Audience minimums and filler accounts.
Every ad platform has a floor. Here is what each one actually publishes, how 1:1 ABM vendors pad small lists to clear it, and what that padding costs you.
Ad platforms refuse to run an audience below a minimum size. The floor that decides most B2B programmes is LinkedIn's, and LinkedIn publishes it: 300 matched member accounts. If your target list is 37 people, 263 people you did not choose have to be added to get the campaign live. Those are filler accounts. Any vendor running a below-floor list on LinkedIn has added them — including us. That is arithmetic, not an accusation.
Last reviewed: 2026-07-29.
This is the single most useful thing I say on a sales call, and it costs us deals. I keep saying it because the alternative is a customer discovering it three weeks into a flight.
Here is the version I give on the phone. You come to us with a list of 37 named people at 12 target accounts. You want ads in front of exactly those 37. We upload the list. LinkedIn looks at it and says no — the minimum is 300. So someone has to put 263 more people into that audience before it will run.
The way I describe what gets added — and I am exaggerating to make the shape of it obvious — is 263 retired people from Sweden. They used to work at Volvo and Ericsson. They are 80 years old. They are cheap to match, they are alive in the identity graph, and they will never buy your DevOps tool. Filler gets picked for how easily it resolves, not for who it is.
So when a vendor promises ads to one named person on LinkedIn, the question to ask is what the other 299 are. There are only three answers — pad the audience, widen the list, or move to a platform with a lower floor — and all three are below. The arithmetic does not leave a fourth.
What is a filler account?
A filler account is a record added to your ad audience for the sole purpose of getting that audience past a platform's minimum size threshold — not because you wanted to reach that person. The vendor picks them, you pay for their impressions, and they appear in your reach and frequency numbers as though they were prospects.
Filler is not random. It is whatever the vendor can match cheaply and at volume. Cheap-to-match records skew toward profiles that have been in identity graphs longest and have the least commercial value: lapsed accounts, retired professionals, people who left the workforce a decade ago, and people in countries you do not sell into. A record that is easy to resolve and worth nothing to you is the perfect filler from the vendor's side of the trade.
Three terms get mixed up here, so let me separate them:
- Filler accounts — records added to clear a minimum. What this page is about.
- Lookalike expansion — the platform generating a modelled audience from your seed list. A different mechanic, disclosed by the platform, and covered in B2B lookalike audiences.
- Audience expansion / Advantage+ — the platform's own opt-out setting that delivers beyond your selected audience when it predicts better performance. Also disclosed, also different.
Only the first one is invisible by default.
What are the audience minimums on each ad platform?
I went back to each platform's own documentation on 2026-07-29 rather than repeating what the industry says. The result is more interesting than the numbers: three of the five publish a floor you can quote, and two of them publish nothing you can plan against.
| Platform | What the platform actually publishes | Applies to | Where it is published | Verified 2026-07-29 |
|---|---|---|---|---|
| "The minimum audience size required for an ad set is 300 member accounts"; separately, "your list must have at least 300 rows for a successful upload" and "your list must match a minimum of 300 member accounts to be used in an active ad set" | Ad set delivery and list upload — two separate gates | Target audience size best practices, Contact list targeting | Yes — quoted verbatim | |
| "To stay eligible, a list must have at least 100 members added or updated within the last 540 days"; segments on Search, Display and YouTube "must have a minimum of 100 active visitors or users within the last 30 days"; "We recommend audience list sizes of at least 100 users to prevent ads from not serving" | Customer Match list eligibility, then per-network serving | About Customer Match, How your data segments work, Fix Customer Match issues | Yes — quoted verbatim | |
| X | The minimum audience size is 100 users post-match, and "if an audience with less than 500 users is matched, it will not be available for targeting in the X Ads UI" | Audience creation at 100, then Ads UI availability at 500 | X Ads API — Audiences | Yes — 500 figure quoted verbatim |
| Meta | No minimum stated for using a custom audience. The Marketing API custom-audiences guide documents maximum limits but no floor. Meta does publish a floor for lookalike source audiences — "Origin audiences must have at least 100 members" — which is a different mechanic | — | Custom audiences guide, Lookalike audiences guide | Guides checked; no targeting floor published |
| Not retrievable. The custom-audiences help article returns an error shell to automated retrieval | — | Reddit Ads Help | No |
Two things to take from that table.
Google is the escape hatch and almost nobody uses it that way. A 100-member floor is a third of LinkedIn's, and it is the difference between a 120-person ABM programme running clean and the same programme needing 180 strangers bolted on. If your buying-group campaign is small, start on Google Customer Match and treat LinkedIn as the platform you graduate to — see Google Customer Match for B2B.
"Not published" is a finding, not a gap. When a platform does not state a floor, the floor still exists in the delivery system; you simply cannot plan against it. That is a reason to test with a real list before you commit budget, not a reason to assume there is no limit.
What does a 37-person list actually cost you?
Here is the math nobody puts in a deck. All rows assume LinkedIn's verified 300-matched-member floor and a $3,000 flight.
| Your real target list | Records that must be added | On-target share of the audience | Ceiling on wasted spend at $3,000 | What you are actually buying |
|---|---|---|---|---|
| 37 contacts (12 accounts) | 263 | 12.3% | ~$2,630 | A 300-person campaign with 37 people you care about in it |
| 50 contacts | 250 | 16.7% | ~$2,500 | Same shape, marginally better |
| 300 contacts, matching at 90% | 30 | 90.0% | ~$300 | A campaign that nearly works as sold |
| 300 contacts, matching at 70% | 90 | 70.0% | ~$900 | Blocked until padded — see below |
| 335 contacts, matching at 90% | 0 | 100% | $0 | The smallest list that clears the floor honestly |
| 430 contacts, matching at 70% | 0 | 100% | $0 | The smallest list that clears the floor at a poor match rate |
Read the "ceiling" column as a ceiling, not a measurement. Delivery is not perfectly proportional to audience share — platforms optimise toward whoever engages, and 80-year-old lapsed profiles do not engage. Some of that $2,630 will drift back toward the people who click. But you cannot control the drift, you cannot see it in the reporting, and the reporting itself is computed across the padded audience, so your reach, frequency and CTR are all describing a campaign you did not design.
The row most people miss is the fourth one. LinkedIn's 300-row upload rule and its 300-matched-member rule are separate gates. A list of exactly 300 contacts does not clear the second one, because contact lists do not match at 100%. Native uploads of B2B contact lists match poorly — your CRM holds work emails and people register for consumer platforms with personal ones, which is the whole subject of B2B match rates. At the 70-99% range we see with identity enrichment, a 300-row list lands 210 to 297 members. You need to upload roughly 305 to 430 contacts to reliably land 300.
That is the practical planning number, and I have never seen it written down anywhere: if you want a LinkedIn matched audience with no filler in it, build a list of at least 400 real people.
LinkedIn says the same thing from the other direction, more bluntly than most vendors will repeat: it recommends uploading "lists with at least 10,000 personal or professional email addresses to help ensure your list matches with at least 300 verified email addresses on LinkedIn" (LinkedIn Marketing Solutions Help, checked 2026-07-29). Ten thousand emails to reliably reach three hundred people. That is LinkedIn's own guidance for a native upload, and it is the clearest published admission of how bad native B2B matching is.
And the floor is not even the number LinkedIn thinks you should be at. The same help page that gives 300 as the minimum says LinkedIn "suggest[s] a minimum of 50,000 to drive results" (LinkedIn). The gap between 300 and 50,000 is the gap between "the system will let you" and "the system is designed for this."
How do vendors get past the minimum?
There are three methods. All three are legitimate uses of the platform. The difference between them is who pays and whether you were told.
| Method | What actually happens | Who pays for the extra people | How to spot it |
|---|---|---|---|
| Pad with cheap matched records | The vendor appends records it can resolve at volume until the audience clears 300 | You do, in delivered impressions | Reported audience size is far larger than your list; exported audience contains countries, titles or companies you never selected |
| Combine several lists into one audience | Several targeting lists are selected together so the combined audience clears the floor. LinkedIn documents this as the supported fix | You, for the people on your own lists | Ask for a per-list breakdown of the live audience. If the vendor can produce one, you can see exactly whose records are in there |
| Run on a lower-floor platform | Campaign moves to Google Customer Match (100) instead of LinkedIn (300) | Nobody — this is the clean option | The vendor proposes it before you ask |
LinkedIn itself points at the second method: "If your selected company or contact targeting list has fewer than 300 member accounts, you can select multiple lists to meet the minimum audience size requirement" (LinkedIn Marketing Solutions Help, checked 2026-07-29). One limit worth knowing: a single contact list audience segment still has to come from one uploaded file — "you can only create a contact list audience segment from a single list" (LinkedIn, checked 2026-07-29) — so the combining happens in the ad set's targeting rather than inside one list.
Combining lists you own is ordinary, documented use. I am not claiming that any vendor combines one client's list with another client's, and I have no evidence that anyone does. The point is narrower: a per-list breakdown of the live audience is the artefact that would settle the question either way, which is why it is worth asking for.
The third method is the one to push for, and it is the one that almost never gets proposed, because LinkedIn is where B2B advertisers expect to be and proposing Google reads as a downgrade. It is not a downgrade if it is the only platform where your 120-person audience runs without strangers in it.
What should you ask a vendor before you sign?
Four questions. Every one of them has a short honest answer and an evasive one, and the evasion is easy to hear.
| Ask this | A good answer sounds like | A bad answer sounds like |
|---|---|---|
| "My list is 37 people. LinkedIn's floor is 300. What happens to the other 263?" | "We add matched records to clear the floor. Here is exactly what gets added and how to exclude it." | "Our matching technology handles that." / "We target your list one-to-one." |
| "Can you export the audience you are actually running, not the list I sent you?" | "Yes, here it is." | "The platform doesn't allow exports." / a delay |
| "Is the padding included in the reach and impression numbers you report to me?" | "Yes, and here is the on-target subset separately." | A single blended reach number with no breakdown |
| "What is the smallest list you will run without padding, on each platform?" | A specific number per platform, and a recommendation to widen the list | "There's no minimum with us." |
The fourth answer is the tell. LinkedIn publishes a 300-account floor and applies it to every advertiser on the platform. So "there is no minimum with us" cannot mean the floor does not apply. It can only mean the vendor is handling the floor on your behalf — which may be completely reasonable. Ask them to put in writing how they handle it, and judge that answer rather than the sales line.
We publish feature-by-feature comparisons against Influ2, Vector and AudienceLab. None of those pages says anything about how those companies handle platform minimums, because we have not seen any of their own documentation on the subject and we are not going to characterise it for them. Ask them the four questions above and go on what they tell you.
How do you check your own audience for filler?
Four checks, in order of how fast they work.
1. Compare the sizes. Take the number of contacts you handed over and the audience size the vendor or platform reports. If you gave 37 names and the audience reports 300+, the gap is padding. This takes thirty seconds and settles most cases.
2. Export the running audience and diff it. Not the list you sent — the audience the platform is delivering against. Diff it against your source list. Everything in the difference is either a legitimate match expansion of your own records or filler.
3. Score the difference on three fields. For every record in the difference, check: is the country outside your sales territory, is the job title one you never selected, is the company one you never targeted. Filler fails all three at once, and it fails them in a pattern — you will see clusters, not scatter.
4. Read the demographics tab. LinkedIn Campaign Manager breaks delivery out by job function, seniority, company and industry. If your ABM audience is supposedly 12 named accounts and the company breakdown shows 200 companies, you have your answer without exporting anything. This is the same reporting surface covered in ABM metrics.
If the vendor cannot complete step 2, stop at step 2. An audience you are paying to deliver against and cannot inspect is not an audience you control.
Where this still bites at ContactLevel
I would rather write this section than have you find it out later.
We use filler when a list is below the floor and you ask us to run anyway. There is no technical trick that exempts us from LinkedIn's 300. Identity enrichment raises the share of your list that matches — 70-99%, against a native CSV upload that runs roughly 60-85% on LinkedIn but collapses to 2-20% on Google and Meta with work emails — which shrinks the gap you have to fill. It does not remove the gap. A 37-person list is 37 people at any match rate.
Enrichment helps most exactly where the floor hurts most, and least on LinkedIn. LinkedIn is the one platform where a clean contact list already matches well, so enrichment buys you less there than the marketing usually implies. Where it changes the arithmetic is everywhere else: a 400-person list at a 30% native match — ordinary on Meta or Google with work emails — lands 120 members and needs 180 filler records. The same list at 85% lands 340 and needs none. That is the real argument for contact-level advertising on this specific problem — not "we can target one person," which is not true on LinkedIn for anyone, but "we can get your real list over the line without strangers in it."
We do not have a way to exclude filler from platform-side reporting. You can exclude it from targeting. LinkedIn's own reach and frequency numbers are still computed across whatever audience is live.
We do not track ad impressions at the individual person level, which is directly relevant here: it means we cannot tell you that a specific named buyer saw the ad rather than a filler record. Clicks and website visits we do resolve to the individual. The full, dated version of that boundary is on what we track, and it is the page that wins when any other page on this site disagrees with it.
The honest recommendation, in order:
- Widen the list to real ICP contacts until it clears 300 matched on its own — roughly 400 uploaded at a good match rate. Not padding. Real people you would be happy to reach.
- If you cannot get to 400, run the campaign on Google Customer Match, where the published threshold is 100.
- If it has to be LinkedIn and it has to be 37 people, accept the filler, get it exported and excluded where possible, and price the campaign knowing that as little as 12% of delivery may be on target.
- If none of those work, the campaign is too small for paid, and the answer is sales, not ads.
Option 4 loses us the deal. It is still the right answer often enough that I am putting it in writing.
Why the floor exists at all
Worth understanding, because it changes what you argue for.
No platform publishes a rationale for the number itself, so treat this as the reading that fits the evidence rather than an official explanation. Minimums behave like a privacy control, not a revenue control. If an advertiser could run an audience of one, ad reporting becomes a way to confirm facts about a specific named individual — that this person works here, opened this, is in market for that. Aggregate the reporting across a small enough audience and delivery data leaks identity. LinkedIn applies the same logic one layer up and does say so: its demographic breakdowns "will only show if there is enough data per reporting facet to protect member identity" (LinkedIn Marketing Solutions Help, checked 2026-07-29). A floor of 300 makes the reporting non-identifying by construction.
Which means the floor is not going away, and no vendor is going to negotiate it down for you. The people arguing that platform minimums are anti-competitive are arguing against the mechanism that stops ad platforms becoming a person-lookup service.
It also means the interesting question is not "how do I get under 300." It is "how do I get 300+ real people who are worth reaching." That is a targeting question and a buying group question — twelve accounts with three contacts each is 36 people and will never clear the floor; twelve accounts with the full buying group mapped — Forrester puts the average at 13 people per decision — is 156, and forty accounts at that depth clears it comfortably.
The audience minimum is quietly telling you that your account list is too shallow. Most of the time it is right.
Go deeper.
→ B2B match rates — why native contact-list uploads match so poorly, platform by platform, and what enrichment changes.
→ Buying group marketing — how to build a target list deep enough to clear the floor on real people.
→ What we track today — the dated statement of what ContactLevel measures in production, including what it does not.
→ Google Customer Match for B2B — the lower-floor platform, and how B2B lists behave on it.
→ ABM metrics — the reporting surfaces where padding shows up, and how to read them.
→ Contact-level advertising — how identity enrichment shrinks the gap between your list and the floor.