How do you turn a customer persona into a site list?
Part of Lookalike websites
The short answer
Persona based site selection translates what a persona does, wants to learn and likes to read into content topics, formats and publisher types, then keeps sites whose articles cover them. It uses no personal data and assigns no income, health or other sensitive traits to readers: the persona shapes the brief, and each site's content has to prove the fit.
Key takeaways
- A persona earns its place in site selection only through traits that describe content: activities, questions, skill level, preferred formats and language.
- Income, age, health, beliefs and similar traits do not map to topics, and the FTC warns that non sensitive categories combined can become proxies for sensitive ones.
- Each site is tested against the persona's questions by reading its articles, not by assuming who its readers are.
- One primary persona per brief, with at most two secondary ones searched as separate themes, keeps the list coherent; this is an editorial recommendation.
How do you turn a persona into a site list?
Brand teams can describe customers as personas: a name, a short story, what the person does and what they worry about. Media planners receive the persona in the brief and are asked to find websites for it. The trap is to treat the persona as a description of who reads a site. A site list built on content can only test what a site publishes, so the persona has to be translated into subjects first.
Definition
- Persona based site selection
- A planning method that converts a customer persona's activities, information needs and reading habits into content topics, formats and publisher types, then selects websites by reading their articles against those topics, without data about any individual.
The middle steps carry the work. A persona who "wants to get back into painting" becomes questions such as which beginner paint set to buy and how to mix skin tones, then topics such as watercolor technique and supply reviews, then publisher types such as art instruction blogs and hobby magazines. Only then does anyone search for sites, using the approach on the lookalike websites hub or a topic search.
Which persona traits map to content topics?
IAB Tech Lab's implementation guidance (opens in a new tab) for its taxonomies separates the categories that describe a page's "aboutness" from "an additional set of orthogonal content attributes / "vectors" such as content language, format." That split is a useful frame for a persona. Some traits describe what the persona wants to read about, some describe how they like to read it, and some describe the person in ways no page can confirm.
| Persona trait | Maps to | Fictional illustration | Use in site selection? |
|---|---|---|---|
| Activity or hobby | Topic (aboutness) | Paints in watercolor on weekends | Yes, the core of the search |
| Questions the persona asks | Article types and subtopics | How to stop paper from buckling | Yes, as the test each site must pass |
| Skill level | Depth of coverage | Returning beginner | Yes, beginner tutorials over technical treatises |
| Preferred format | Format attribute | Step by step posts with photos | Yes, as a secondary filter |
| Language | Content language | Reads in English | Yes, checked on the articles themselves |
| Age, life stage or generation | Nothing on the page | Recently retired | No; keep it in the creative brief |
| Income, wealth or spending power | Nothing on the page | Comfortable budget | No; never translated into sites |
| Health, beliefs, orientation, ethnicity | Nothing the plan should use | Any such detail | No; remove it from the media brief |
Google's own description of contextual targeting (opens in a new tab) works on the same principle: "Google's system analyzes the content of each webpage to determine its central theme." The same page notes that Google also weighs "your language and location targeting, a visitor's recent browsing history, and other factors," so a platform's contextual campaign is not the same as a list built from content alone. A persona list checked page by page keeps the selection on the subject.
What persona claims should a planner avoid?
Some persona details look harmless but describe a person the plan cannot see. The FTC staff report A Look Behind the Screens (opens in a new tab) (September 2024) warns that the assumptions and inferences behind targeted advertising can "lead to individuals being misidentified or misclassified," and its footnote 194 adds that targeting "based on categories that are not sensitive on their own but can be combined to produce proxies for sensitive categories can also result in consumer harm." A persona that stacks age, neighborhood and spending habits is that kind of combination.
Platform rules draw similar lines. Google's policy on restricted targeting in personalized advertising (opens in a new tab) lists sensitive interest categories such as "Health", "Negative financial status", "Religious beliefs" and "Sexual orientation", and, for the US and Canada, access to opportunities categories: "Consumer finance", "Employment" and "Housing". That policy governs personalized ads, not content lists, but a persona leaning on those traits is a warning sign for any plan. In Canada, the Office of the Privacy Commissioner's behavioural advertising guidelines (opens in a new tab) name "health and financial data, ethnic and racial origins" among sensitive information.
Persona claims to rewrite before the search
- No sentence says the readers of a site are wealthy, older, younger, healthy or ill.
- Income and spending power are removed; a premium product is described by the subjects its buyers read about.
- Life stage appears only as a subject the persona reads about, such as retirement planning articles, and only when the brief is about that subject.
- No trait from a sensitive category is used to choose or rank a site.
- Every remaining trait can be checked by reading a page.
- The persona's questions are written down in the persona's own words, ready to test sites against.
How do you check sites against a persona?
The persona's questions become the test. A site fits when its recent articles answer those questions at the right depth and in the right format, whatever anyone assumes about its readers.
A five part persona fit test for each candidate site
- Read five recent articles. Note what each is about in one sentence and how many answer at least one persona question.
- Check the depth. Confirm the coverage suits the persona's skill level, such as beginner tutorials for a returning painter.
- Check the format. Confirm the site publishes the formats the persona prefers, such as step by step posts with photos.
- Check the language on the page. Google's placement report FAQ explains that "the language targeting selects the language of the user and not the content of the site," so read the articles' language directly.
- Record the verdict and the evidence. Write fit, partial fit or no fit, with the two articles that decided it.
| Result | What it means | Next step |
|---|---|---|
| Answers the questions, right depth and format | Core fit for this persona | Run the standard quality, suitability and seller checks |
| Right topic, wrong depth | Covers the subject for experts or for casual readers only | Keep only if the persona has a second skill level |
| Right topic, occasional coverage | A general site with an occasional matching story | Seed with the section, or leave it out |
| Answers questions the persona never asked | Adjacent subject | Keep apart as an idea to test, not on the core list |
Delivery reports come later and answer a different question. Google's Where ads showed report (opens in a new tab) helps judge whether a placement "is a good fit for your brand message," but warns that its performance data "is generally not statistically significant" at low volume. Use it to confirm content after launch, not to infer who the readers were.
How many personas should one brief carry?
The practical limit comes from the search: each persona brings its own questions and topics, and a list built for four personas at once is pulled toward general sites that serve none of them well. The lookalike drift guide shows the same pull toward the middle when seeds are mixed.
How does PlacementGraph turn a persona brief into sites?
What does persona based selection look like on a real brief?
Fictional example
A fictional watercolor paint maker in Alaska
An unnamed Anchorage maker of watercolor paints and sketchbooks hands its agency three personas for a winter display campaign geotargeted to Alaska: a returning painter who stopped after school and wants to start again, a daily sketcher who fills a journal with drawings of food and city scenes, and an affluent retiree with time for classes.
The planner rewrites the third persona. Affluent is removed, since no page can show it. Retiree is removed from selection and kept in the creative brief. What remains is a need: someone looking for classes and structured lessons, which merges into the returning painter.
Two themes go to search. The returning painter's questions become beginner technique, paint and paper reviews and lesson series. The daily sketcher's become sketchbook reviews, quick sketch methods and illustration journals. Each candidate gets the five part test. A large general crafts portal answers the questions in a handful of articles out of its sample and is left out; an Alaska arts magazine's workshop section passes and is seeded by section. The list labels each site with its persona theme.
Frequently asked questions
Can a persona come from the client's customer database?
The persona can be written from what the client learned in aggregate, such as the hobbies and questions customers mention in reviews. The site search itself needs no customer file and no upload. Hand the planner the persona's needs in plain words, and keep records about individuals out of the media brief entirely.
Should persona names appear in the plan the client sees?
Yes, as labels. Tag each site with the persona theme it serves and the questions it answered in the test, so the client can see which part of the plan supports which persona. Avoid any wording that describes the readers of a site as that person, since the plan only checked the content.
Does a persona replace the topics to avoid in a brief?
No. A persona says what to look for; the avoid list says what the brand will not appear beside. A returning painter persona does not tell the planner whether the client accepts sites that cover art market scandals. Write both, and check each candidate against both.
What if two personas want the same sites?
That is a sign they are one persona with two descriptions. Compare their questions; if the same articles answer both sets, merge them and keep the richer set of questions. If only a few sites overlap, keep the personas apart and let those sites carry both labels.
Sources
- IAB Tech Lab Taxonomies: implementation guidance (opens in a new tab)IAB Tech Lab (GitHub)
- Contextual targeting (opens in a new tab)Google Ads Help
- A Look Behind the Screens: Examining the Data Practices of Social Media and Video Streaming Services (opens in a new tab)Federal Trade Commission staff report, September 2024
- Restricted targeting in Personalized advertising (opens in a new tab)Google Advertising Policies Help
- Guidelines on privacy and online behavioural advertising (opens in a new tab)Office of the Privacy Commissioner of Canada (modified 2025-08-11)
- Privacy Implementation Notice 2024-01: Digital Advertising (opens in a new tab)Treasury Board of Canada Secretariat (effective April 2, 2026)
- About the "Where ads showed" report (opens in a new tab)Google Ads Help
Researched and written by PlacementGraph Research. Read how we research, source and update these pages. Corrections: diana@placementgraph.com.