How do you find websites by topic when there is no seed site to start from?
Part of Lookalike websites
The short answer
To find websites by topic without a seed, write a topic map of subtopics, reader questions and specialist terms, search with exact phrases, exclusions and date limits, mine the resource pages of clubs and associations, then sample recent articles on each candidate. Keep real publishers that mostly cover the topic, and log where and why each site was found.
Key takeaways
- Without a seed site, the topic map does the job a seed would: it defines what an on topic article looks like before any searching starts.
- Exact phrase searches on specialist vocabulary target the words specialists use, while broad subject terms also match general sites.
- Clubs, associations and event organizers can link to the publications their members read.
- A site earns a place on the candidate list only after a sample of its recent articles shows the topic is its main subject, not an occasional story.
- Recording the source, query and date of every find makes the list reviewable and lets the next planner repeat the search.
How do you find websites about a topic without a seed site?
Some briefs arrive without any example publications. The client is new to the open web, the category is new, or the client simply does not know who writes about its subject. The lookalike websites hub and the find similar websites guide start from sites a client already trusts. This method starts from the topic itself and builds the list from nothing.
The first step is a written topic map. It replaces the seed: a planner who cannot point to an example site can still describe, in plain sentences, the articles the brand wants to appear beside. The map also exposes disagreements early. If the client says "astronomy" and means backyard stargazing while the planner hears space industry news, the map shows it before an hour of searching is wasted.
| Part | What to write | How it is used |
|---|---|---|
| Core subject | One sentence describing the ideal article, such as a hands on review of a beginner telescope | The yardstick for sampling: each candidate article either fits this sentence or does not |
| Subtopics | Five to ten narrower themes the core subject breaks into | Each subtopic gets its own searches, so one theme does not crowd out the others |
| Reader questions | The practical questions the audience asks, in their own words | Searched as exact phrases; publishers that answer them in depth are strong candidates |
| Specialist terms | Vocabulary an outsider would not use, such as product parts, techniques or event names | Distinctive terms that point the search toward specialist coverage |
Which search methods surface specialist publishers?
Broad subject terms can return large general sites near the top. Specialist publishers are easier to find when the search uses the words specialists use. Google's search help (opens in a new tab) documents the operators that make this practical: quotation marks for an exact phrase, a minus sign to exclude a word, site: to limit results to one domain, and before: and after: to limit results by the date a document was last updated.
Six search moves for building a topic list from scratch
- Search specialist terms in quotes. Put each term from the topic map in quotation marks, alone and paired with a reader question, and note every independent publication in the first pages of results.
- Exclude the shops. Add a minus sign before words such as cart, price or shipping to push store pages down, then repeat without the exclusion to catch publishers that review products.
- Limit to recent work. Add after: with a date about a year back so that abandoned blogs drop out and active publishers remain.
- Search inside the community. Look for the resources, links or recommended reading pages of clubs, societies, trade associations and event organizers in the field, and list the publications they point to.
- Follow the bylines. When a strong article appears, check where else its writer publishes; a specialist writer may contribute to several outlets in one field.
- Check each find with site:. Run site: on the candidate's domain with a specialist term to see how much of the site covers the topic, before opening any articles.
How many articles should you read before shortlisting a site?
Enough to tell whether the topic is the site's main subject or a passing interest. A practical working rule is to read the five most recent articles first, then read five more only if the result is mixed. Recent articles matter more than the best article, because a plan buys the site as it is now, not the story that made it famous.
The site's own feeds make sampling fast and fair. An RSS feed (opens in a new tab) lists items with a pubDate that "indicates when the item was published" and can carry category elements that place each item in one or more categories. Taking articles in the order the feed lists them avoids the bias of picking only the stories that look relevant from the homepage. Score each sampled article against the core subject sentence from the topic map: fits, partly fits or does not fit.
How do you tell a publisher from a store or brand site?
Topic searches also return sites that are about the subject but are not publications: retailers' buying guides, manufacturers' blogs, marketplaces and forums. Some are useful to know about, but a media plan needs publishers that sell advertising beside their own editorial work. The page markup can say which is which. Schema.org defines a Product (opens in a new tab) as "any offered product or service" with an offers property, which is what store pages carry, while a NewsMediaOrganization (opens in a new tab) can declare policies such as a masthead, an ethics policy and a corrections policy.
Signs a topic site is a publisher
- Articles carry bylines and dates, which Google's helpful content guidance (opens in a new tab) encourages wherever readers might expect them.
- The site covers products or services from many makers rather than one company's catalog.
- Most pages are articles, guides or reviews, not product listings with prices and a cart.
- An about or masthead page names an editorial team or an owner separate from any retailer.
- The site publishes on a regular schedule, visible in its feed or archive.
- The site shows advertising or has an ads.txt file, which suggests it sells ad space.
- Original photos, measurements or reporting appear in the articles, not only manufacturer copy.
Edge cases need a decision, not a rule. A retailer that runs a genuinely independent magazine section may be worth a direct conversation but may not be buyable programmatically. A forum with lively discussion may suit the audience while failing a brand's suitability rules on user content. Record these as set aside with a reason, so they can be revisited if the plan needs more scale.
How do you record why each site was found?
A topic list built without a seed is only as defensible as its notes. When a client asks why a site is on the plan, the answer should be one line: the query or source that surfaced it, the date, and the sampled articles that proved the fit. The same notes let another planner rerun the search next quarter and see what changed.
| Field | Example entry | Why it matters |
|---|---|---|
| Domain | examplestargazing.example | The unit that will be planned and bought |
| Found by | Exact phrase search for a specialist term, or a society's recommended reading page | Shows whether the find came from search ranking or from the community |
| Query or source URL | The exact query text, or the page that linked to the site | Lets anyone repeat the step |
| Date found | The day the search ran | Search results and sites change; the date frames the evidence |
| Subtopic | Which part of the topic map the site matched | Keeps the list balanced across subtopics |
| Sample result | Articles read and how many fit the core sentence, with two links | The proof of topic fit, readable by the client |
| Status and reason | Candidate, set aside or left out, with one line of reason | Explains every exclusion as well as every inclusion |
How does PlacementGraph find websites from a written brief?
What does topic first discovery look like on a real brief?
Fictional example
A fictional telescope retailer in Ontario
An unnamed Ottawa retailer of telescopes and astronomy cameras wants English display ads across Ontario ahead of the autumn viewing season. It has never bought open web media and cannot name a single publication. The planner writes a topic map with one core sentence, a hands on review or observing guide for amateur astronomers, and subtopics for beginner telescopes, astrophotography, dark sky sites and sky event calendars.
Exact phrase searches on specialist terms such as mount types and camera sensor terms surface a dozen independent astronomy blogs, a magazine and several equipment stores. Excluding cart and price drops most of the stores. The recommended links pages of amateur astronomy clubs add four publications the searches missed, including a dark sky travel blog.
Sampling five recent articles per site removes a space industry newsletter whose recent pieces were about launch contracts rather than observing, and a photography blog that had stopped publishing. The discovery log shows each remaining site's source, date, subtopic and two matching articles, and the list moves on to quality and seller checks.
Frequently asked questions
Should a topic list include large general sites at all?
Sometimes, as a section rather than a whole domain. A national outlet with a regular science column may suit an astronomy brief if the buying system can target that section. Record it separately from the specialist publishers, because its topic fit depends on the section, not the site.
Can AI chat assistants help find websites by topic?
They can suggest starting points, but every suggestion must be opened and checked. Assistants can name sites that no longer publish, have changed subject or never existed. Treat any suggested domain as an unverified lead and log it with its source like any other find.
How long does topic first discovery take compared with starting from a seed?
It can take longer, because the topic map has to be written and agreed first, and searches on several subtopics replace one comparison against a known site. The time is front loaded: once the map and log exist, later refreshes of the same topic are quick.
What if the topic searches return very few publishers?
Widen in steps and label each step. Add neighboring subtopics, include publications in the other languages the brief allows, and look at newsletters or podcasts with companion websites. If the list is still short, tell the client; a short, honest list is better than one padded with loosely related sites.
Sources
- Refine Google searches (opens in a new tab)Google Search Help
- Google Search operators (opens in a new tab)Google Search Central
- RSS 2.0 Specification (opens in a new tab)RSS Advisory Board
- Product (opens in a new tab)Schema.org
- NewsMediaOrganization (opens in a new tab)Schema.org
- Creating helpful, reliable, people-first content (opens in a new tab)Google Search Central
Researched and written by PlacementGraph Research. Read how we research, source and update these pages. Corrections: diana@placementgraph.com.