Keyword Research From Scratch: No Data, No Tools

Almost every keyword research guide starts the same way: open Google Search Console, pull your top queries, sort by impressions. That advice is useless if your site launched last week. GSC only shows queries tied to impressions your site has already earned, so a brand-new domain, or one that is not live yet, opens to an empty report.

This guide skips that step entirely. It covers four free, no-signup methods for finding real keywords before you have a single data point: Google Autocomplete, the "People Also Ask" and "Related Searches" boxes on a results page, the forums and communities where your actual customers already talk, and a look at what topics your competitors already consider worth covering. None of it requires a login or a credit card.

Quick Navigation

• The Problem With Most Keyword Research Advice
• Method 1: Google Autocomplete
• Method 2: People Also Ask and Related Searches
• Method 3: Where Your Customers Already Talk
• Method 4: What Competitors Already Cover
• Turning Raw Phrases Into a Keyword List
• What Changes Once You Have Real Traffic
• Frequently Asked Questions

---

The Problem With Most Keyword Research Advice

Most keyword research guides assume a resource a new site does not have: existing search performance data. Google Search Console can only report on queries that already generated an impression for your domain, so a site with no history, or no live pages yet, has nothing to pull.

Paid keyword tools solve a different problem. They estimate search volume by sampling clickstream and API data across millions of sites, then charge a monthly fee for access to that estimate. A new business with no marketing budget yet does not need an estimate of how many people search a term nationally. It needs to know what real, current searchers actually type, in their own words, right now. That is a different problem, and it has a free solution.

Method 1: Google Autocomplete

Google Autocomplete predicts full search queries as you type, and Google's own documentation confirms those predictions reflect real searches other people have already run, filtered by language, location, and current trends.

How to Use It

1. Open Google in an incognito or private browser window (this avoids your personal search history skewing the suggestions).
2. Type a broad "seed" term related to your business — a product, a service, or a problem you solve — and stop before pressing enter.
3. Write down every suggestion Google offers. These are phrases with real search demand, not guesses.
4. Add a single letter after your seed term (your seed term plus "a", then "b", then "c", and so on through the alphabet) and repeat. This surfaces suggestions Google would otherwise hide behind the first few results.
5. Try the seed term with common question words in front of it: "how," "what," "why," "best," "cheap," "near me."
6. Try it with the seed term at the end instead of the start: "[question] for [your seed term]."

A dog-walking business launching a new site might seed with "dog walker" and surface suggestions like "dog walker insurance," "dog walker rates near me," and "dog walker vs pet sitter." None of that requires a login or a paid account — it is Google showing you, directly, the phrasing real searchers already use.

Tips

• Run the same seed term on Google Images and Google Maps search boxes too. Each surface pulls from a different prediction set and often turns up phrasing the main search box does not.
• Mobile autocomplete and desktop autocomplete sometimes differ slightly. If most of your customers search on a phone, check there too.
• Keep a running spreadsheet as you go. You will generate more raw phrases than you expect, and half the value of this method is having them all in one place to sort later.

Method 2: People Also Ask and Related Searches

"People Also Ask" boxes and the "Related Searches" list at the bottom of a results page are both built from real query data tied to the search you just ran, and both are visible without any account or subscription.

How to Use It

1. Search your seed term normally, without incognito mode this time — you want the full results page, not just the box.
2. Look for the "People Also Ask" box, usually a few results down. Click one question to expand it, and Google will load two or three additional questions beneath it. Keep clicking. This can surface a dozen or more related questions from a single starting search.
3. Scroll to the very bottom of the results page and record every phrase in the "Related Searches" grid.
4. Click into a few of those related searches and repeat the process. Related searches often branch into a completely different angle on the topic than your original seed term did.

Where Autocomplete captures how people start typing a query, "People Also Ask" and "Related Searches" capture what people wonder about once they are already looking at results on that topic. That distinction matters for content planning: Autocomplete phrases tend to make good page titles, and "People Also Ask" phrases tend to make good subheadings and FAQ entries within that page.

Tips

• "People Also Ask" boxes are dynamic and change based on what you clicked. If you expand a question, the box refreshes with new ones related to what you just read — treat it like a rabbit hole and follow it a few levels deep.
• Some searches show no "People Also Ask" box at all. That is a real signal too; it usually means the topic is more transactional than informational.

Method 3: Where Your Customers Already Talk

Reddit threads, niche forums, and public Facebook groups contain the exact language real customers use to describe their problem, which is often noticeably different from the phrasing a business would choose to describe the same thing.

How to Use It

1. Search "[your industry] reddit" or search directly within Reddit for subreddits related to your niche, your local area, or the problem you solve.
2. Read threads where people ask for recommendations or describe a problem your business solves. Do not skim for keywords — read the actual sentences people write when they are frustrated, confused, or looking for help.
3. Do the same in relevant Facebook groups (local community groups and industry-specific groups both work) and in niche forums that still exist in your industry.
4. Note the exact phrases, not paraphrases. A customer asking "why does my dog pull on the leash every single walk" is different from the tidier "leash training tips" a business might assume people search for.

This is the method most likely to surface language that a business would never think to target on its own, because it comes from outside the business's own vocabulary. A plumber thinks in terms of "water heater installation." A customer with a cold shower at 6 a.m. searches "why is my hot water suddenly gone."

Tips

• Pay attention to the words customers use to describe the wrong solution, not just the right one. Complaints about a bad past experience often contain the exact search phrase someone would type before choosing a business — "how to tell if a plumber is overcharging you," for example.
• Local Facebook groups are especially useful for local businesses. National subreddits skew toward national brands and general advice; a city or neighborhood group skews toward exactly your market.

Method 4: What Competitors Already Cover

Reviewing the page titles, headings, and FAQ sections on a competitor's site tells you which topics that competitor considered important enough to invest time writing about, which is a useful signal even without knowing whether any of it ranks well.

How to Use It

1. Identify two or three direct competitors — businesses that solve the same problem for the same customer, in your area or your niche.
2. Visit their blog, resource, or FAQ pages and list every distinct topic covered, not every individual post. Multiple posts often circle the same underlying topic.
3. Look specifically at their page titles and H2/H3 headings. These represent the phrases the competitor's own SEO effort was built around, whether or not that effort worked.
4. Compare that list against your own Autocomplete, "People Also Ask," and forum research from the earlier steps. Topics that show up in a competitor's content and in real search behavior are strong candidates. Topics a competitor covers that never showed up anywhere else in your research may be worth a second look, or may simply not be a topic your future customers search for.

This is not about copying a competitor's content. It is a fast way to sanity-check your own list against what someone else in the same market already decided was worth writing about, and to catch topics your other methods missed entirely.

Tips

• Check competitor sitemaps directly (usually at competitor-domain.com/sitemap.xml) for a complete list of every page they have published, including ones not linked prominently from their main navigation.
• A topic covered by every competitor in your niche is probably table-stakes content you need too, not a differentiator. A topic covered by none of them, but that shows up repeatedly in your forum research, is often the more interesting opportunity.

Turning Raw Phrases Into a Keyword List

By this point you have a spreadsheet full of raw phrases from four different sources, and the next job is sorting them by what a searcher actually wants, not by how many words they contain.

Group each phrase into one of three buckets:

• Informational: the searcher wants to learn something ("why does my dog pull on the leash")
• Commercial: the searcher is comparing options ("best dog walker near me reviews")
• Transactional: the searcher is ready to act ("book a dog walker today")

Then group phrases within each bucket by topic. Several differently worded phrases usually point at the same underlying topic — "dog walker rates," "how much does dog walking cost," and "dog walker pricing near me" are all one topic, not three.

Those topic groups are the raw material for a content plan. Our guide on structuring content into topic clusters covers how to turn a list like this into a set of pillar pages and supporting posts instead of one page per keyword phrase, which is a stronger structure once you get past a handful of pages. For a broader look at how page structure and internal linking work together on a growing site, see our content architecture guide.

What Changes Once You Have Real Traffic

Everything above works with zero data because it observes public search behavior instead of your own site's performance, and that is also its limit: it tells you what people search for in general, not what is already working for your specific site.

Once a site has been live long enough to generate real impressions and clicks, Google Search Console becomes the better source, because it shows your actual keyword data instead of an inference about likely search behavior. That is the point where the process in what keywords should I target takes over: finding keywords you already rank for on page two, keywords with high impressions but a low click-through rate, and other opportunities visible only in your own performance history.

Brass-SEO reads that Google Search Console data directly, cross-referenced against Google Analytics 4 traffic data, and surfaces those opportunities without manual spreadsheet work. Both GSC and GA4 need to be connected for that next-level analysis to run — the four free methods in this guide are what get a site to the point of having that data to analyze in the first place.

---

Try Brass-SEO Once You Have Data to Analyze

The methods above work with nothing but a browser and time. Once your site is live and generating search impressions, Brass-SEO connects to your Google Search Console and Google Analytics 4 accounts and answers plain-English questions about your actual keyword performance — no spreadsheets required. It costs $25/month with a 3-day free trial, and setup takes about two minutes once both accounts are ready to connect.

---

Frequently Asked Questions

Do I need any paid tools to do keyword research on a brand-new site?

No. Google Autocomplete, "People Also Ask," "Related Searches," relevant forums and Facebook groups, and competitor content review are all free and require no account or signup. Paid keyword tools add search volume estimates, which are useful once you are choosing between similar-sized opportunities, but they are not required to build an initial keyword list for a new site.

How many keywords should I have before I start writing content?

There is no fixed number, but aim to group your raw phrases into at least five or six distinct topics before writing your first page. A single page built around one narrow keyword phrase, with nothing else planned around it, tends to leave obvious follow-up questions unanswered. Grouping phrases into topics first gives you a small content plan instead of one page in isolation.

Is Google Autocomplete the same on every device?

Not always. Autocomplete predictions factor in language, location, and trending interest, and mobile and desktop can surface slightly different suggestions for the same seed term. If most of your customers are likely to search from a phone, check autocomplete on a mobile browser as well as desktop.

Should I trust competitor content as a keyword source?

Treat it as one signal among several, not the primary one. A competitor's page titles and headings show what that business decided was worth writing about, but that decision may have been a good one or a poor one. Cross-check competitor topics against what you found through Autocomplete, "People Also Ask," and forum research — topics that show up in multiple sources are the stronger bets.

When should I switch from these free methods to Google Search Console?

As soon as your site is live and has accumulated a few weeks of search impressions. GSC data reflects your actual performance rather than general search behavior, which makes it more precise once it exists. The methods in this guide are designed to get a new site to that point. They are the starting line, not a permanent substitute for first-party data.