How keyword research works

Quibo scores every keyword it discovers and surfaces the most attractive ones first:

• How many people search it: monthly Google search volume for the site's language and market.
• How hard it is to rank: a 0 to 100 difficulty score (lower is easier).
• What's already ranking: we take a snapshot of the pages a web search returns for the keyword and look for shortcuts. Forum threads (Reddit, Quora, Stack Exchange) among the top results mean a beatable page; a Wikipedia page among the top three means a hard one. The snapshot is a proxy for Google's page one, not a live Google result. Informational queries get a small boost because new content ranks for them more easily.

These combine into the Opportunity score, the default sort of the table: high volume and low difficulty win, a weak page one pushes a keyword up, a page one owned by Wikipedia pushes it down.

Each run also labels its results:

• Gem: in the top quarter of that run by opportunity, with a beatable page one (difficulty 35 or less, forum results on page one, or no Wikipedia page in the top 3) and at least 50 searches a month. Start here.
• Long-tail: four or more words, or a three-word question. Lower volume, stronger intent.
• Standard: everything else.

What you actually do

1. Click + Discover keywords. Pick the site, type a seed (for example marketing automation, red wine pairing, ceramic frying pan) and set Max keywords this run (5 to 60; the default is set from the articles you have left). Click Start discovery.
2. Wait while it runs. The dialog shows how long recent runs took; a run takes longer while articles are being generated. You can leave the page: the result shows above the table, and the table updates on its own when the new rows land.
3. Use the Gems, Long-tail and Standard pills and the volume, difficulty and status filters to narrow the list. Tick the keywords you would actually write about and click Accept, or use the check icon on a single row.
4. The banner above the table counts your accepted keywords against the articles you have left. Click Plan this month to generate. See Monthly generation.

That's it. No formulas to tune, no knobs to turn. The defaults work well for B2B SaaS, consumer SaaS, e-commerce, and service businesses. If you have a special case, contact us.

What a run does behind the scenes

Each run expands the seed into related concepts using your brand profile, pulls suggestions and related keywords with their volume and difficulty, drops brand names, misspellings, past-year and near-duplicate phrases, mines top-ranking pages for long-tail phrases when the topic is narrow, drops anything clearly off-topic for your brand, and takes a snapshot of page one for the keywords it keeps. Phrases already in the library for that site and language are skipped, and so are phrases that target the same search as one already there, so a run never creates duplicates. A phrase dated to a past year does not block this year's version: with best ai detectors 2025 on the site, a run can still add best ai detectors 2026.

A run stops at Max keywords this run or when your keyword quota runs out, whichever comes first: it adds only as many new keywords as remain, from the monthly quota first and then from any extra keywords you bought. Only keywords new to your account count. A phrase that is already in your account, on any of your sites and in any status, never uses quota again when a run returns it. Owners, admins and editors can run discovery.

Research runs per month

Each plan includes a fixed number of discovery runs per cycle, on top of the keyword quota:

• Solo and Starter: 1 run a month.
• Growth: 2 runs a month.
• Agency and Business: 4 runs a month (weekly).
• Lifetime tiers 1 and 2: 1 run a month. Tier 3: 2 runs. Tier 4: 4 runs.
• Free: 1 run in total. Like the rest of the Free allowance, it never resets.

The counter next to + Discover keywords shows how many runs you have used this cycle. A run counts only when it adds at least one keyword: runs that add none, fail or stop are not counted. Once they are spent the button is disabled until your next cycle starts, which resets the counter together with the article and keyword quotas: on each renewal for monthly subscriptions, on the monthly anniversary for annual subscriptions and lifetime tiers (see Plans and quotas). Runs are also rate limited to 5 every 5 minutes. Manual additions and CSV imports never use a run.

Used vs. not used

Plan this month takes the highest-opportunity accepted keywords for each site, up to the count you set, skipping accepted keywords that target the same search as another one in the plan or as an article you already have, and marks them used. Two keywords target the same search when they differ only by plural, by a form of the same word (ai checker, ai checking, ai check), by small words or by word order that does not change the meaning (impact of ai on economic growth, ai impact on economic growth), or when their page one shares at least 4 of its top 10 pages. Word order that changes the meaning keeps them apart (flights rome milan and flights milan rome, dog food for cats and cat food for dogs), and an article dated to a past year does not block this year's version. The table marks such keywords with a Same search pill. Used keywords are never suggested again. Accepted keywords the plan did not reach stay accepted for the next plan; discovered and rejected keywords stay in the library with their scores and sort alongside the next run's results. See Monthly keyword cycle.