mix you are buying, and how much of your file is composed of strong versus weak identifier matches. Without historical delivery data from prior campaigns against the same file, a reasonable starting assumption is that some meaningful portion of the planning estimate will not convert into billable delivery, and budgeting against the full estimate sets you up to over-promise reach to stakeholders.

If you have run prior campaigns against the same file, your actual delivery logs are the best calibration tool you have. Comparing the planning estimate from a prior campaign to the delivered unique reach that campaign actually produced gives you a file-specific ratio that is far more useful than any estimate a new planning tool will generate.

Why This Matters Beyond Scale

Audience size estimates affect more than reach projections. They shape frequency assumptions, budget pacing logic, and decisions about whether a file is large enough to support the campaign structure you have planned. An inflated estimate can lead to underfrequency if actual addressable reach is significantly smaller, because your budget spreads across fewer real people than expected and the per-person exposure count rises unexpectedly.

It can also affect audience suppression logic. If you are suppressing existing customers from a prospecting campaign and your suppression file looks large but is mostly non-addressable records, the practical suppression effect will be weaker than the file size suggests, and some existing customers will see prospecting creative anyway.

Treating the planning estimate as a starting point for questions rather than a confirmed answer is a small adjustment that tends to produce more realistic campaigns and fewer surprises when delivery reports come back.