What is Word Error Rate (WER)?
Definition
Word Error Rate (WER) is the standard metric for measuring the accuracy of speech-to-text and transcription systems. It calculates the percentage of words incorrectly transcribed by comparing the AI output to a reference human transcription.
How It Works
WER is calculated by counting the total number of substitutions (wrong words), insertions (extra words), and deletions (missing words), then dividing by the total number of words in the reference text.
A WER of 5% means the system gets 95 out of every 100 words correct. State-of-the-art systems typically achieve WER between 3-10% depending on audio quality, accent, and domain.
glossaryWordErrorRateHowItWorks3
Why It Matters
WER is how transcription quality is objectively measured. Lower WER means more accurate transcriptions and better meeting notes.
When comparing meeting AI tools, WER helps evaluate which system will produce the most accurate transcriptions for your specific use case.
glossaryWordErrorRateWhyItMatters3