How to Do Keyword Research Without Paying for Ahrefs

Do you actually need Ahrefs to do useful keyword research?

Short answer: no, not for the first one to three years of a small site. Ahrefs and SEMrush are excellent at scale, but their core value (massive backlink and keyword indexes) compounds for sites publishing hundreds of pages a month or analysing thousands of URLs. A solo founder with twenty pages and one article a week does not have the throughput to use ten percent of what they surface. The data you actually act on is small: queries your audience types into Google, questions on Reddit, and terms showing impressions in your site console. All three are free at the source. The trick is not finding more keywords, it is picking the ten that fit your business this quarter and writing those well, which is a workflow problem rather than a data problem.

What free signals beat paid keyword data for small sites?

Five free signals together cover almost everything a paid tool delivers for a site under fifty pages. Google autocomplete shows you the live phrasing real searchers use, including long-tail variations a static database cannot guess at the same freshness. People Also Ask surfaces the exact questions Google itself thinks are related, which is pure gold for FAQ blocks and for mapping intent. Google Search Console reveals what queries your own site is already showing for, including impressions you have never noticed, which is the single highest leverage source on this list. Reddit, niche forums, and community Q and A sites expose pain in the user's own words before it ever becomes a polished search query. And the SERP itself, read carefully, tells you what format Google rewards (long guide, listicle, comparison, video) before you commit to a draft.

Can AI cluster intent and uncover gaps without a paid tool?

Yes, and this is where the workflow stops looking like a chore. A capable AI marketing employee can take a raw list of two hundred autocomplete and People Also Ask queries, group them by underlying intent, label each group as informational, commercial, or transactional, and rank the groups by how well they fit your actual product. The same employee can read your top five competitor pages for each cluster, summarise what they cover, and flag the subtopics every competitor misses. That last step is where small sites win: gaps are not in the volume column of a paid tool, they are in the silence between competing articles. A free pipeline of Google plus an AI employee finds those silences faster than a $129 subscription, because the employee reads pages, while a tool only counts them.

1. Collect raw queries — Pull autocomplete, People Also Ask, Search Console, and Reddit threads for your seed topic into one list.
2. Cluster by intent — Ask the AI employee to group queries by intent and label each as informational, commercial, or transactional.
3. Score against your product — Rank each cluster by fit with what you sell, so research maps to revenue rather than vanity traffic.
4. Read the SERP — Have the employee read the top five ranking pages per cluster and list every subtopic they share.
5. Flag silences — Ask what no competitor covers, what is outdated, and what a founder voice can uniquely add.

The workflow above is easy to write down, but founders rarely run it because every step demands a different surface (browser tabs for autocomplete, Search Console for impressions, a spreadsheet for the list, a notebook for clusters). The friction kills the habit by week two. Hiring an AI marketing employee collapses the five steps into one chat thread, where the same coworker pulls the queries, clusters them, reads the SERP, and hands back a ranked list. That single change moves keyword research from a chore to a short conversation.

Once you have a ranked list of clusters, the next risk is writing the wrong article. A keyword can look perfect on paper, score high on every free signal, sit in a clear gap, and still flop because the SERP rewards a format you cannot credibly produce, or the audience does not buy anything. Validation is the cheap insurance step solo founders skip and regret. The next section is the short checklist I run before any draft begins.

How do you validate before writing?

Validation is a five-minute habit that prevents most wasted drafts. Check the top ranking pages first: personal blogs, big brands, or directories. If Forbes and Wikipedia dominate the SERP, do not waste your time. Check format: are the rankers long guides, listicles, or comparison tables, and can you credibly produce that format with depth. Check freshness: if every page is over a year old and the topic moves fast, an updated article wins almost by default. Check buying intent: would the searcher plausibly use your product within ninety days, or is this pure traffic with no revenue path. And check your own credibility: can you write this from experience, or are you paraphrasing what already ranks. Two failing is a soft no. Three failing is a hard pass.

1. Check who is ranking — Big brands and directories on page one mean a small site has no realistic shot. Pick where personal sites still rank.
2. Check the format — If Google rewards a long guide and you can only ship a 600-word blog, drop the query. Match the rewarded format.
3. Check freshness — Old top results in a fast-moving topic mean an updated piece wins on recency alone. Note the dates first.
4. Check buying intent — Would a searcher plausibly use your product within ninety days? If not, the article is decoration, not revenue.
5. Check your own credibility — Write only what you can back with experience, first-hand data, or a perspective competitors cannot copy.

What is a one-hour weekly keyword routine that compounds?

Keyword research breaks when it becomes a quarterly sprint instead of a small weekly habit. The shape that compounds is one focused hour, the same hour every week, run as a five-step ritual. Skim Search Console for queries that gained impressions on page two or three. Run autocomplete and People Also Ask on the seed terms you covered recently. Drop the raw list into your AI marketing employee for clustering and gap notes. Pick one cluster to commit to for the coming week. Brief the writing employee in the same conversation, so research and draft live on one thread. Sixty minutes, fifty-two times a year, and the back catalogue grows on its own.

1. Skim Search Console — Open the past seven days and look for queries gaining impressions on page two or three. Cheapest wins.
2. Refresh autocomplete — Run live autocomplete and People Also Ask on seed terms you covered recently. Copy raw results into one list.
3. Hand to AI for clustering — Drop the list into your AI marketing employee. Ask for intent clusters, gap notes, and a ranked recommendation.
4. Pick one cluster — Commit to one cluster for the week. Saying no to nine clusters is what makes the habit sustainable.
5. Brief in the same thread — Brief the writing employee inside the same conversation. Research and draft stay glued together.

Frequently asked questions

Keyword research is the upstream half. The downstream half is publishing the article, the social post, and the follow-up, week after week, without burning out. That second half is where most solo founders quietly stall, because one piece is easy and fifty is brutal. If you want the playbook for delegating that whole loop to a small AI marketing function, the next read is the companion piece.

The honest framing on keyword research is the one I keep coming back to on every small-business question: tools never make a founder ship. A free pipeline of Google signals plus an AI marketing employee replaces the paid stack for years, but only if the habit gets run weekly, briefs get written tightly, and articles get published whether or not they feel perfect. Paid tools are a luxury earned after the workflow compounds, not a prerequisite for starting it. Pick a seed topic this week, spend the hour, hand the cluster to a coworker that can actually draft, and let the back catalogue do the slow work of ranking. Almost everything else in this category is decoration on top of that single weekly hour.