How to Find Winnable Keywords for a New Website (When You Have Zero Domain Authority)

A founder opens Ahrefs for the first time and types in their main product category. The difficulty score reads 72. The next ten ideas all sit above 40. The site launched three weeks ago and has two backlinks, both from free directories.

Most new sites stall right there. This guide shows you how to find winnable keywords for new websites without relying on authority you don't have yet. The process is built on reading the results page yourself, checking intent, and picking battles a small site can win.

The Goal

By the end, you'll have a ranked list of 10 to 15 keywords your new site can realistically win. Each one will have a checked SERP, a confirmed intent, an AI Overview note, and a one-line reason it's winnable. You'll know which to write first.

Before You Start

• A live, indexed site. Search site:yourdomain.com in Google. If nothing shows up, fix indexing before you do keyword work.
• Google Search Console, verified. It's free. Use a Domain property so you capture every subdomain and protocol.
• One keyword tool. Ahrefs' free Keyword Generator or a free Semrush account is enough to start. Paid plans for both start above $100/month, and you don't need them for this tutorial.
• A spreadsheet. Google Sheets works. Create these columns: Keyword, Source, Volume, KD, SERP Notes, AI Overview (Y/N), Intent, Format, Verdict.
• A Chrome incognito window. You'll use it to see results without your own search history skewing them.
• Time. Plan 3 to 4 hours for the first pass. After that, 30 minutes a week keeps the list fresh.
• Knowledge. You need to know your product and your customers. You don't need an SEO background.

Step 1: Your Best Seed Keywords Come From Customers

Seed keywords are the short phrases you'll expand later. Starting from customer language keeps your list tied to what you sell. It also surfaces wording that keyword tools miss, because the search volume looks too small for them to report.

1. List every product, service, or core offer in 2 to 3 words. A Shopify store selling linen bedding might write "linen sheets," "linen duvet cover," and "stonewashed linen."
2. Pull 10 real questions from support emails, DMs, sales calls, or product reviews. Copy the wording exactly, including the clumsy parts.
3. Search your category on Reddit. Add thread titles that people keep posting. "Are linen sheets worth it" is a perfect seed.
4. Add problem phrases. These describe what the customer is trying to fix before they find you. For the linen store, that's "sheets too hot at night."
5. Put each seed in the Keyword column and note its source.

✅ Check: You have 20 to 30 seeds. At least a third come straight from customer wording, not from a tool.

Step 2: Long-Tail Questions Are Where a New Site Can Compete

Long-tail keywords are longer, more specific phrases with lower search volume. They matter for new sites because large sites rarely bother with them. This step expands each seed into dozens of specific questions and modifiers you have a real shot at.

The case for the long tail is strong. BrightEdge (updated 2022) reports that long-tail keywords make up over 90% of search queries. That figure is old, and the method behind it isn't clear. Still, the direction matches what any search data shows: most searches are specific.

The cost of ignoring this is also clear. Ahrefs (2023) found that 96.55% of pages get no organic traffic from Google, based on roughly 14 billion pages. A big share of those pages chase terms nobody searches or terms the site can't win.

1. Type each seed into Google. Write down the autocomplete suggestions.
2. Add a letter after the seed ("linen sheets a," "linen sheets b") to pull more suggestions.
3. Expand the People Also Ask box. Each click loads two or three more questions.
4. Paste your seeds into Ahrefs' free Keyword Generator or Semrush's Keyword Magic Tool. Open the "Questions" view.
5. Keep phrases with 4+ words or a clear modifier: best, vs, for, cost, near me, how to, alternative, worth it.
6. Record volume and KD for each. Don't delete zero-volume keywords yet.

Difficulty scores have a blind spot you need to understand. Ahrefs' Keyword Difficulty is a 0 to 100 score based mainly on referring domains pointing to the top 10 results (Search Atlas, 2026). It measures links, not whether your site can rank. Moz's Domain Authority has a similar limit. Google's John Mueller has said repeatedly that DA is a third-party metric Google doesn't use for crawling, indexing, or ranking.

Semrush has responded to this gap. Its Personal Keyword Difficulty (PKD%) scores a keyword against your specific domain instead of an average site. That helps, but a brand-new domain gives it little to work with.

✅ Check: You have 60 to 100 candidate keywords. Most are 4+ words or carry a buying modifier.

Step 3: A Low KD Score Means Nothing Until You Check the SERP

Difficulty scores estimate backlinks, not beatability. A KD 8 keyword can still be locked up by Amazon and Wirecutter. This step has you read the actual results page for each candidate and judge whether a zero-authority site could break into it.

This is the core of how to find winnable keywords for new websites. Nothing replaces looking.

1. Open a Chrome incognito window. Search the keyword.
2. Scan the top 10 organic results. Skip ads and shopping carousels.
3. Note any weak signals from the table below.
4. Check suspicious domains with Ahrefs' free Website Authority Checker. A site with a few dozen referring domains is beatable.
5. Count the weak signals. Three or more means winnable. One or two means maybe. Zero means skip.
6. Write a one-sentence reason in SERP Notes. Example: "Two Reddit threads in top 6, #3 last updated 2022."

| Weak signal | What it tells you | |---|---| | Reddit, Quora, or forum thread ranking | Google lacks a dedicated page that answers the query well | | Small site with few referring domains | Links aren't what's holding the top spots | | Page last updated 2+ years ago | A current, accurate page has room to replace it | | Page answers a different question | The intent is underserved | | Short page on a question that needs depth | A thorough answer can outperform it |

Forum results are the clearest signal. A study from February 2024, cited by QuickSEO, found Reddit appearing on 7,509 of 10,000 product-review queries. Google showed a "Discussions and forums" feature on 77% of them. Reddit takes up space on those pages. When a thread ranks #4, Google is often telling you no specialist page exists yet.

✅ Check: Every candidate has a verdict and a written reason. Expect 15 to 30 to survive.

Step 4: AI Overviews Can Erase the Clicks You'd Win

Some easy keywords now send almost no traffic, even from position one. Google answers the question in an AI Overview above the results. This step flags those keywords so you spend your first months on queries that still produce visits.

The numbers are hard to ignore. Ahrefs data from December 2025, reported by MediaNama in February 2026, shows click-through for the top result drops about 58% when an AI Overview appears. An earlier Ahrefs estimate in 2025 put the drop at 34.5%. Seer Interactive, reporting in Search Engine Land in November 2025, found organic CTR on informational queries with AI Overviews fell 61% since mid-2024.

1. For each surviving keyword, note whether an AI Overview appears. Mark Y or N.
2. Read the overview. Does it fully answer the query? "What is stonewashed linen" gets a complete answer. "Best linen sheets for hot sleepers" usually leaves the reader wanting a comparison.
3. Deprioritize keywords marked Y where the answer is complete and the intent is purely informational.
4. Look at which sources the overview cites. If small sites appear there, being cited is a realistic goal.

My recommendation: for your first 20 articles, lean toward keywords with no AI Overview or a clear buying angle. Keep a few cited-source opportunities, though. Tools like CocoSEO track citations across ChatGPT, Perplexity, and Google AI Overviews, so you can see whether that visibility turns into anything.

✅ Check: The AI Overview column is filled in. No more than a third of your top picks are informational queries that get a complete AI answer.

Step 5: Search Intent Decides Whether a Keyword Pays You Back

Intent is the job the searcher wants done. Ranking for a keyword with the wrong intent brings visitors who leave. This step labels each keyword and matches it to the page format Google already rewards, so your article fits before you write a word.

| Intent | Example modifiers | Format that usually ranks | Value to a new store | |---|---|---|---| | Informational | how to, what is, why | Guide or explainer | Low to medium | | Commercial | best, vs, review, alternative | Comparison or buyer's guide | High | | Transactional | buy, price, discount | Product or collection page | High, but hard | | Local | near me, in [city] | Google Business Profile, local page | High for local businesses |

1. Look at the format of the top 5 results. Note whether they're listicles, product pages, guides, or videos.
2. Label the intent in your sheet.
3. If the top 5 are all store category pages, a blog post won't rank. Mark it for a collection page or skip it.
4. Move commercial keywords like "linen vs percale for summer" toward the top.
5. Write the expected format in the Format column.

Don't write thin pages just because a keyword looks easy. The Helpful Content system has been part of Google's core algorithm since March 2024 (Krawly, 2026). Small niche, recipe, and travel publishers took heavy losses during that period (impact.com, 2024). An easy keyword with a weak article is a liability.

✅ Check: Each keyword has an intent label and a format. You can name the exact page you'd build for it.

Step 6: Page-2 Rankings Are Your Fastest Wins

Once your site has a few months of data, Search Console shows keywords where Google already sees you as relevant. Positions 11 to 20 are the sweet spot. You rank #11, and one good article or update can win page 1.

1. Open Search Console. Go to Performance > Search results.
2. Set the date range to the last 3 months.
3. Turn on Average position and Impressions at the top.
4. Open the Queries tab. Add a filter: Position, greater than 10.
5. Sort by Impressions. Ignore anything with an average position above 20.
6. Click a query, then open the Pages tab to see which URL ranks.
7. Choose an action. If the ranking page fits the query, update it with a direct answer and a better title. If it doesn't, write a dedicated article.
8. Add internal links to that page from 2 or 3 related pages on your site.

Know the limits. The Queries table shows at most 1,000 rows and leaves out rare, anonymized queries (Analytify, 2026). A new site may show very little for its first couple of months. That's normal and not a sign that something is broken.

✅ Check: You have 5 to 10 page-2 queries, each with a planned action. If you don't have that data yet, put a reminder on your calendar for 8 weeks out.

Step 7: A Short, Focused List Beats a Long Scattered One

You now have more winnable keywords than you can write this quarter. This step trims them into one or two topic clusters and puts them in order. That way each article supports the next, and your site builds focus in one area.

1. Group keywords by theme. "Linen care," "linen vs other fabrics," and "sleeping cool" are separate clusters.
2. Pick the cluster closest to revenue.
3. Order your list. Put commercial keywords with 3+ weak signals first. Page-2 updates come next. Supporting informational pieces go last.
4. Schedule 1 to 2 articles per week.
5. Link every new article to at least two others in the same cluster.

Set honest expectations. An older Ahrefs study (around 2017, cited by Whitehat SEO) found only 1.74% of new pages reached the top 10 within a year. The average #1 page was about 5 years old, and only 22% of top-10 pages were under a year old. The data is dated, but the lesson holds: plan in months, not weeks. Expect volatility during core updates too. Google's March 2026 core update ran from March 27 to April 8 with no specific guidance published (Search Engine Journal, 2026).

Writing a reason for every keyword is the habit that makes this work. CocoSEO attaches that reasoning to each suggested topic automatically. Here's what an example topic card looks like:

Topic: best linen sheets for hot sleepers Why it's winnable: Two Reddit threads in the top 6. The #3 result was last updated in 2022. No AI Overview on this query. Intent: Commercial. Top results are comparison lists. Suggested format: Buyer's guide with a comparison table and cited sources. Fit: Matches three products in your catalog.

You approve or skip it in about 60 seconds. You decide based on the reasoning, not a bare difficulty score.

✅ Check: You have 10 to 15 keywords in order. Each has a written reason, and the first four are on your calendar.

Troubleshooting: Fixes for How to Find Winnable Keywords for New Websites

Why does every keyword in my tool show zero volume?

Zero search volume in a keyword tool is usually a reporting gap, not proof that nobody searches. Keyword tools under-report long-tail phrases, so a new website should confirm demand through Google autocomplete, People Also Ask boxes, and Reddit threads before discarding a phrase that matches its offer.

If Google autocompletes the phrase, a People Also Ask box shows the phrase, or Reddit threads keep asking the same question, real people search for it. Write the article when the searcher's intent matches your offer.

Why hasn't my "easy" keyword moved after 90 days?

A stalled "easy" keyword is a page that has not climbed after 90 days despite a low keyword difficulty (KD) score. The usual causes are new competitors entering the results or a KD score that hid strong pages, so the first step is to run the SERP check again.

Next, add internal links from your strongest pages to the stalled page. One relevant link from another site pointing to that specific page can do more for the stalled page's ranking than general authority building.

Why do I rank on page 1 but get almost no clicks?

A common cause of a page-1 ranking with almost no clicks is a Google AI Overview that answers the query before searchers reach your link. Search the keyword yourself to check. If an AI Overview appears and answers the query fully, rewrite your title to promise what the overview can't deliver.

Good title angles include a comparison or a specific use case. If no AI Overview appears for the keyword, shift effort toward commercial keywords instead.

Why does Search Console show only a handful of queries?

A short query list in Google Search Console is normal for a brand-new site. Search Console data takes time to accumulate, so give the site 6 to 12 weeks of publishing before judging keyword performance. Until then, filter Search Console reports by page instead of by query to see more complete data.

Page-level data is less affected by two Search Console limits: the 1,000-row cap and anonymized queries. Query-level reports leave out anonymized queries, while page-level totals still reflect the clicks and impressions those hidden queries generate. Once a specific page starts showing impressions, open that page's query list to see which long-tail phrases Google already associates with the page. Use those phrases to plan the next article or to expand the existing one, because the phrases come from real searches rather than tool estimates.

Why do visitors arrive but nobody buys?

Visitors who arrive but never buy signal an intent mismatch: the keyword attracts people researching a topic, while the page expects buyers. The fix is to compare the keyword's intent label in your keyword tool with what the page actually offers, then add a clear path from the page to the relevant product.

Informational keywords bring readers who are still learning, so a page targeting an informational keyword should guide readers toward the product page rather than leave them with no next step. After fixing existing pages, rebalance your keyword list toward commercial-intent formats: "best" keywords, "vs" keywords, and "for [use case]" keywords. Searchers using those phrases are comparing options, which makes those keywords a better match for a page designed to sell.

What to Do Next

Your first cluster of keywords is the start of a system. Next, structure those articles so they link together and reinforce each other. Our guide on building topic clusters for small sites walks through that step by step. After that, add factual answer blocks under your headings. Short, direct answers raise your chances of being quoted in Perplexity and AI Overviews. For more on that, read how to get cited in AI search engines. Then repeat this process every month as Search Console fills in.

Further reading

• Google Search Central, "Creating helpful, reliable, people-first content": the standard your articles will be judged against.
• Ahrefs, "SEO Statistics": current benchmarks on traffic, page age, and CTR, updated regularly.
• Semrush's Personal Keyword Difficulty documentation: explains how PKD% scores keywords against your own domain.
• Google's Search Console Performance report help page: covers filters, data limits, and how anonymized queries are handled.
• Seer Interactive's AI Overview CTR research (via Search Engine Land, Nov 2025): the clearest data on how AI answers change the value of informational keywords.

Sources

1. 96.55% of Content Gets No Traffic From Google. Here's How to Be in the Other 3.45% [New Research for 2023] — Ahrefs, 2023. Supports: 96.55% of pages get no organic traffic from Google, based on about 14 billion pages.
2. Ahrefs Study: Google AI Overviews Cut Clicks By 58% — MediaNama, 2026. Supports: Ahrefs data from December 2025 shows the top result's click-through rate falls about 58% when an AI Overview appears.
3. Google AI Overviews drive 61% drop in organic CTR, 68% in paid — Search Engine Land, 2025. Supports: Seer Interactive found organic CTR on informational queries with AI Overviews fell 61% since mid-2024.
4. Keyword Difficulty: How to Score and Pick Winnable Terms — Search Atlas, 2026. Supports: Ahrefs' Keyword Difficulty is a 0–100 score based mainly on referring domains pointing to the top 10 results.
5. John Mueller Rebuts Idea that Google Uses Domain Authority Signal — Search Engine Journal, 2018. Supports: Google's John Mueller has said Google doesn't use Moz's Domain Authority for ranking.
6. How to find low-competition keywords with Semrush — Semrush, 2025. Supports: Semrush's Personal Keyword Difficulty (PKD%) scores a keyword against your own domain instead of an average site.
7. Google Confirms March 2026 Core Update Is Complete — Search Engine Journal, 2026. Supports: Google's March 2026 core update ran from March 27 to April 8, and Google called it "a regular update" without publishing specific guidance.
8. A deep dive into Search Console performance data filtering and limits — Google Search Central Blog, 2022. Supports: The Search Console interface exports at most 1,000 rows, and the report table leaves out anonymized queries even though chart totals include them.