Feedback Panel
A reference guide to every metric in the SpeakNow session feedback panel.
The Feedback panel appears on the right side of the lesson screen during a live session and updates in real time. After the session, the same metrics are shown on the Progress screen and included in the downloadable PDF report.
All metrics are derived from the session transcript. They are coaching signals, not certified acoustic measurements.
Metric reference
Your turns
The total number of times you spoke during the session.
More turns mean more practice opportunities. A healthy benchmark is 8–10 turns in a 5-minute lesson. Low turn counts usually mean the tutor is doing most of the talking — try giving longer answers and asking clarifying questions to keep the conversation balanced.
Turn share
Your turns as a percentage of all turns in the session.
A balanced two-way conversation falls in the 40–60% range. Below 30% typically indicates very short answers that end the exchange quickly. Above 70% can indicate the AI is not getting enough input to generate useful follow-up questions.
Word share
The percentage of all spoken words that were yours.
Word share is a more meaningful indicator than turn share because it accounts for answer length, not just answer frequency. A session where you gave 15 short one-sentence replies will show a high turn share but a low word share. Aim for 40% or above.
Avg words per turn
Your average number of spoken words per answer.
| Level | Target |
|---|---|
| A1 / A2 | 10–20 words |
| B1 / B2 | 25–40 words |
| C1 / C2 | 40+ words |
Longer answers signal deeper engagement and allow the AI to evaluate a wider range of vocabulary and grammar structures.
Speaking pace (WPM)
Approximate words per minute derived from your transcript.
| Pace | Interpretation |
|---|---|
| Below 80 WPM | Sounds hesitant; long gaps between words |
| 80–120 WPM | Measured; acceptable for lower levels |
| 120–180 WPM | Conversational range for native English speakers |
| Above 200 WPM | May reduce clarity; the AI may miss words |
Pace is affected by audio quality and connection speed, so treat it as a trend indicator across multiple sessions rather than a precise measurement.
Filler words
Total count of hesitation markers used during the session: um, uh, er, like, you know, basically, literally, right.
Some filler words are natural in conversational speech. The concern is density, not raw count. Use this number alongside filler density (below) to assess whether fillers are a pattern worth addressing.
Filler density
Filler words per 100 words spoken.
This is the more actionable metric because it adjusts for how much you said. A session where you spoke 500 words and used 10 fillers has a density of 2.0 per 100 — unremarkable. The same 10 fillers in a 100-word session gives a density of 10.0 — a clear signal.
Above 5 fillers per 100 words is worth addressing with deliberate practice (see How to Improve Faster).
Response coverage
The percentage of the tutor's questions that you answered.
High coverage (above 80%) indicates strong attentiveness and engagement. Low coverage can mean you missed questions, gave answers that were too short for the AI to register as a response, or that audio cut out during the session.
Reading the metrics together
No single metric tells the full story. A useful quick read after each session:
- Word share below 40%? → Prioritise longer answers next session.
- Filler density above 5? → Focus on deliberate pauses (see Improve Faster).
- Response coverage below 70%? → Check audio quality and ensure you are waiting for the tutor to finish before answering.
- Pace below 80 WPM? → Warm up with a short spoken passage before your next session.
Note: All metrics are transcript-based coaching signals. Audio quality, background noise, and connection stability all affect accuracy. Results from 3+ consistent sessions are more reliable than any single data point.