We graded 13,000+ sales calls against the same scoring template, applied the same way across a set of B2B teams. One skill came out weakest more often than any other, and it was the same skill from one team to the next.
This is not a talent problem. The skill they miss sits in the final minutes of the call, the part reps almost never practice.
What sales call skill do reps fail most often?
Across 13,000+ scored sales calls, the weakest skill is closing with clear next steps. It succeeds on just 63% of calls, meaning a score of 3 or more out of 5, the lowest of any skill. Put differently, more than 1 in 3 calls end without a clear next step. The strongest skill, communication, succeeds 96% of the time.
The full ranking is consistent. Reps are strong on the human skills: communication at 96%, rapport at 91%. Discovery is already shakier, at 71%. And they are weakest at the exact point where a meeting turns into a next step, at 63%.
One number makes the gap concrete. Roughly 1 call in 60 earns a perfect 5 on closing. That is 1.7%.
Reps stay sharp for forty minutes, then go vague in the three that count.
How we scored 13,000+ sales calls
Every call was recorded, transcribed, and graded by Ergo against one shared scoring template used by a subset of B2B teams: six comparable skills, each scored 0 to 5, from discovery to objection handling and closing. Managers can override any AI score, and they did on 23% of grades. We used the manager's score wherever one exists, so the ranking reflects human-validated judgment, not raw model output.
This matters for one reason. Every team in that set was graded on the same six skills, on the same scale. When one skill ranks worst across teams selling different products in different markets, the problem is not a single team. It is the role itself.
That is what we found. Closing ranks as the single worst skill on team after team, and in the bottom two almost everywhere. The large majority of teams share the same weakest skill.
Different teams, different markets, one identical blind spot.
Why do reps fail to close the sales call?
The easy answer is discipline. The data, and the job itself, point to two better reasons.
Reason one: timing. Look at the skills reps master: communication (96% success), rapport (91%), objection handling (88%). These run through the whole call, so a rep practices them hundreds of times a week. Closing happens once, in the final minutes, under time pressure, with the next meeting already due. The buyer is wrapping up, the demo ran long, and "I'll send over some times" feels close enough. It is not. On the rubric, and in the pipeline, a vague promise scores a 2 and leaves the deal without a defined next step.
Reason two: weak discovery upfront. This one comes from sales experience rather than the scorecard. A buyer agrees to a next step when they are convinced they have a problem worth solving, and that your product maps to it. That case gets built in discovery, not in the last three minutes. When discovery truly surfaces the pain, proposing another call is easy. When it does not, no closing technique will save it: why would a buyer book a second meeting they do not believe will help them? Notice that discovery is also the second-shakiest skill on the chart, at 71%.
The call doesn't end when time runs out. It ends when the next step exists.
How to close a sales call with clear next steps
Only 1.7% of calls earn a perfect closing score. Here is how they do it: a four-step method you can run on your next call. The goal is simple, agree on a clear next step in the last five minutes, rather than chase one afterward.
Start before the close: earn the next step in discovery. The four steps below only work if the first half of the call did its job. A buyer books a second meeting because call one convinced them the problem is worth solving. Dig into the pain, size it, and tie your product to it, and the close becomes a confirmation instead of a plea.
- Reserve the last five minutes, out loud, at the start. "I want to save the last five minutes to agree on next steps." Announced at minute one, the close stops being a time-squeeze casualty and becomes an agenda item the buyer expects.
- Book the meeting on the call, not after it. Calendar open, invite sent before you hang up. "I'll send times" is where next steps go to die, and it is the single most common pattern in 2-scoring calls.
- Name an owner and a date for every commitment. Not "we'll get you the security doc" but "Sarah sends the security doc Thursday, you route it to IT by Monday." A next step without an owner and a date is a wish.
- Confirm what happens after the next step. One sentence: "If the technical review goes well, what happens next on your side?" That single question separates a 4 from a 5, because it converts one next step into a path.
Run those four on your next call, then score yourself against the rubric. Have every rep do the same, because the data is clear: this is not your weakest rep's problem. It is the weakest minute of the job.
Methodology
We analyzed 13,000+ call evaluations from B2B organizations that share one scoring template, graded between February and July 2026, each skill scored 0 to 5 by Ergo's call scoring, with manager overrides applied on 23% of grades. We always use the effective score: the human override where one exists, the AI score otherwise. The template scores eight fields; we exclude the overall-evaluation summary and call preparation, which cannot be judged from a call transcript, leaving six comparable skills. "Succeeds" means an effective score of 3 or more out of 5.
Cross-team consistency was computed on the teams with enough graded volume to rank, and it holds only within this set: teams using their own custom rubrics are excluded because their criteria are not comparable. Scoring launched in early 2026 so the window covers several months, and the grader is an AI reviewed by humans, not a fully human panel. We also checked whether a weak close predicts a lost deal and found no reliable signal in the linkable subset, so we make no outcome claim here.
