Every paid performance metric you report has a ceiling above it that nobody set on purpose.
Your media team optimizes bids, creative, frequency, pacing. All of that work happens inside a boundary drawn earlier and somewhere else: how many of your customers the destination can actually recognize. You can improve targeting indefinitely. You cannot bid on someone the platform does not know is there.
Call that number your addressable base. It is the share of an audience that arrives at a destination as a person that destination can act on. Not the size of your customer list. Not the size of the segment you built. What survives the trip.
If you own the CDP, you are one of the few people who could find out. Most teams have never been asked to.
Why the ceiling goes unexamined
The people who feel it cannot move it. A media buyer sees a match rate, concludes that match rates are roughly what match rates are, and returns to the levers they control.
The people who can move it are not measured on it. Identity resolution and destination configuration sit with you, and your scorecard is uptime, governance, and delivery. No quarterly goal anywhere reads “raise the percentage of our audience The Trade Desk can recognize.”
So a low ceiling never presents as a problem. A campaign underdelivers and the response is a media response: refresh the creative, raise the bid, widen the segment. Widening is the telling one. It means paying to reach strangers because the known customers cannot be reached, which is the exact inverse of why the brand invested in first-party data.
This can run for years without anyone escalating it, and often it does.
The ceiling is lower than most teams assume
Take The Trade Desk. It does not accept email addresses or phone numbers, raw or hashed. It accepts its own cookie ID, which depends on a sync almost nobody has in place, mobile device IDs, and UID2. It will also take a decrypted ID5 ID if you already run ID5.
So unless you are generating UID2 or already paying for a separate ID vendor, your addressable base there is whatever share of your audience carries a usable device ID. The identifier you have on nearly every customer counts for nothing.
That is not a targeting problem, which means that no amount of media optimization can actually correct it.
One large streaming customer switched on UID2 for a master segment that had been activating on device IDs alone. Addressable reach on that segment rose 91%, before any enrichment was added.
Two levers, both upstream of the campaign
Raising the ceiling means two distinct things, and they compound.
The first is how many people in the audience are reachable in the first place. Someone signed up with one email and now logs in with another. Someone changed phones two years ago. Someone has only ever interacted with you through the app. They belong in the segment, but you cannot act on them, because the single identifier you hold is not one the destination can use. Match Boost resolves further identifiers for those same people, inside the audience, before anything is forwarded. Membership does not change. What changes is how many members arrive somewhere useful.
The second is whether the destination can act on what you send. UID2 normalizes an email or phone number, hashes it with a rotating salt, and produces a value that identifies a person deterministically without carrying their identity in it. The Trade Desk accepts it. That is the entire reason your emails become worth something.
Until recently, using UID2 in mParticle meant generating the values yourself, storing them on the profile, and maintaining them as salts rotated. Many teams built it once and moved on, which meant the values aged out quietly and the ceiling drifted back down without anyone noticing. It's now a toggle, and the refreshing happens whether anyone remembers it or not.
The two questions to ask before you turn it on
Any platform can say it supports UID2. Two things decide how much work it actually takes off your plate.
The first is who runs it. Someone has to convert your emails into UID2 values and maintain them. mParticle runs its own UID2 operator, so conversion happens inside the mParticle network and the upkeep is ours. The alternative is that you maintain it yourself. The credentials, the mappings, and the refresh cycle stay with you, and you forward the values in. Both produce a valid UID2. Only one keeps your data in one place and turns on with a single toggle.
The second is what it costs to turn on. Generation here runs on mParticle's existing partnership agreement with The Trade Desk, so there is no separate contract, no addendum, and no added cost. Other platforms convert through a hosted operator, and several of those require you to register with The Trade Desk and manage your own credentials.
The key question to ask to your media team
Take a real, named segment and ask what its true addressability is before it ever runs in a campaign. What share of it reaches the destination as someone that destination can bid on, and what does the same segment look like with the ceiling raised? That is a sharper question than the match rate on its own, because it separates what the audience contains from what survives the trip.
You own a number the whole media budget sits on top of. At most companies nobody has been asked for it, which means nobody has had to explain why it is where it is.
What’s next: mParticle-generated UID2 support for event integrations
The identifiers you generate for audiences do not stop being useful once a campaign runs. The same UID2s can carry through to the conversion events you send back to The Trade Desk, so the customers you made addressable are also the ones you can measure.
