How UK addressable TV targeting actually works

A conventional television spot buys everyone watching a programme at a time. Addressable television buys the households you selected, wherever in the schedule they happen to be, by substituting your ad for the one that would otherwise have played. In the UK that substitution happens at the Sky or Virgin set-top box on linear inventory, and in the stream on broadcaster video-on-demand.

That mechanism is why the vocabulary is what it is. If you are buying a slot, you talk about programmes, dayparts and audience delivery against a demographic. If you are buying households, you talk about attributes the broadcaster's data partner holds against an address, and about the postcode geography those addresses sit in. Everything awkward about addressable TV targeting follows from that one substitution, including the fact that the selection is made before anybody sits down in front of the television.

The second consequence is commercial. Because the buy is a household selection rather than a share of an audience, it is sold with a minimum burst rather than a minimum weight. Sky AdSmart's is £3,000, and AdBuyMCP enforces it in code: a request under the floor is rejected before any booking document is generated, with the amount converted to sterling first so a non-GBP budget cannot sail under a sterling floor.

An addressable plan on AdBuyMCP is an intersection of up to four household attributes with a list of postcode districts and the broadcast regions those districts sit in. The attributes come from the Sky AdSmart catalogue, which is Experian Mosaic and CACI-derived, and the compiler picks them by matching your persona's own terms against catalogue keywords rather than by asking a language model to name a segment.

The distinction matters more than it sounds. A model that emits a segment identifier can emit one that does not exist, and a plausible-looking identifier is indistinguishable from a real one until a sales house rejects the booking. Registry lookup can only ever return an attribute that is in the catalogue.

Three of the seven channels AdBuyMCP compiles for have a fidelity score that varies with the quality of the brief — CTV, paid search and paid social. Addressable TV is not one of them. It is fixed at 55 out of 100 with a household grade, and the compiler's own rationale says why: household-level proxy, Mosaic and CACI attribute combinations intersected with postcode districts, strong geographic precision, firmographics approximated by household affluence and occupation attributes.

A fixed score is unusual enough to be worth dwelling on. It means there is no version of your brief that makes addressable TV a person-level buy, and no amount of budget that changes the grade. If you need to know that the person watching holds a particular role, this channel cannot tell you that, and the number is the honest summary of how far from that claim it is.

One thing moves the number, and it only moves it downward. Naming a city the geography table cannot map costs fifteen points and the rationale names the city, so an unmappable geography is visible on the plan rather than quietly dropped. Separately, and without changing the score at all, where your brief carries firmographics the compiler attaches a warning in plain words: B2B on addressable TV is a household proxy, and the attributes describe the household, not the job title. That warning travels with the plan into the approval screen.

Sky AdSmart's minimum burst is £3,000, checked before anything else happens. On top of that, AdBuyMCP enforces its own plan floor of £1,000 on addressable and linear TV when the plan is built, which is a different thing from the vendor minimum and bites earlier: the plan floor stops a seven-channel allocation putting a token £200 on television, and the vendor minimum stops a booking that would be rejected.

There is no data CPM uplift on addressable TV in the way there is on CTV, where consent-chained B2B segments carry roughly £2.50 to £3.00 per thousand on top of the media. The AdSmart attribute set is part of the product rather than a marketplace purchase. What there is instead is a production cost you should budget separately: television copy has to be delivered as a broadcast-quality mezzanine, and it has to clear.

AdBuyMCP's fee sits on top of the media at 10%, dropping to 5% plus £749 a month above £25,000 of trailing 30-day spend, and it is disclosed per line rather than blended into a CPM. On a £3,000 minimum burst that is £300, which is worth checking against what an independent agency would charge you for the same booking, because at that level the honest answer is sometimes that it is not worth either party's time.

The thing that surprises buyers new to UK television is that the media is rarely the constraint. Clearcast is. Every commercial has to be cleared before it can be scheduled, the clearance carries a clock number that the booking is made against, and contentious copy takes longer than clean copy. AdBuyMCP's AdSmart workflow puts a lead time of about fourteen days on a booking including clearance, and the ITV Planet V workflow about ten.

That is why the creative pre-check exists and why it runs before anything is commissioned rather than after. It predicts the flags Clearcast, the CAA and publishers are likely to raise, and where you declare that you hold substantiation for a claim, it records the declaration and states plainly that it has not assessed your evidence. A pre-check is not a clearance, and we do not pretend it is one; what it buys you is finding out that your headline claim needs substantiation before you have paid for the film.

The practical planning rule: work backwards from the on-air date by a fortnight for the clearance and the booking, and add the production time for a broadcast master on top of that. A campaign that needs to be live next Monday is not an addressable TV campaign.

Addressable television has no self-serve buying API in the UK, and we do not pretend otherwise. What it automates is the paperwork and the state machine around it. The workflow generates a Sky AdSmart booking document naming the attribute combination, the postcode districts, the flight and the estimated addressable household universe; drives the Clearcast submission; prepares the submission to the accredited-partner portal; waits for the sales house to confirm; and finally reconciles box-level impressions against the booked universe.

It is the right line when your audience is geographically concentrated, when household affluence or occupation is a decent proxy for the person you want, and when you want television's attention profile without national television's price. A regional retailer with twenty catchments, a B2B business whose buyers cluster in a handful of city centres, a service that is only sold in three broadcast regions: all of these are better served by an addressable buy than by a spot buy, and the geography is what makes them measurable afterwards.

It is the wrong line when the account list is the audience. If the requirement is to reach eleven named companies, addressable television cannot do it and nothing in this market can. It is also the wrong line at the bottom of the budget range: at a £3,000 minimum burst against a total monthly media budget of £5,000, television is the whole plan rather than a channel in it.

The reason addressable TV keeps its place in a seven-channel plan despite a 55 is the same reason cinema keeps its place with a 40. A visible approximation you can argue with beats an unlabelled demographic buy, and the geography that makes it addressable is also the geography that makes a matched-market test possible later.