How does the title tag CTR predictor calculate its score?
It applies a rule-based scoring model with eight signals that each map to a delta against a baseline of 50. Length (50-60 chars = +20, 40-50 or 60-70 = +10, otherwise -10), contains a number (+10), power words like "best" or "proven" (+5 each, capped at 15), brackets/parens (+5), recent year ≥ 2025 (+5), emotional triggers (+5 each, capped at 10), all-caps words >2 chars (-5 each, capped at -15), and a brand suffix like "| Brand" (+3). The total is clamped to 0-100, then mapped to a predicted CTR by multiplying by 8% - our roughly calibrated upper bound for organic SERP CTR on featured queries.
Where do these scoring rules come from?
They're distilled from public CTR studies - Backlinko's 11.8M-headline analysis, Advanced Web Ranking's position-1 CTR curves, and HubSpot's headline emotional-value research. The signals (numbers, power words, brackets, freshness markers) all show double-digit CTR lift in those studies. The model is intentionally simple so the score is explainable; a black-box ML model would predict slightly better but you couldn't see why a title scored what it did.
Why does the predicted CTR cap at 8%?
Because organic CTR for the #1 result on a typical informational query in 2024 averaged around 27% - but for transactional or branded queries it can be much lower (2-8%) due to ad blocks and SERP features (featured snippets, People Also Ask, Knowledge Panels). 8% is a realistic upper bound for an arbitrary SaaS / ecommerce title in a competitive niche. Adjust your expectations upward for branded "Brand X login" queries and downward for "what is" queries dominated by featured snippets.
Does the score guarantee a higher click-through rate?
No. CTR depends on factors this tool can't see: your search position, the SERP features Google shows above you, your meta description, your domain's familiarity, and even the user's location and device. A high score means your title has more of the patterns associated with high CTR, not that those clicks will materialise. Treat it as a copywriting checklist, not a forecast - and always A/B test in Search Console once the page has indexed.
Why does my title get truncated in the SERP preview?
Google truncates title tags at roughly 580 pixels wide, not at a fixed character count - wide letters like W, M, and capital letters take more pixels than thin letters like i and l. We approximate at 9 pixels per character, which means a 65-character title with average letter widths usually fits, but the same length in all caps may be cut. The fix is either trimming the title or moving the most important keyword to the front so it survives truncation.
Is anything sent to a server when I use this tool?
No. The entire predictor runs in your browser. Your title strings - which may include unreleased product names, internal codenames, or campaign drafts - never leave your device. There's no analytics on the input box, no server round-trip, no logging. You can paste sensitive copy from a launch deck without privacy concerns; close the tab and the input disappears.
