To assess sales competencies, mark a few defined skills on the rep's real calls against written levels, check that two markers agree, and judge the result by how well it ranks reps. Self-rated questionnaires measure confidence.
Most sales competency assessments are surveys. The rep rates themselves, sometimes the manager rates them too, and the result is a radar chart that says the team is good at rapport and weak at negotiation. It tells you what people believe about their skills. A call shows what they do.
What assessing a sales competency involves
Assessing a competency means deciding, from evidence, what level a rep is at on a defined skill. The evidence that matters most is what the rep does on calls with buyers.
Enablement usually owns the assessment design: the competencies, their levels and how calls get marked. First-line managers do most of the marking, because they will coach what it finds. Setting it up for one role takes two to three weeks. After that, marking takes a manager about fifteen to twenty minutes per call, or much less with an AI scorer doing the first pass.
The assessment has two uses. One is a baseline: where each rep is today, so coaching starts in the right place. The other is change: whether a skill moved after coaching or training. The second use is the one most programs never get to, because the first assessment was a survey and there is nothing to compare it with.
Assessing on calls, step by step
The six steps below take a team from a competency list to an assessment that a manager can repeat each quarter.
- 01BriefAgree the skill and its levels
- 02ObserveMark it on real calls
- 03DebriefShow the moment and the level
- 04Observe againCheck it moved on later calls
- 01BriefAgree the skill and its levels
- 02ObserveMark it on real calls
- 03DebriefShow the moment and the level
- 04Observe againCheck it moved on later calls
1. Pick four to six competencies for the role
Start from the role's sales competency model if you have one. Pick the four to six competencies that most affect results for that role, such as discovery, quantifying the problem, reaching the economic buyer and agreeing next steps. An assessment of twenty competencies at once takes too long to mark and gives the rep too many things to work on.
2. Write levels a manager can mark
For each competency, write three to five levels. Each level should describe something a person could see or hear on a call. "Asks good questions" cannot be marked. "Asks how the problem shows up in the buyer's own numbers, and gets a figure" can.
3. Add marker reasons to each level
Next to each level, add a short reason from a real call: why that call sat at that level and not the one above. In our call-scoring research, the trained assessors' own reasons for their marks, which they called marker reasons, gave the biggest single gain in how closely AI scores matched human ones. Textbook descriptions of good behavior added nothing (field guide). Marker reasons help human markers in the same way: they show where the line between levels sits.
4. Mark real calls
Pick calls where the competency should show up. Discovery is assessed on first meetings, negotiation on late-stage calls. Use at least two or three calls per rep per competency, because one call can be a bad day. Mark each competency on its own, with the moment in the call that justifies the level.
5. Check that two markers agree
Have a second person mark some of the same calls and compare. Disagreement is normal: in our research, two trained experts marking the same calls disagreed on 18% of scores. Where markers disagree on the same call, the level description is usually the cause. Rewrite it and mark again.
6. Judge the assessment by ranking
The test of an assessment is whether it puts reps and calls in the right order. A scorer can look accurate on average while telling you nothing: in our research, a scorer that never read the transcript landed an average error of 0.41, against 0.43 for the best real model. Ask instead whether the assessment ranks calls the way your best managers would. If it does, it is good enough to coach from.
Assessing in the team's own method
The levels should use the words the team already sells with. A team that runs MEDDPICC should assess "identifies the economic buyer" and "tests the champion" in those terms, because those are the words managers coach in and reps hear in deal reviews. A generic competency list on top of a method creates two languages, and reps learn to perform for the assessment.
The same goes for the calls. Assess on calls from the team's own market, with its own buyers and its own products, so the evidence matches the job.
New hires are the exception to the two or three call rule. In the first weeks they have few real calls, so assess them on practice conversations at first, then move to live calls as soon as they have them. Our guide on the same scorecard for practice and live calls covers why the levels should not change between the two.
A worked example
This example is illustrative. Arrowmere, a 600-person logistics firm, wants to assess discovery across its eight account executives before a coaching program starts.
Competencies. Enablement picks four: uncovering the problem, quantifying it, finding the decision process, and agreeing a next step.
Levels. For quantifying the problem, level 1 is "accepts the buyer's description with no number"; level 3 is "asks for a figure and gets an estimate"; level 5 is "gets a figure, ties it to a cost the buyer's leadership tracks, and confirms it back". Each level carries one marker reason from a real Arrowmere call.
Marking. Managers mark three first meetings per rep. Enablement marks a third of the same calls. On quantifying the problem, the two disagree on four calls; all four sit between levels 2 and 3. The level 2 description is rewritten to say the buyer offered a number without being asked, and the second round of marking agrees on all but one.
Result. Six of eight reps sit at level 2 or below on quantifying the problem, so that becomes the first coaching focus. Three months later the same managers mark three new calls per rep using the same levels. Four reps moved up a level. Two did not, and their managers change how they coach it.
Where it goes wrong
Using a survey as the assessment. Self-rated skills measure confidence. Use surveys, if at all, to ask reps where they want help.
Too many competencies. Twenty competencies marked on one call produce twenty weak judgments. Four to six marked well produce a coaching plan.
Levels nobody can mark. Adjectives like "strong" and "effective" mean something different to each manager. Rewrite any level two managers mark differently on the same call.
Assessing once. A baseline with no second assessment cannot show change. Put the repeat assessment in the calendar before the first one starts.
Assessment without coaching. The assessment only matters if a manager acts on it. Across 1,246 pharmaceutical sales reps, managers' coaching skill predicted goal attainment (Dahling et al. 2016), and what happens back at work is more strongly tied to whether a skill gets used than what happens in training (Saks and Belcourt 2006). Both are summarized in our Game Winning Kickoffs field study. Our sales coaching framework covers the brief, observe and debrief loop that turns an assessment into practice.
Assessing every rep, every quarter
The hard part of call-based assessment is the marking time. Three calls per rep, four competencies each, for a team of eight is a lot of manager hours each quarter, and most teams stop after the first round.
AI scoring can take the first pass if it is built and tested carefully. Our research program ran more than 5,000 tests and more than 250,000 scoring decisions across 15 AI models on real enterprise discovery roleplays. The final method ranked calls close to the order trained experts did, and only 1.7% of scores changed when the same calls were scored again (the checklist model alone). The method is in how to build a call scoring rubric. On Replicate Labs, reps then work on the gap with Caty, one of our AI sales experts, by voice or video on a call, in the team's own method. For enablers shows how teams set it up, or talk to us.
FAQ
How do you assess sales competencies?
Assess sales competencies by marking a few defined skills on the rep's real calls against written levels. Use two or three calls per rep, check that two markers agree, and repeat each quarter.
What is a sales competency assessment?
A sales competency assessment is a structured judgment of a rep's level on defined sales skills. The most useful versions are based on real calls with buyers, marked against levels a manager can see.
How many competencies should you assess at once?
Assess four to six competencies per role at a time. More than that takes too long to mark and gives each rep too many things to work on.
Can AI assess sales competencies?
AI can assess sales competencies on calls if the scorecard is written carefully and the scorer is tested on how well it ranks calls. In our research, the final method ranked calls close to the order trained experts did.
How often should you assess sales competencies?
Assess sales competencies once a quarter on the same levels, so each assessment can show whether a skill moved. A single baseline with no repeat cannot show change.