Hook: Every paragraph of the draft comes back scored, and the weak ones come back rewritten. One careful read against the buyer costs a senior editor EUR 68 to EUR 135 an article.
What You'll Be Able To Do
- Score a draft paragraph by paragraph against one buyer and one topic cluster, both taken from your own scan
- Read the four numbers that hand you a send-back list before you have read a word of the draft
- Open the rewrite on any weak paragraph, with the reason it scored low sitting underneath it
- Keep every run in Content Fit History, filtered by persona and by cluster, so a decision made today still has a record in six weeks
- Re-check a page that has been live for months, which is the run almost nobody makes
Runtime: 4:40 | Difficulty: Intermediate
You sign off on drafts. Every article waits for you to read it against the buyer it was written for, mark the passages that miss, and send it back with a note. You are the queue. Your standard slips when you are tired, and weak paragraphs ship anyway.
What it costs today
Assumption: a senior editor at EUR 90 an hour.
| Line item | Time | Cost |
|---|---|---|
| One careful read against a persona document, weak passages marked | 45 to 90 min an article | EUR 68 to 135 |
| Ten articles a month | 7.5 to 15 h | EUR 675 to 1,350 |
| A content-grading seat, per person | monthly | EUR 90 to 450 |
Two editors reading the same piece mark different paragraphs, so the standard moves with whoever is on shift. The grading seats score a draft against the pages that already rank in search; they hold no model of your buyers and no map of your own topics. Marketer is EUR 196 a month against EUR 68 to EUR 135 an article, so the plan clears its own cost on the second article at the dear end and the third at the cheap end.
What the video shows
- Content generation opens in the sidebar, then Content Fitter. One panel does the whole job: pick who the piece is for, pick where it sits on your topic map, paste the draft.
- The Cluster picker lists the topic groups the scan found on the brand's own website, each one carrying its coverage bar. Pick the cluster this draft belongs to.
- The Persona picker lists the buyers the same scan produced, name and role. Pick the one the draft was written for, so the score is against your audience and your topics.
- Paste the draft into the content box, headings and all. Rankfor.AI cuts it into chunks of roughly 500 characters, close to a paragraph each, and scores every chunk on its own.
- Click Run Content Fitter. Each chunk is compared against every page in that cluster and against the persona, and the two scores are blended by what this buyer decides on.
- Four numbers fill in: Overall Fit Score, Overall Confidence, Good segments, Need Improvement. The last one counts every paragraph that did not land, amber ones included, and it is your send-back list.
- The segment stack scrolls. Every paragraph carries a badge, GOOD FIT, MODERATE FIT or NEEDS IMPROVEMENT, with its fit score and its confidence. Green cards say no improvements needed, so leave them alone.
- See improvements opens on a weak card. The rewrite appears in a panel headed IMPROVED, with the reason the original scored low in small type below it. That is what you hand the writer.
- The History pill opens the log. Every run is kept and filtered by persona and by cluster, each row showing the score, the two names and the date.
- The closing frame holds the report at full width: four numbers across the top, one rewrite open below.
What you need
Marketer, EUR 196 a month. Content generation opens on Marketer and above, and a free-tier scan produces a single persona, which makes the picker look broken.
You need a project whose scan has finished, including the metrics pass, because that pass is what produces the personas and clusters the two pickers offer. You need the draft on your clipboard. Six or more segments makes a stack worth scrolling, so aim for roughly 3,000 characters. One content-fit job runs at a time on a scan; a second one is refused while the first is in flight.
Questions people ask
Does the Loose, Medium and Strict setting change my score? No. The picker is stored with the run and never reaches the scoring. Good is 70 and up, needs improvement is below 50, on every run. The History row prints the word you picked, which makes that easy to misread, so leave it on Medium.
Can I upload a file? No. Content Fitter takes pasted text and nothing else. There is no export either, so the report is the screen and the History row.
Is this an SEO score? No. It measures how close each paragraph sits to the buyer you picked and to the topics already on your website. It says nothing about search rankings, and the confidence number tells you how sure Rankfor.AI is of that fit.
The page came back at 81% with nothing flagged. Wasted run? No. Above the threshold means the page still matches the buyer and the topic map you picked. Your voice moves as the year's articles pile up, so a page written in January can fall out of voice by October without a word of it changing. Running a page that has been live for months is the check almost nobody does.
Watch the four minutes, then paste the draft sitting in your inbox.