Speaker Hub uses necessary cookies for secure accounts. With your permission, first party analytics help us improve signup and onboarding. We do not use advertising or cross site tracking cookies. Read the Cookie Policy.
A Speech Analyzer score means nothing by itself. Compare is where it starts to mean something. · Speaker Hub
A Speech Analyzer score means nothing by itself. Compare is where it starts to mean something.
Analyze produces one number for one talk. Compare and Talk Library are the other two-thirds of the tool, and they are what turn that number into evidence of whether you actually got better. Here is what each part does and how to use them.
•6 min read
Run a talk through Analyze and you get one number. Maybe it is a good number. Maybe it is not. Either way, a single score cannot tell you if you are improving, because improving is a comparison, and a comparison needs a second talk.
That is not a gap in the tool. It is the other two thirds of it.
Analyze is one view of three, not the whole tool
Speech Analyzer is three connected views that share the same shell and the same navy accent: Analyze, Compare, and Talk Library. They are not three separate features bolted together. They are one segmented switch, because the job only finishes once a talk has somewhere to sit and something to sit next to.
Analyze is where a talk goes in. Talk Library is where it lives afterward. Compare is where two of them get read against each other. Most of the value sits in the second and third view, and most speakers only ever open the first.
Every talk you run already lives in the library
You do not save anything to get a talk into Talk Library. Run it through Analyze once and it is there, searchable by title, twenty to a page. There is no separate archive step, no export, no extra click.
The search box only matches the title you gave the talk, not the transcript underneath it. If you named every upload "Keynote" or left the default filename in place, search will not help you tell them apart later. A few extra words in the title when you upload, the client name, the date, the version, pay off the next time you are hunting for one talk among thirty.
Compare is not two scores side by side. It is three specific numbers
Pick Talk A and Talk B and Compare shows three stat cards: overall engagement out of 100, cognitive load out of 100, and mind wandering out of 100. The first two run the way you would expect. Higher is better.
Mind wandering runs the other way. A lower number there means the audience stayed with you longer, not less. It is the one card on the page built to be read backward, and it is easy to glance at three numbers going in different directions and assume something broke. Nothing broke. Check which one it is before you decide a talk regressed.
None of the three is a transcript of what the room felt. They are estimates built from your recording and transcript, the same way the single Analyze score is. What Compare adds is not more accuracy. It is a second data point, which is the only thing that turns an estimate into a trend.
The comparison lives in the link, not just the screen
Pick two talks and the pair does not just sit in the page's memory. It goes into the URL itself, as the two talk ids. That is deliberate: the timeline chart and the written report are their own pages now, and each one rebuilds the same comparison from that link instead of losing the pairing the moment you navigate away.
Practically, that means the link you are looking at after picking Talk A and Talk B is a real address for that specific pairing. Copy it into a message to a coach, a co-presenter, or your own notes, and opening it again puts the same two talks back in place.
Three ways to actually use this
Build the library before you need it. A talk that went fine is not a talk worth skipping. It is the baseline the next one gets measured against.
Compare your most recent talk to your best one, not your worst one. Beating your worst talk tells you almost nothing. Closing the gap to your best tells you where the ceiling actually is.
Read mind wandering last, and read it backward. Check overall engagement and cognitive load first, since those move the way you expect, then look at mind wandering knowing the direction is flipped.
A single score is a snapshot. Whether you are getting better is not something one number can answer, no matter how good that number looks. That question only has an answer once there are two talks to put next to each other, which is what the other two thirds of Speech Analyzer are for.
Do I need to manually save a talk before I can compare it?
No. Every talk you run through Analyze lands in Talk Library on its own. There is no separate save step, and nothing extra to do before it shows up as an option in Compare.
My newer talk got a lower mind wandering number. Is that good or bad?
Good. Mind wandering is the one number on Compare that runs backward: a lower score means the audience stayed with you longer. Overall engagement and cognitive load both work the normal way, where higher is better, so mind wandering is the one to read in reverse.
Can I send a comparison to a coach or a co-presenter?
Yes. The two talks you picked live in the page's own link as a pair of ids, not just in the screen state. Copy that link after picking Talk A and Talk B and send it on. Opening it lines up the same two talks again.
The Speech Analyzer pairs your recording with a transcript and an estimated response timeline, and none of that is a measurement of how the room actually felt. Here is what the number is built from, what to check before you act on it, and why Compare matters more than any single score.
Two psychologists spent years studying um and uh and concluded they are words, not mistakes. Here is what to cut instead, and the one filler that actually costs you.
