Headline Analyzer – Score Any Title, Subject Line or Video Title
A headline is the smallest piece of writing with the largest consequence. Most readers see it and nothing else, so it decides whether the work behind it gets read at all — and yet most headlines are judged by feel, written last, and shipped unrevised. This headline analyzer replaces the guesswork with a score out of 100 and, more usefully, with the six sub-scores that produced it, so you can see exactly which part of the headline is weak instead of being told that it is.
The six things it measures
Length is scored against the surface the headline will appear on, because the budgets differ wildly: roughly 600pixels — about 60 characters — for a Google title, about 40 characters before a phone truncates an email subject line, and about 60 before YouTube cuts a video title in mobile search. The score is a triangular curve peaking at each platform's ideal, so the same 41-character line scores full marks as a subject line and loses points as a blog title.
Word balance checks the mix. Headline research consistently finds the same shape: around 20–30% common words, 10–20% uncommon, 10–15% emotional, and at least one power word. Common words carry the grammar and stop the line reading like a telegram; uncommon words are what stop the eye on a page of similar titles; emotional words move it from informative to compelling; a power word — proven, effortless, definitive — is the promise the reader is clicking for.
Headline type detects the structure: list, how-to, question, comparison, warning, quote or plain statement. Lists and how-to headlines travel further than generic statements because they promise a bounded, specific payoff. Emotional Marketing Value reports emotional words as a share of the whole, split into the classic intellectual, empathetic and spiritual appeals. Readability applies Flesch Reading Ease and the Flesch–Kincaid grade level to the headline itself. SEO checks an optional focus keyword for presence, position and repetition — position matters most, because snippets are read left to right.
Truncation, previews and the parts nobody checks
Two headlines of identical character count can truncate differently, because Illinois is narrower than Wallmount. The tool estimates rendered width from a per-character width table and cuts the Google SERP snippet preview at the real pixel boundary, then shows what was hidden. Email subject lines get an inbox preview at both phone and desktop widths; YouTube, X, LinkedIn and Facebook get mock cards at their own limits.
Alongside that it runs the checks that quietly cost engagement: a spam-trigger dictionary for subject lines, a clickbait detector for stacked superlatives and curiosity-gap phrasing, ALL-CAPS and exclamation-stacking flags, and a first-three / last-three word scanning-zone analysis — because readers do not read headlines, they scan the ends. Each word is rendered as a colour-coded chip, and clicking a weak one offers stronger swaps that rewrite the headline and re-score it instantly.
Comparing variants instead of chasing a number
The score is most reliable as a relative measure, which is what A/B comparison mode is for: up to six variants ranked side by side against the same platform, keyword and dictionaries, with a per-metric table showing which one wins on length, emotion and keyword placement. Bulk mode takes up to 200 headlines pasted one per line and returns a scored table with CSV export, which is how an editorial team audits a content calendar in one pass rather than one title at a time.
In practice the fastest gains are structural. Adding a number turns a generic statement into a list; moving the focus keyword from word nine to word two costs nothing and fixes the SEO sub-score outright; swapping one vague word for a specific one moves balance, emotion and the scanning-zone reading at once. Length is usually the last thing to tune, not the first.
What is estimated, and how
Syllables — and therefore the grade level — come from a table of spelling traps plus a vowel-group heuristic, so a word can be off by one. Because a headline is a single sentence, its Flesch–Kincaid grade is dominated by word count and runs lower than the same formula on prose; the readability sub-score therefore awards full marks at or below your target level and only penalises headlines that read harder. The word lists are curated for this tool rather than licensed, so absolute EMV percentages will not match a paid grader, and every dictionary-driven measure is English-only — a non-Latin headline gets its length, structure and truncation scored and the rest skipped, with the weight redistributed so the total is still out of 100.
Everything runs in your browser: no upload, no request, no signup, so an unpublished or embargoed title never leaves the machine. Results export as a plain-text report or as CSV, and a share link reopens the headline and settings exactly as you left them.