Desk sources and panels for connected TV advertising competitor research

What to take away

Desk sources first

Start with published material. Trade bodies, platform ad libraries and regulator filings can provide a quick, citable starting point, though access and reuse terms vary. A planner in Birmingham can map rival activity in an afternoon without commissioning fieldwork. For UK context, use BARB for television audience measurement, Ofcom's published reports for media-market context and ASA rulings for advertising complaints and standards decisions. These sources serve different purposes and should not be treated as a complete record of CTV campaign spend.

Search the Google Ads Transparency Center and Meta Ad Library for public creative on their respective platforms; neither is a complete record of connected-TV campaigns. For a UK competitor map, list relevant services and platforms such as ITVX, Channel 4, Sky, Prime Video, Roku and YouTube, then verify where each rival actually advertises.

Commercial ad-intelligence sources to ask about include Nielsen Ad Intel, Kantar Media and Comscore. Confirm the supplier's current UK CTV coverage, measured services and formats, reuse rights and licence fee before commissioning; do not assume any dataset gives complete competitor-spend data.

The Google Ads Transparency Center and Meta Ad Library can show public creative on their respective platforms, but neither is a comprehensive register of CTV campaigns. They show creative, not total spend, so pair them with rate card estimates.

Thinkbox publishes an overview of content partnerships that shows how advertisers tie programming to brand messages. Read it before assuming a rival's sponsorship is purely a media buy. The content partnerships overview for TV advertisers explains the mechanics you will be comparing against.

Trade press release notes and filings help you date a campaign, so record the publication date each time you cite one.

LinkedIn's ad documentation is useful even for connected TV work. It describes targeting and video formats that often sit alongside streaming buys in the same campaign. The LinkedIn advertising documentation helps you see where a competitor's video budget may be split.

Panel methods second

Desk sources cannot tell you why a household switched from one streaming service to another. That needs people. Recruit a sample that mirrors your target market, not just your existing customers.

For UK fieldwork, ask YouGov, Kantar, Ipsos or Savanta about respondent panels and recruitment. Brief them on the CTV viewing behaviour relevant to the question, and confirm panel composition and recruitment method before treating results as representative.

A short survey can test awareness, recall and preference. Keep it under ten minutes and avoid leading questions. Qualtrics and SurveyMonkey are examples of survey platforms; they collect responses but are not, by themselves, a source of representative respondents. Our guide to survey tools in England covers platform choices and fieldwork practicalities for English samples.

For preference between packages, choice modelling beats a simple rating scale. Nielsen's marketing optimisation solutions describe how choice and mix modelling inform campaign decisions. The Nielsen marketing optimisation solutions page is a sensible starting point for method design.

Agree sample quotas with your client before recruitment opens. Retrofitting quotas after fieldwork rarely works.

Fieldwork timing matters as well. Avoid the weeks around major sporting events if your category advertises heavily then, because recall skews.

Checklist for the brief

Common pitfalls

Treating a platform's own case studies as neutral evidence is the most common error. They show best practice, not market averages.

Small samples produce confident-sounding nonsense. If you cannot afford a properly sized panel, say so and lean on desk sources instead.

Comparing your own click-through rates with a rival's published case study is another mismatch. Formats and placements rarely compare like for like.

Another trap is ignoring regional differences. Streaming take-up and channel preference vary across England, Scotland, Wales and Northern Ireland. A London-only sample will mislead a UK-wide plan.

Glossary

Common questions

How many respondents do I need?

For a directional read on preference, a few hundred respondents with quotas set properly is usually enough. For sub-group claims, you need enough per group to report separately.

Can I rely on platform case studies?

No. They illustrate what is possible, not what is typical. Use them alongside neutral sources and your own panel data.

Should I research competitors or my own audience first?

Start with your own audience. Understanding your viewers makes competitor comparison meaningful. The connected TV advertising audience research guide sets out that sequence in detail.

How often should I repeat the research?

Annually for market structure, and after any major platform or pricing change. Preference data ages faster than sponsorship data.