By Jay ShahQuick answer:
- Sales rep certification is a set of gates a new rep passes before talking to buyers alone: a knowledge check, a recorded pitch, a scored call roleplay, an objection roleplay and one supervised live call.
- Score every gate on the same 0/1/2 rubric. Pass at 80 or higher, call 90 or higher Merit, and fail anyone with a 0 on a must-pass line, whatever the total.
- Calibrate graders first, re-score real calls at day 30, and recertify when something changes, not only once a year.
Joelle has her first live discovery call at 10 tomorrow. She finished the product videos, aced the quiz and sat in on four calls. Her manager asked if she felt ready and she said yes, because what else do you say in week three? Nobody has heard her answer “we already have a tool for that.”
Below is how to certify sales reps without an enablement team: gates, a sample scorecard, pass bars, grader calibration, a fail path, recertification triggers and a one-page checklist.
What is sales rep certification, and why bother?
Sales rep certification is a short series of scored tests a rep must pass, in order, before taking a given kind of call alone. It replaces “do you feel ready?” with a written standard and a number.
Forrester’s 2018 research on certifying the sales force has three levels: content mastery, simulated application and field execution. Many teams stop at the quiz, and McKinsey’s 2019 article on sales training describes the result: “People who complete the training get a piece of paper.”
Certify experienced hires too. The Bridge Group’s 2026 AE report found companies now want 3.7 years of experience at AE hire, up from 2.7 in 2022, yet average AE ramp hit 6.2 months, the highest in the study’s history. At the report’s $200K median OTE, that’s roughly $103K of on-target earnings spent before an AE is fully productive (our math, not theirs). For the wider onboarding plan, see our guide to AI sales onboarding.
What should you certify? The five gates
Certify the moments a new rep will face in their first month of calls, one gate at a time, each with a pass bar and a named grader. Keep SDR and AE versions separate.
| Gate | SDR version | AE version | Pass bar |
|---|---|---|---|
| 1. Knowledge check | Product, ICP, personas, process | Same, plus pricing | 90%+, retake after review |
| 2. Talk track | Recorded 30-second opener | Recorded 3-minute first-call pitch | 80+, no false claims |
| 3. Core call roleplay | Cold call: opener, one pain question, book the meeting | Discovery call (scorecard below) | 80+, no must-pass 0 |
| 4. Objections | Top 5 objections, random order, unscripted | Same | 2 on at least 4 of 5, no 0s |
| 5. Live-call gate | Shadow 3 calls, then 1 supervised live call | Same | 80+ on the gate 3 sheet |
The hiring manager grades gates 2 to 5. Keep gate 1 friendly: Matt Doyon, quoted by the Sales Enablement Collective, says the knowledge check should be “a check, not a trap.” Run gates 3 and 4 as sales role play with a manager or peer as the buyer.
GitLab publishes a real version in its handbook. Its Revenue Academy ends in a stand-and-deliver capstone, and AEs keep a shadow log reviewed at week 3 before a manager-graded recorded pitch. Our SDR ramp plan shows where the gates fit in a 30-day schedule.
How do you score it? A 0/1/2 rubric and pass bars
Score each behavior 0 (didn’t happen), 1 (partly or late) or 2 (clearly and on time), turn the total into a percentage and pass at 80. Mark two lines must-pass, so a high total can’t hide the step that loses deals.
The bands come from GitLab’s Revenue Academy: Merit at 90 or higher, Pass at 80 to 89, Needs Development below 80. They’re the only public company pass bars we found.
Here’s a sample scorecard for gate 3, the AE discovery roleplay. It uses the same scale as our sales call scorecard template, so certification and live-call scores line up, but each line spells out what 0, 1 and 2 sound like so two graders hear the same thing.
| Behavior | Must-pass | 0 | 1 | 2 |
|---|---|---|---|---|
| Set an agenda | No | Jumps into questions | Mentions it, doesn’t check | States agenda and time, asks “does that work?” |
| Asked about the current situation | No | Pitches first | Closed yes/no questions | Open questions plus follow-ups like “how often?” |
| Found the pain in the buyer’s words | Yes | No problem named | Rep names it, buyer half agrees | Rep plays it back and the buyer confirms |
| Quantified the impact | No | No number | “A lot of time” | Hours, money or risk tied to the pain |
| Learned the decision process | No | Never asks | Asks who decides, not how | Knows who’s involved and what happens next |
| Handled the surprise objection | No | Argues or drops it | Answers before asking why | Acknowledges, asks a question, then answers |
| Booked a next step with a date | Yes | “I’ll send some info” | Agrees to talk, no date | Meeting on the calendar before hanging up |
Seven lines give 14 points, which makes the bands clean: 13 or 14 is Merit, 12 is Pass, 11 or less is Needs Development. A 0 on either must-pass line fails the gate even at 12.
What happens when a rep fails a gate?
The rep doesn’t move on, practices only the missed lines and retests the same gate within a few days.
- Same day: share the scored sheet and the moment that cost points (“at minute 6 the buyer said ‘it’s slow’ and you moved on”).
- Next 2 to 3 days: at least three practice reps on each missed line.
- Retest: same gate, different buyer persona.
- Second fail: a panel retest. GitLab’s virtual re-assessments use a panel of at least the hiring manager and an Enablement Lead. Write down any exception and why.
No “pass with notes” before live calls. That’s what the must-pass lines are for.
How do you calibrate graders before certifying anyone?
Have every grader score the same recordings alone, compare line by line, and rewrite anchors until you agree on at least 80% of lines.
This is frame-of-reference training, which a 2012 meta-analysis by Roch and colleagues found improves rating accuracy. The 80% target comes from McHugh’s 2012 paper on interrater reliability, which notes many texts treat 80% agreement as the minimum. For a stricter test, it calls any kappa below 0.60 inadequate.
- Pick 3 recordings: one strong, one weak, one in between.
- Score alone.
- Count exact matches. Seven lines times three recordings is 21 scores; you need 17 identical.
- Fix the splits. Rewrite the anchors on any line you split on in more observable words.
- Repeat with new recordings until you reach 80%.
- Keep checking. Double-score every fifth certification and recalibrate if agreement slips.
With one manager, the second grader can be your strongest rep or a founder.
How do the live-call gate and day-30 re-score work?
A roleplay pass shows a rep can do it in practice; the live-call gate checks they do it with a real buyer. Score one supervised live call on the gate 3 sheet, then re-score 3 to 5 live calls around day 30.
A meta-analysis of behavior modeling training (Taylor, Russ-Eft and Chan, 2005) found large effects on knowledge and skills but modest effects on job behavior, with better transfer when supervisors were trained too. Jonathan Mahan of The Practice Lab, in the same Sales Enablement Collective piece, says passing should require a live-call recording.
At day 30, average the rep’s last 3 to 5 calls. At 80 or higher, they’re fully certified. Below that, the weakest line goes into your weekly loop; our guide on how to coach sales reps covers the routine.
When should reps recertify?
Recertify when something changes, with a 6-month floor for small teams. On a yearly cycle alone, reps can pitch old pricing for months.
| Trigger | What to recertify |
|---|---|
| New product, feature or pricing | Gates 1 and 2 for everyone, within 2 weeks |
| A scorecard line averages below 1 on the team’s live calls for a month | That gate, whole team |
| A rep’s live-call average stays under 80 for 3 weeks | That rep’s weakest gate |
| SDR moves to AE | All AE gates |
| Rubric or methodology changes | Recalibrate graders, then the affected gates |
| 6 months since last certification | Gates 3 and 4 |
In McKinsey’s case study, reps had to “recertify every year.” Gartner’s 2026 research on sticky sales training says sellers start forgetting shortly after training and recommends “verifying application of learned behaviors.” Short, spaced re-tests do that better than one big exam: in Roediger and Karpicke’s 2006 study, students who were repeatedly tested recalled 61% after a week, versus 40% for those who kept restudying.
Copy this one-page sales certification checklist
Paste this into a doc or sheet, one copy per rep. The hiring manager grades every step except the quiz.
| Done | Step | Pass bar | Due | Score and date |
|---|---|---|---|---|
| ☐ | Graders calibrated on 3 recordings | 80%+ agreement | Before day 1 | |
| ☐ | Gate 1: knowledge check | 90%+ | End of week 1 | |
| ☐ | Gate 2: recorded talk track | 80+, no false claims | Week 2 | |
| ☐ | Gate 3: cold call or discovery roleplay | 80+, no must-pass 0 | End of week 2 | |
| ☐ | Gate 4: top 5 objections | 2 on 4 of 5, no 0s | End of week 2 | |
| ☐ | Shadow log: 3 calls with notes | Reviewed | Week 3 | |
| ☐ | Gate 5: supervised live call | 80+ on gate 3 sheet | Week 3 | |
| ☐ | Day-30 re-score: 3 to 5 live calls | Average 80+ | Day 30 | |
| ☐ | Recertification date set | 6 months or next trigger | Day 30 |
How TrackPoint runs the gates
TrackPoint is where we’d run gates 3 to 5 and the recertification loop. You build each roleplay from your own talk track, playbook or a past call in under 30 minutes, then send it as an assignment with a due date and a passing score that matches your rubric. The AI buyer keeps details like budget or the real objection back until the rep asks the right question (we call it Semantic Reasoning), so a memorized script doesn’t pass.
A team view shows each rep’s readiness per competency, and you can export a CSV of who passed. When a recertification trigger fires, reassign the gate.
For the live-call gate and day-30 re-score, a notetaker joins Zoom, Google Meet or Microsoft Teams calls and scores each one on the same scorecard, so you don’t have to listen to every call. Calls already in Gong import and get scored the same way. Weak moments on real calls become assigned practice roleplays.
Frequently asked questions
What is a sales certification program?
It’s a set of scored gates a rep must pass before taking a type of call alone. A good one tests knowledge, practice and a real call against one rubric with a written pass bar.
How long should sales onboarding certification take?
About three weeks for the go-live gates: knowledge in week 1, roleplays by the end of week 2 and a supervised live call in week 3, then a re-score at day 30. Full ramp runs longer; the Bridge Group’s 2025 SDR report puts average SDR ramp at 3.0 months.
What should be in a sales readiness checklist?
Grader calibration, a knowledge check, a recorded pitch, a core call roleplay, an objection roleplay, a shadow log, a supervised live call, a day-30 re-score and a recertification date. Give each line a pass bar, a grader and a due date.
What is a good passing score for sales certification?
80 out of 100, with 90 or higher as Merit, matching the bands in GitLab’s public handbook. Use 90% for knowledge checks, and fail any gate with a 0 on a must-pass line.
How often should sales reps recertify?
At least every 6 to 12 months, and sooner when something changes: new pricing or products, a team-wide drop on a scorecard line, a rep’s live-call average under 80, or a move from SDR to AE.
Who should grade sales certification roleplays?
The hiring manager, with a second grader double-scoring a sample. Calibrate both on the same recordings first and aim for at least 80% exact agreement per line.
Run the discovery gate with your next hire
Start with gate 3 and the sample scorecard, and calibrate with one other person before your next rep starts. Add the other gates once you trust your own scores.
Talk to our team to see certification run on your own talk tracks and scorecard, or try a free roleplay and see how a practice call gets scored line by line.
