Before we shipped Closerbell, we tested it the only way that mattered: scoring my own calls. 30 in a row. The dimensions revealed four patterns I'd been blind to. Here's the log.
Building a sales-call scoring tool means dogfooding it on your own calls. So that's what I did.
Over six weeks, I scored every B2B sales call I ran — discovery, demo, pricing, follow-up, the lot. Pasted the transcript into Closerbell. Read the ticket. Wrote down the dimensions that flagged.
The first 10 calls averaged a closer-score of 7.5 / 14. Mid-range. The kind of number that suggests "you're not bad" — and tells you nothing about how to be better.
What changed was reading the same dimension flag four times in a row.
I'd assumed my weak dimensions were random — bad call here, good call there. They weren't. Across 30 calls, four dimensions kept flagging. Same four. Different calls, same four.
Talk-listen ratio across my first 10 calls averaged 62:38 — me at 62. The rubric flags anything over 55:45 on a discovery call. I'd been telling myself I was a "consultative" rep. The dimension said I was a monologue.
The decision-maker-confirmation dimension flagged on 8 of my first 10 calls. I'd been pitching to whoever booked the meeting and assuming they'd "bring it back to the team." 6 of those 8 calls died at the "let me check with the team" stage. I never asked who the team was.
"I'll send a follow-up email" / "Let's circle back next week" / "I'll send some materials over." All flagged. The rubric only credits next-step-clarity if the rep names a specific date, specific people, and a specific deliverable. I was hitting one of three on average.
This was the embarrassing one. The rubric tracks whether the buyer described their problem in their own words during the call — a signal that they actually have the problem they think they have, vs. having it framed for them. My buyers were articulating their pain on 40% of calls. The 60% where they didn't were the calls I lost.
Pattern-spotting is one thing. Drilling against it is another. What I actually changed:
I started timing my discovery answers with a stopwatch. Anything over 90 seconds, I cut the next time around. I also added a deliberate 3-second pause after every prospect answer before responding. Felt unnatural for a week. Now it's automatic — and the prospect almost always fills the silence with more useful information than my next question would have surfaced.
One question, asked in the first 10 minutes of every call: "Who else weighs in on a decision like this?" Sometimes it's just them. Sometimes it's a budget-holder. Sometimes it's a committee. Either way, you know before you pitch.
The biggest change. Last 3 minutes of every call: I open the calendar and offer a specific Tuesday 2pm. If they can't, I offer a specific Wednesday 10am. We pick a slot. I send the invite from my screen, on the call, before we hang up. Closer-score jumped 1.5 points the week I started doing this.
When a prospect mentions a problem, I started asking "What does that look like on a Tuesday?" Forces them to describe the day-to-day texture of the problem. If they can't, the problem isn't acute enough to buy. If they can, you have your case study, your ROI calculator, and your follow-up email written for you.
4 points in 30 calls. ~6 weeks. No coaching, no manager review, no expensive enterprise platform. Just the rubric, applied consistently, with the discipline to actually drill the dimensions that flagged.
Three deals closed in the last month that I'd have lost in February. Not because I got better at "selling" — I got better at four specific things the rubric kept telling me to fix.
Some of you will read this and think "of course, that's basic sales." Talk less. Confirm DM. Lock the next step. Get the buyer to articulate pain. None of it is novel.
The point isn't that the advice is new. The point is that I'd read all of it before — in books, in newsletters, in manager 1:1s — and I still wasn't doing it. Reading doesn't change behavior. Pattern-spotting in your own data does.
Closerbell didn't make me a better salesperson. It showed me which dimensions of "salesperson" I was actually weak in, and gave me the same flag enough times that I had to change.
If you've been reading sales advice for a year and your numbers haven't moved, that's the gap. Run your last 10 calls through the engine. Read the dimensions that flag. Find your four.
Drop your last 10 transcripts. The pattern shows up by call 5. Free tier is 1 score per 24h.
Score a call →— Logged across 30 of my own scored calls, Mar–May 2026.