The commonest mistake in affiliate content is choosing keywords by volume. High-volume terms are high-volume because everyone wants them, and a new site will not win them.
The useful question isn't "how many people search this" but "can I rank for this, and will the people who find it buy anything".
Intent comes first
Sort every keyword into one of four buckets:
Transactional — "buy X", "X pricing", "X discount". Ready to purchase. Rare, competitive, converts best.
Commercial investigation — "best X", "X vs Y", "X review", "X alternatives". Deciding between options. This is the sweet spot for affiliate content: high intent, more volume than transactional terms.
Informational — "what is X", "how does X work". Learning. Converts poorly on its own but builds topical authority and captures people earlier.
Navigational — "X login", "X dashboard". Looking for a specific site. Worthless to you.
Aim the bulk of your effort at commercial investigation, use informational content to build authority around it, and don't chase navigational terms at all.
Then filter by winnability
Difficulty scores from SEO tools are a rough guide. Reading the actual results page is better, and free.
Search your target term and ask:
- Who ranks? If the first page is all major publications and established review sites, pick something else.
- Are there forum or community results? Reddit or Stack Overflow ranking on page one signals that no one has published a good dedicated resource. That's an opening.
- How old is the content? If the top results are three years stale, current content has a real advantage.
- Does the content actually answer the query? Frequently the top result is a thin listicle that doesn't address what was asked. That gap is your opportunity.
- How many ads are there? Heavy advertising means commercial value, which means competition — but also that the traffic is worth having.
Where the winnable terms hide
Add a qualifier. "Best CRM" is impossible. "Best CRM for freelance designers" is not, and the searcher is far closer to buying.
Go geographic. "Best hosting Australia", "European payment processors". Global sites handle these badly.
Go by use case. "Email marketing for podcasters", "accounting software for Etsy sellers".
Target comparisons. "X vs Y" terms are lower volume and dramatically higher intent — someone comparing two specific products is close to a decision.
Target alternatives. "Alternatives to X" catches people actively dissatisfied and shopping.
Chase problems, not products. "Why is my site slow" reaches people who will shortly need hosting, a CDN, or an optimisation tool.
Watch for new products. A tool launched three months ago has almost no content written about it. Being early is one of the few reliable advantages available to a small site.
Building a content plan around clusters
Isolated articles rarely rank. Groups of related articles that link to each other do, because they demonstrate depth on a topic.
Pick a narrow area — say, email marketing tools for creators — and plan:
- One broad piece: "best email marketing tools for creators"
- Several comparisons: "X vs Y", "Y vs Z"
- Several alternatives pieces: "alternatives to X"
- Supporting informational pieces: "how email deliverability works", "list segmentation basics"
- Practical pieces: "migrating your list from X to Y"
Link them to each other, and link them to the relevant category pages in your directory. Ten interlinked articles on one topic will outperform fifty scattered ones.
Free ways to find keywords
- Google autocomplete. Type your seed term and read the suggestions. Add a letter and read them again.
- "People also ask". Every question is a heading, and often an article.
- Related searches at the foot of the results page.
- Reddit and niche forums. Real questions in the words people actually use.
- Review sites for the products. Complaints in reviews are keywords: "X is too expensive" becomes "affordable alternatives to X".
- Search Console, once you have traffic. The most valuable source there is — terms you already rank for on page two are the cheapest wins available.
A realistic target profile
For a site under a year old, look for terms with modest volume, weak or outdated top results, clear commercial intent, and at least one page-one result that isn't a major brand.
You will not find many. That's fine — ten such articles, properly executed and interlinked, will outperform a hundred pieces targeting terms you cannot win.