CTV buyers have spent years asking for more program-level metadata. And the request made sense: linear plans were built using titles and genres. Streaming supply arrived with far less of both.
In fact, 60% of CTV bid requests carry no usable program-level signal, so a declared genre is often the only content information a buyer can still plan against.
That reliance holds up as long as each declared genre describes the program it is attached to. When a single impression declares nine, ten or twenty genres, however, the field stops narrowing anything. Instead, it matches whichever line items happen to target any one of the genres listed, which widens inclusion targeting and triggers exclusions that buyers did not intend.
In other words, more metadata can produce a less-precise signal.
How a declared genre becomes a targeting decision
When a buyer sets genre in the DSP, say on a line item built around Home & Garden content, and each bid request arrives with a genre string, the DSP compares every value in it against the line item's settings; any match makes the request eligible. The DSP reads the field as declared and does not test whether the combination of genres is plausible for an actual program.
For most requests, this works as intended. Peer39's bidstream analysis of September 2026 covered 751 billion CTV bid requests that declared a genre across 12 DSP accounts: 74% carried exactly one genre, with an average of 1.56 per request.
When one impression qualifies for every plan
Now consider a hypothetical show that declares reality, sports, action, drama and comedy all at once. No single program fits all five, but the DSP has no reason to question the string, so a reality line item, a sports line item, and a comedy line item each treat the impression as a valid match. That same impression can therefore qualify as the right content for several different plans.
The same logic scales. A documentary line item accepts a request that lists documentary alongside eight other genres as readily as one that describes a documentary, so the longer the list, the more campaigns it qualifies for and the less the genre setting narrows delivery.
We call requests that declare nine or more genres “Genre Overload.” A related pattern, “Implausible Pairs,” occurs when two genres that rarely appear on the same title are declared together.
What Genre Overload would look like if a show actually had five genres: Kids, Humanitarian, Humor, Home and Garden, Film-Noir
Every Thursday, three comedy writers must negotiate to finish the garden patio by spring at a kids' slime lab, under strict union rules. Meanwhile, a hard-boiled private eye in a fedora explains humanitarian supply logistics, and the judges take careful notes.
How bad genre data distorts both targeting and exclusions
The problem does not stop at inclusion targeting. Exclusions read the same genre field the same way.
A strict “avoid news” setting matches any request carrying a news tag, including entertainment content that lists news among many other genres. That can remove inventory with little connection to news. The same problem works in reverse: a kids and family campaign can match any request carrying a kids tag, including content that lists it beside mature genres.
One error costs reach; the other weakens protection. Both come from the same underlying issue: the buying control treats every declared genre as an accurate description of the program.
Why program-level authentication matters
A declared genre is metadata passed through in the bid request. To know whether it actually describes the content, buyers need to be able to connect that signal back to the program itself.
In Peer39’s September data, requests whose content ID matched a partner catalog averaged 1.40 genres, and 82% carried a single genre. Genre Overload was most concentrated among requests with no content ID at all, appearing at two-to-six times the rate of requests that carried one.
The important distinction is whether the genre is derived from the identified program or simply passed through from the bid request.
What buyers can do
Low effort: a fast check anyone can do
Medium effort: what to ask your partners
Higher effort: verification in delivery
Verify the genre before you target it
Genre remains a useful CTV control; however, problems arise when it’s assumed that every genre declared in the bidstream accurately describes what is on screen.
When that assumption breaks, more metadata can widen targeting, trigger unintended exclusions and weaken the precision buyers expected from the field in the first place. The answer is to know where the signal comes from and to verify it against the program it claims to describe.