# Voice drift · writing glossary · ScriptGrain > Voice drift is the gradual movement of a writer's, brand's or client's published writing away from its own established habits: sentences lengthen or flatten, contractions disappear, the register formalises, new fillers creep in. It is measurable as the change in voice-match score between a baseline and later pieces, and it is what multi-writer teams, ghostwriters and AI assistants most reliably introduce. Canonical: https://scriptgrain.com/glossary/voice-drift # Voice drift Voice drift is the gradual movement of a writer's, brand's or client's published writing away from its own established habits: sentences lengthen or flatten, contractions disappear, the register formalises, new fillers creep in. It is measurable as the change in voice-match score between a baseline and later pieces, and it is what multi-writer teams, ghostwriters and AI assistants most reliably introduce. ## The numbers - Voice-match band below which a piece reads as drifting from its reference: **under 0.80** ([ScriptGrain: how Voice Match is calculated](https://scriptgrain.com/reference/voice-measurement-framework)) - Pages ScriptGrain's voice monitor scores against a profile each month on the Writer and Operator plans (15 on Studio, 50 on Agency): **5** ([ScriptGrain pricing](https://scriptgrain.com/pricing)) - Mean voice match of GPT-5 drafts written from a one-page measured prompt, against 0.83 for drafts from the profile (SGR-002): **0.62** ([SGR-002: ChatGPT vs a measured voice](https://scriptgrain.com/research/chatgpt-vs-a-measured-voice)) ## What voice drift is Drift is the gap between how a brand used to write and how it writes now, when nobody chose that gap on purpose. A [voice match](https://scriptgrain.com/glossary/voice-match) score tells you how close two pieces of writing sit to each other. Drift is what you call it when that score falls over time, and the fall wasn't a decision anyone made. ## Where drift comes from Most drift has a boring cause. Several writers work on one brand, each with their own habits, and the average creeps somewhere nobody intended. A ghostwriter brings their own rhythm to a founder's voice, quietly, post by post. AI drafts flatten sentence variance and strip out contractions, because that's what unmeasured generation tends to do. A new editor arrives with different instincts about commas and length. The brand starts writing for a new audience and the tone shifts to match. And sometimes a writer's own voice just changes over years: that's not an error, it's growth, and it isn't drift to correct. ## How drift is measured You need a baseline first. That's either a profile built from a writer's own back catalogue, or one reference piece that everyone agrees sounds right. Later pieces get scored against that baseline. Drift is the fall in that score, plus the named attributes that moved: sentence length, contraction rate, formality, the lot. The score returns per-feature deltas, so you're not told "it's off", you're told what changed. The bands, for a pairwise comparison: 0.80 and above is on voice. 0.55 to 0.79 is drifting. Under 0.55 is off voice. ScriptGrain's voice monitor scores up to 5 pages a month on the Writer or Operator plans, 15 on Studio, 50 on Agency, against a set profile, and emails the movement when it happens. If you just want a quick check across clients with no account, the [multi-client voice drift check](https://scriptgrain.com/tools/multi-client-voice-drift-check) compares up to three pieces against a reference for free. For the mechanics behind the scoring itself, see the [voice measurement framework](https://scriptgrain.com/reference/voice-measurement-framework). ## What it cannot tell you A score tells you what moved. It doesn't tell you why, and it doesn't tell you whether the move was wrong. A brand voice can shift on purpose over years and that's fine; the numbers won't distinguish "growth" from "drift" without a human reading both pieces. And a single low score against one reference piece is a data point, not a verdict. Treat it as a prompt to look closer, not a final judgement. Agencies managing several brand voices at once find this especially useful; see [for agencies](https://scriptgrain.com/for-agencies) for how that scoring gets applied across clients rather than one brand. ## Worked example a client's reference piece measures average sentence length of 15 words, contractions at 22 per 1,000 words, and a formality score of 4. Three months later, a new writer has produced a run of posts. Those measure 21 words average sentence length, 4 contractions per 1,000 words, and formality at 6. the comparison sits between 0.55 and 0.79, which lands it in the drifting band, not off voice. Three habits explain most of the movement: sentences have got longer, contractions have nearly disappeared, and the tone has formalised. None of these is dramatic on its own. Together, they're enough to notice. drift is specific, so the fix is specific. Nobody needs to "sound more like the brand" in the abstract. Someone needs to shorten sentences, put contractions back in, and loosen the tone by two points. That's a note you can hand to a writer, not a feeling you have to argue about. For more on how these thresholds get set and applied consistently, see [brand voice measurement](https://scriptgrain.com/reference/brand-voice-measurement). ## Sources - [ScriptGrain: how Voice Match is calculated](https://scriptgrain.com/reference/voice-measurement-framework) - [SGR-002: ChatGPT vs a measured voice](https://scriptgrain.com/research/chatgpt-vs-a-measured-voice) - [ScriptGrain: brand voice, measured](https://scriptgrain.com/reference/brand-voice-measurement) ## More terms - [Voice Match](https://scriptgrain.com/glossary/voice-match) - [Burstiness](https://scriptgrain.com/glossary/burstiness) - [Perplexity](https://scriptgrain.com/glossary/perplexity) - [Sentence Length Variance](https://scriptgrain.com/glossary/sentence-length-variance) - [Flesch Reading Ease](https://scriptgrain.com/glossary/flesch-reading-ease) - [Type-token ratio (TTR)](https://scriptgrain.com/glossary/type-token-ratio) - [Hapax legomenon](https://scriptgrain.com/glossary/hapax-legomenon) - [Function words](https://scriptgrain.com/glossary/function-words) - [Lexical density vs lexical diversity](https://scriptgrain.com/glossary/lexical-density-diversity) - [Stylometric fingerprint](https://scriptgrain.com/glossary/stylometric-fingerprint) - [Idiolect](https://scriptgrain.com/glossary/idiolect) - [Voice vs tone vs style](https://scriptgrain.com/glossary/voice-tone-style) - [Passive voice](https://scriptgrain.com/glossary/passive-voice) - [Humanised AI content](https://scriptgrain.com/glossary/humanised-ai-content) - [Tone matching](https://scriptgrain.com/glossary/tone-matching) - [Brand voice](https://scriptgrain.com/glossary/brand-voice)