On 18 November 2026 IAB Tech Lab shipped OpenRTB v2.6-202611. Among the dated-snapshot changes, the implementation guide gained a pointer to CTV Ad Portfolio signaling for pause, menu, overlay, and squeezeback formats. Those formats do not live in a new JSON object on the bid. They show up as familiar imp.video integers whose meanings moved into AdCOM lists between OpenRTB 2.5 habit and the 2.6 train.
If you traffic video or audit bid requests, you already touch this cluster daily. Buyers filter on plcmt. Legacy paths still emit placement. Viewability vendors still read pos. Muted outstream still sends playbackmethod. This sheet is the four-field map I keep open when a request parses clean and still loses instream deals.
Who this sheet is for
Ad ops and bid-ops engineers who normalize publisher video imps into DSP-ready OpenRTB 2.6, SSAI partners who rebuild device requests at the edge, and anyone writing traffic-shaping rules that mention "instream" without naming the exact AdCOM list behind the integer.
The four fields on imp.video
OpenRTB still transports the auction. AdCOM owns the enumerated lists. The same integer can be valid on one list and nonsense on another, and there is no automatic translation from the deprecated list to the current one.
| Field | JSON path | AdCOM list | What it answers | Do not confuse with |
|---|---|---|---|---|
placement |
imp.video.placement |
Placement Subtypes - Video (deprecated) | Legacy "where is the video unit" label from the 2.5 era |
plcmt: different list, stricter instream definition |
plcmt |
imp.video.plcmt |
Plcmt Subtypes - Video (1–9 since 1.0-202607) | How the ad relates to content the viewer chose (instream, accompanying, pause, overlay, etc.) |
placement: copying the same number mislabels muted outstream as instream |
pos |
imp.video.pos |
Placement Positions (0–17) | Where the slot sits on the screen (fold, footer, fullscreen, CTV geometry) |
plcmt: above/below fold is not instream vs outstream |
playbackmethod |
imp.video.playbackmethod[0] |
Playback Methods (1–11) | How playback starts and whether sound is on at start | Only first array element counts per 2.6 guidance; multiple entries do not combine |
Two neighbors belong on the same sticky note but are not AdCOM placement enums: linearity (linear vs non-linear VAST expected) and startdelay (offset inside content). A pause buy can be plcmt 5 with linearity 2 while pos 7 says fullscreen presentation.
plcmt values you see on live traffic
| Value | Label | Typical playbackmethod[0]
|
Pairing note |
|---|---|---|---|
| 1 | Instream | 1 (autostart, sound on) or 3 (click-to-play) | Requires sound on or clear intent; muted autoplay beside editorial is usually 2 or 4, not 1 |
| 2 | Accompanying Content | 6 (in-view, sound off) | Article and scroll-to-play players that used to ride placement 1 |
| 3 | Interstitial | 1 or 2 | Often with imp.instl 1 on mobile; full-screen attention, not content-relative instream |
| 4 | No Content / Standalone | 2 or 6 | Outstream, feed, and floating players without a content video |
| 5 | Pause (CTV portfolio) | 8 (pause, sound on) or 9 (pause, sound off) | Added AdCOM 1.0-202607; see plcmt 5 |
| 7 | Overlay | 1 or 2 while content plays | Non-linear creative; pos may use 8–17 for band or corner geometry |
Values 6, 8, and 9 cover screensaver, squeezeback, and in-scene inventory on the same list. Full per-value notes live on the plcmt subtypes enum pages.
Legacy placement vs current plcmt (different tables)
placement (deprecated) |
Name | Rough modern bucket (reclassify, do not copy integer) |
|---|---|---|
| 1 | In-Stream | Often plcmt 1 if sound on or intent holds; else 2 or 4 |
| 3 | In-Article | Usually plcmt 4 with playbackmethod 6 |
| 5 | Interstitial / Slider / Floating | Often plcmt 3 or 4 depending on scroll behavior |
The copied-integer failure mode is still common: "placement": 1, "plcmt": 1 on a muted in-article player satisfies range checks and breaks instream semantics.
pos gotchas that trip video audits
| Value | Meaning today | Historical trap |
|---|---|---|
| 1 | Above the fold on first paint | Not "instream"; fold position only |
| 2 | Locked / sticky | OpenRTB 2.0 used 2 for "may be off-screen"; AdCOM repurposed the integer |
| 7 | Fullscreen | Interstitial presentation; pair with instl when appropriate |
| 8–17 | CTV layouts (double box, L-shape, etc.) | Added with portfolio formats; buyers on pre-202607 code may drop unknown values |
The placement positions list is shared with Banner.pos; video sellers sometimes leave pos at 0 while claiming premium instream.
playbackmethod first element only
| Value | Meaning (post AdCOM 1.0-202607 wording) | Common plcmt mismatch |
|---|---|---|
| 1 | Autostart with sound on | Paired with instream; wrong if player is muted |
| 2 | Autostart sound off | Instream plcmt 1 is suspect unless intent is documented elsewhere |
| 6 | In-viewport start, sound off | Typical accompanying / outstream; see enum playback methods |
| 8–9 | Pause-initiated | Expect plcmt 5 on CTV pause inventory |
| 10–11 | Idle / screensaver | Expect plcmt 6 |
Rows that validate and still fail in production
plcmt 1 with playbackmethod 6. Range checks pass. The buyer's instream model reads muted in-viewport start and drops or downgrades the imp. The JSON never looks "invalid."
Same integer on placement and plcmt. Both fields accept 1. Only the legacy list treated muted web players as in-stream. Modern filters read plcmt and playbackmethod together.
plcmt 5 with playbackmethod 1. Pause inventory should carry pause-initiated methods 8 or 9. A linear break method on a pause slot confuses creative and measurement paths, especially as OM SDK coverage expands for portfolio formats.
OpenRTB 2.4 mental model on a 2.6-202611 hop. Early snapshots used different playback labels and had no plcmt at all. If your edge still emits 2.4-shaped video objects, compare against the dated snapshot you claim in ext or in partner docs, not against a generic "2.6" schema. The 2.4 snapshot pack is useful when you audit what changed before plcmt landed.
How you catch it
An LLM patching bid requests, ARTF mutations, or agentic buy objects can emit integers that match a JSON Schema and still disagree with the AdCOM list for the OpenRTB version on the wire. The RPC can return OK while a traffic filter no-bids.
rtblint is the deterministic check on that OpenRTB object: version-aware enums, deprecated placement warnings, and AdCOM rules such as imp.video.plcmt on 2.6-202303+. It is independent of IAB Tech Lab and of AgenticAdvertising.org. The spec does not require it; it is the package I use when a model cannot be the source of truth for enum membership.
$ rtblint validate request.json
Run the same command in CI on sampled traffic after any migration from placement to plcmt.
Enums on the request are only half of servability. Once a bid wins, validate the VAST the buyer returned: paste a live tag URL into the VAST tag tester for fetch, preview, and tracking URLs, then walk wrapper chains in the VAST inspector. For SIMID or IAB sample interactive fixtures on CTV portfolio paths, the IAB-style VAST tester workbench is an independent fork for handshake debugging (not an IAB Tech Lab product).
Where to go next
- When a copied integer breaks plcmt for before/after request pairs.
- Which plcmt values CTV sellers use when pod and pause inventory mix on one app.
- Pause plcmt subtype 5 tied to playback methods 8 and 9.
- Field reference for
video.plcmtwhen you need version notes beside each value.
Top comments (0)