Keyword Research: Finding Demand Worth Building On

Everything before this article was preparation. Keyword research is where the business actually starts — it is the skill of finding demand that is real, reachable, and worth building on. Get this right and even a mediocre site earns; get it wrong and the best-built tool sits at zero traffic.

Search volume comes from consensus

The core theory is simple: a keyword has search volume because a group of people collectively agreed — without coordinating — to describe the same need with the same words. Nobody voted on it. Millions of individual phrasings converged on a few winning phrases, and those phrases are the keywords.

This matters because it tells you where new keywords come from: new consensus forming around new things. A product launches, a meme spreads, a regulation appears — and within weeks, thousands of people are typing the same new phrase into Google. Your job is to spot the consensus while it is forming, not after it has settled.

A working definition: a “new keyword” is one that appeared within roughly the last 12 months. Older than that, and incumbents have had time to entrench. Newer than that, and you are racing the clock described in the “too late” article — the window is open, but it narrows fast.

Finding new keywords: six methods

1. Google Trends, related queries. Look up a seed topic in Google Trends and check the “related queries” section set to “rising.” Breakout terms there are consensus forming in real time, measured directly from search behavior. This is the single highest-signal free method — start here before anything else.

2. Root-word lists. Keep a list of productive roots — Translator, Generator, Convert, Calculator, Checker, Maker, Remover — and systematically combine them with trending nouns. New tools, new media formats, and new platforms constantly create fresh combinations nobody has built for yet. The method is mechanical, which is its virtue: you can generate hundreds of candidates in an afternoon and then filter.

3. Social platforms. TikTok, Reddit, and X are where new consensus is born. A phrase that starts appearing across posts — a new AI model name, a viral workflow, a slang term — will become a search keyword within weeks. Watch what people repeat. Social is upstream of search; by the time a phrase shows clean data in keyword tools, the early window is already closing.

4. Big sites’ new pages via Similarweb. Large sites publish constantly; Similarweb shows which of their new pages are gaining traffic. A new page on an established site catching fire reveals a keyword with proven demand — and you can serve it with a more focused tool than their generic article. Let their content teams do your discovery for you.

5. Vercel.app subdomains. Builders deploy experiments on *.vercel.app subdomains, and some of them catch search traffic. Browsing what is getting indexed there is a window into what other indie builders are testing right now — free competitive intelligence on which ideas are gaining traction.

6. Year and date keywords. Queries containing years (“best X 2026”) or dates refresh annually by definition. They are seasonal and competitive, but they are predictable: you can prepare content ahead of the cycle instead of chasing it. Mark the calendar and build early.

Validating: is it worth building on?

Finding a candidate keyword is half the job. Before you write a line of code, validate:

Estimate volume with a reference word. Google Trends shows relative interest, not absolute numbers. Pick a keyword whose volume you roughly know, compare your candidate against it in Trends, and scale proportionally. If your reference gets around 10,000 monthly searches and your candidate shows half its interest, you are looking at roughly 5,000. Crude, but it separates “worth building” from “ten searches a month” reliably.

Check the trend shape. A healthy keyword shows steady or rising interest. Exclude two patterns: seasonal spikes that collapse to nothing (unless you deliberately want a seasonal site), and fading spikes — one viral moment decaying toward zero. You are looking for a line that will still be above zero a year from now. Screenshots of the trend curve beat any single volume number.

KGR: the Keyword Golden Ratio. Divide the number of Google results with the keyword in the page title (the intitle: operator) by the keyword’s monthly search volume. Below 0.25 is considered easy — few dedicated competitors relative to demand. It is a rough heuristic, not physics, but it filters out obviously contested keywords in seconds.

KD from Ahrefs. Ahrefs’ Keyword Difficulty estimates how many backlinks you would need to crack the top 10, based on the link profiles of current rankers. For a new site with no link profile, treat high-KD keywords as someone else’s problem. The metric exists to tell you which fights to decline.

Always inspect the top 10 manually. Numbers lie by omission; the actual search results do not. Open Google and look at what is ranking: are they homepages or deep inner pages? (Inner pages are easier to displace.) Are brand terms dominating? (Harder.) Do the results actually serve the query well, or is the top 10 full of forum threads and outdated posts? (Weak results mean opportunity.) Ten minutes of manual inspection beats any metric — this is the step beginners skip and veterans never do.

The first-mover advantage is real

GeFei cites a telling case: for a newly heated keyword, a .com domain with 43 backlinks outranked a .org with 145. The difference was not links — it was timing. The .com registered and published right as the keyword heated up, earning relevance and freshness signals while the consensus was still forming. The .org, despite triple the links, arrived after the window had narrowed.

That is the whole game in miniature: spot the forming consensus, validate quickly, ship immediately. Speed here is not hustle-culture posturing; it is the literal mechanism of the advantage. Every week of delay is a week for someone else’s page to age into the authority slot you wanted.

With a validated keyword in hand, the next step is turning it into a live site — building and launching, which the next article covers.

---

Series: Build Overseas Websites with GeFei | Previous: Choosing Your Niche: What Kind of Site to Build | Next: Build and Launch: Getting Your Site Live

Based on GeFei’s methodology (哥飞): https://github.com/qiayue/opengefei