A 14% list match reads as under 5% on LinkedIn

A match rate looks like a measurement of your customer list. On most ad platforms it is a rounded range or an approximate count, and below each platform’s floor it is a placeholder that tells you nothing about your data.

Upload one list of 2,000 customers that matches 280 people, a 14% match. LinkedIn reports it as a match rate under 5%. Google reports “less than 20%”. Pinterest lets you run it. TikTok will not. Nothing about the list changed between those four readings. Only the platform’s rule did.

What each platform tells you

From each platform’s own documentation, read on 8 November 2026.

Platform	Minimum before the list is used	What it reports about the match	Below the minimum
Google Ads	100 active users in the last 30 days, per network. 1,000 matched for lists last uploaded before 1 Feb 2024	A match-rate range in steps of about ten points. Size rounded to two significant digits, active users only	Size shows 0 under 100 members
Meta	No number published	Audience size as a lower and an upper bound. An approximate range of entries that failed on hashing	delivery_status 300: “smaller than it should be”. No figure
LinkedIn	300 matched member accounts. The upload itself needs 300 rows	A match-rate percentage over the past seven days, and a deduplicated audience count	Match rate shows <5%, audience shows < 300 members
Microsoft Advertising	300 people “before Microsoft Advertising will use it for optimizations”	Active members on Search and on the Audience network, empty for up to 48 hours	Not stated
TikTok	1,000 matched identifiers	The audience becomes available, or it does not	Unavailable
Pinterest	100 matches	A status that turns green, and a size updated within 24 to 48 hours	Status stays red
Snapchat	No number published	An approximate user count and a targetable status	TOO_FEW_USERS

Reddit is not in the table. Its help center did not render a readable page on the day, and the figures for it found elsewhere were second-hand.

Three things in that table are worth naming:

• LinkedIn’s <5% is a floor message, not a measurement. Below 300 matched accounts, a list that matched 14% and a list that matched 1% look the same. A low match rate on a small LinkedIn list says only that you are under 300.
• Two of the seven publish no minimum at all. Meta and Snapchat tell you the audience is too small through a status code. Neither documents the number, so you find the floor by uploading and waiting.
• The floors differ tenfold. Google and Pinterest use a list at 100. TikTok needs 1,000. A list that works on Pinterest can be unusable on TikTok without anything being wrong with it.

One list, three outcomes

We ran one constructed upload of 2,000 rows through each platform’s documented rules at three match counts. It is the same list each time, so the only variable is how many people the platform recognised. In practice each platform matches a different number, because each knows different people. Holding the count fixed isolates the reporting rule.

Matched (of 2,000)	Google	LinkedIn	Microsoft	TikTok	Pinterest
280 (14.0%)	0–19% · size 280 · usable	<5% · < 300 members · not usable	not used for optimizations	unavailable	green
990 (49.5%)	41–50% · size 990 · usable	about 50% · usable	used	unavailable	green
1,460 (73.0%)	71–80% · size 1,500 · usable	about 73% · usable	used	available	green

Google’s sizes assume every matched person was active on that network in the last 30 days. Fewer active users means a smaller size, and Google says so: the size “only shows active users”. LinkedIn’s audience count can sit below the match count too, because it removes duplicates and members who opted out.

At 990 matches the list clears every floor but TikTok’s, ten people short. At 1,460, Google’s rounding adds 40 people to the displayed size. Neither is an error. Each is the documented rule doing what it says.

What the number counts, and what it leaves out

A match rate is overlap plus formatting, not a quality score for your list. The documentation adds three things that shape it further:

• Who counts as matchable. LinkedIn matches only members with a verified email address. Google counts only users active on that network in the last 30 days.
• When it was measured. LinkedIn’s rate covers the past seven days and the last list uploaded. A list that stopped refreshing keeps an old number.
• Which rows failed. Google’s job report has no information on failed matches, “since only hashes are compared”. Meta returns the number of badly hashed entries as an approximate range. Neither tells you which customers did not match, so you cannot fix a list row by row from the platform side.

Where the list goes now

On Google, a customer list is less a targeting choice than an input to bidding. Google’s Customer Match page says Smart Bidding and optimized targeting “automatically learn which Customer Match lists are helpful”, and that lists are not used at all with manual bidding. Microsoft’s floor is phrased the same way: 300 people before it “will use it for optimizations”.

That moves the cost of a list below the floor. It used to be an ad set that would not serve. Now it is also a signal the automated bidding never receives, and nothing in the campaign reports that it is missing.

What to do differently

1. Size the upload to the strictest platform you buy. Plan for 1,000 matches on TikTok and 300 on LinkedIn and Microsoft. A 14% match rate on a B2B list means about 2,150 rows to clear LinkedIn’s 300.
2. Read a low rate against the floor first. On LinkedIn, <5% under 300 matches is the floor, not your data. On Google, most ranges are ten points wide and the lowest is twenty. Compare buckets, not decimals.
3. Compare platforms by eligibility, not by rate. The rates measure different populations over different windows. Whether each platform can use the list is the comparison that holds.

Every list in this piece goes up as hashed emails or phone numbers, which is often treated as the step that makes uploading customers safe. If the platform can match a hash to a person, can anyone else?