Sales call scoring rates each recorded call against a scorecard, a short list of steps a good call includes. Done well, it is a coaching tool: it shows each person what to work on next, not a grade to defend.
The hard part is not the idea. It is writing a scorecard people trust, and scoring enough calls that the results mean something. This guide covers both, with a discovery template you can copy, templates for four other call types and a one-month plan to roll it out.
What call scoring is and what it is for
A scorecard turns "that call felt off" into specific findings. Each step is a question about the call, and each answer is a result: met, partial or miss.
The goal is coaching, not grading. A score on its own tells a seller nothing they can act on. A score that says "you asked about budget but never asked who signs off" does.
That changes how you build and use the scorecard:
- Steps describe behavior you can hear. "Asked who signs off on the purchase" is scorable. "Was confident" is not.
- Results point to a moment. The seller should be able to find the part of the call that earned the result.
- Scores feed 1:1s. If nobody discusses them, scoring is a reporting exercise.
Call scores also help managers beyond coaching. Over weeks, they show which steps the whole team skips, which is a training problem and not a people problem.
Manual review vs AI scoring
Most teams start with manual review. A manager picks a few recordings, listens, and fills in a form. It works, and it is the right way to learn what a good scorecard looks like. The limit is time.
A 30-minute call takes about 30 minutes to review, more if you pause to take notes. A manager with a full calendar can review only a few calls a week. If each person on a team of ten has several calls a week, the manager sees a small slice of what happens.
| Manual review | AI scoring | |
|---|---|---|
| Coverage | A few calls a week, usually picked by the manager or the seller | Every recorded call |
| Consistency | Varies by manager, mood and how much time they had | Same steps and same wording on every call |
| Time per call | About as long as the call itself | None for the manager until they open the result |
| Feedback speed | Days later, if the review happens at all | Soon after the call ends, while the seller remembers it |
| Patterns | Hard to see from a handful of calls | Trends per person and per step |
| Best use | Calibrating the scorecard, judging nuance | Covering every call and finding where to look |
The point is not that AI is better than a manager at judging a call. It is that a manager cannot be on every call, and a pattern needs many calls to show up.
A sample also has a bias. Managers tend to review calls that are already known to be good or bad: a big deal, a lost deal, a new hire. The routine calls in the middle, where most habits form, go unseen.
If you want to know how much of your call volume gets scored today, the call scoring calculator takes four inputs and gives you a coverage number.
How to build a call scorecard
You can build a scorecard in a spreadsheet in an afternoon. Follow these five steps.
1. Pick one call type
A discovery call, a demo and a renewal call need different steps. One scorecard that tries to cover all of them ends up vague. Start with the call type that matters most to your pipeline, usually discovery, and build one scorecard for it.
2. Choose 5 to 8 steps
Every step you add makes the scorecard harder to use and the results harder to read. Five to eight steps is enough to cover a call and short enough that a manager can name the one that matters.
If you already train on a methodology, use it as the spine. MEDDICC, SPIN, GPCT and Challenger each give you a ready list of things to cover. If you use SPIN or GPCT, their guides list the questions behind each step.
3. Write each step as a yes/no question
"Discovery quality" cannot be scored. "Did they ask who signs off on the purchase?" can. A yes/no question forces you to name a behavior, and it gives the seller a clear target.
Begin each step with "Did they..." and keep it to one behavior. If a step has an "and" in it, split it or drop one half.
4. Say what "met" and "partial" look like
This is the step most teams skip, and it is why two managers score the same call differently. Write a line for each step:
- Met: the behavior happened fully. "The seller asked who signs off and got a name or a title."
- Partial: it happened, but not enough. "The seller asked about approval in general but did not get a name."
- Miss: it did not happen.
Short examples are better than adjectives. "Got a name or a title" is clear. "Explored the buyer thoroughly" is not.
5. Weight what matters
Not every step is equal. Skipping the next step on a call costs more than skipping a rapport question. Mark the few steps that decide the call as more important, and let the lighter ones count for less.
Use three levels, such as Critical, Important and Encouraged, or a simple 3, 2, 1 weight.
Call scoring criteria: good steps vs bad steps
Most scorecards fail at the wording, not the idea. A step like "built rapport" sounds clear until two managers score the same call and one says met while the other says miss.
The fix is the same every time. Name the behavior, ask it as a yes/no question, and write down what met sounds like. The met line is what turns a list of questions into a call scoring rubric.
Here are common vague steps and how to rewrite them.
| Vague step | Why it fails | Rewritten as a yes/no question | What met looks like |
|---|---|---|---|
| Built rapport | Everyone hears rapport differently, and it rewards small talk. | Did they connect the opening to something specific about the buyer? | In the first two minutes, the seller mentions a fact about the buyer's company or role, such as a new hire or a launch. |
| Discovery quality | It covers ten behaviors at once, so nobody knows what to fix. | Did they ask what the problem costs today? | The buyer names a cost in time, money or risk: "We lose about two days a month to this." |
| Asked good questions | Counting questions rewards volume, not depth. | Did they ask a follow-up on the buyer's first answer? | The seller asks "What happens then?" or "Can you give me an example?" before moving to the next topic. |
| Qualified the deal | Qualification is several steps, so the score hides which part was missed. | Did they find out who signs off on the purchase? | A name or a title, confirmed before the seller moves on. |
| Handled objections well | "Well" is an opinion, and it rewards a fast answer over a good one. | Did they ask about the objection before answering it? | The seller asks what is behind the concern ("What makes you say that?") before giving a response. |
| Built value | Value is a result, not a behavior you can hear. | Did they tie what they showed to a problem the buyer named? | The seller uses the buyer's own words: "You said onboarding takes three weeks. Here is how this changes that." |
| Good talk time | A ratio says nothing about what was said, and you cannot count it by ear. | Did the buyer describe the problem in their own words before the seller pitched? | The buyer talks through the problem, with at least one follow-up from the seller, before any product comes up. |
| Was confident and professional | It scores personality, which nobody can practice. | Did they set an agenda and check it with the buyer? | The seller says what the call will cover and asks, "Does that work, or is there anything you want to add?" |
| Strong close | It means one thing on a discovery call and another on a pricing call. | Did they agree a dated next step before the call ended? | A meeting with a date, a purpose and named attendees, confirmed out loud. |
Two tests catch most bad steps. First, could a new manager listen to the call and score the step without asking you what you meant? Second, could the seller practice it before their next call? If either answer is no, rewrite the step.
Template: Discovery - MEDDICC scorecard
Here is a scorecard you can copy. It follows MEDDICC, with one added step for the close of the call. Each step is a question, with a line on what "met" looks like.
| Section | Step | What met looks like |
|---|---|---|
| Metrics | Did they quantify the result the buyer wants? | The buyer states a number: a target, a current baseline or a cost, in their own words. |
| Economic buyer | Did they find out who signs off on the purchase? | A name or a title, and how that person has approved similar spend before. |
| Decision criteria | Did they learn what the buyer will judge options on? | At least two things the buyer will compare, such as price, integrations or time to launch. |
| Decision process | Did they map how the decision gets made? | The buyer names the steps, the people and a rough timeline from here to a signature. |
| Identify pain | Did they uncover a problem and what it costs? | A specific problem, who it affects and what happens if nothing changes. |
| Champion | Did they find someone who will push for the deal internally? | A named person with a reason to want a change, and a sign they will help, such as an intro. |
| Competition | Did they ask what else the buyer is considering? | The buyer names alternatives, including doing nothing or an existing tool. |
| Next step set | Did they agree a dated next step before the call ended? | A meeting on the calendar with a date, a purpose and the right people invited. |
Eight steps is the top of the range. If your team finds it long, merge Decision criteria and Decision process into one step on how the buyer will decide, or drop Competition for a segment where it rarely comes up.
For weighting, treat Economic buyer, Identify pain and Next step set as Critical, since a call without them rarely moves forward. Make Metrics, Decision process and Champion Important, and Decision criteria and Competition Encouraged.
Here is what a scored discovery call looks like: each step gets a result and the reason behind it.
Notice that each result comes with a reason tied to something said on the call. That is what makes a score usable in a 1:1. The seller sees what was missing, and the manager does not have to argue the number.
Call scoring templates for other call types
Discovery is where most teams start, but the same method works for every call you record. Each template below has five to seven steps. Copy the one you need, then change the wording to match how your team sells.
Cold call scorecard
A cold call is short, so the steps cover the first 30 seconds and the ask. Score it on whether the seller earned the next conversation, not on how much they pitched.
| Section | Step | What met looks like |
|---|---|---|
| Opener | Did they say who they are and why they are calling in the first two sentences? | Name, company and a reason, with no "How are you today?" in front of it. |
| Permission | Did they ask for a moment before going on? | A question such as "Can I take 30 seconds to say why I called, and you tell me if it's worth a chat?" |
| Reason for the call | Did they give a reason specific to this buyer? | A trigger the buyer recognizes, such as a new role, a hiring push or a problem common in their industry. |
| Problem question | Did they ask one open question about a problem the buyer may have? | The buyer answers with detail, not a yes or no: "How are you handling new customer onboarding today?" |
| Brush-off | Did they respond to the first brush-off with a question? | On "Send me an email" or "We're all set", the seller asks one short question before accepting it. |
| Booked meeting | Did they ask for a specific meeting time? | The seller offers a date and time. Met counts the ask, even when the buyer says no. |
Treat Reason for the call and Booked meeting as Critical. A friendly cold call that ends without an ask is still a miss.
Demo scorecard
A demo goes wrong when it shows everything. These steps keep it tied to what the buyer said in discovery.
| Section | Step | What met looks like |
|---|---|---|
| Recap | Did they recap the buyer's problem before showing anything? | The seller restates the problem from discovery, and the buyer confirms it is still right. |
| Tailored to stated pains | Did they show the parts that match the buyer's problems? | Each section links back to a pain in the buyer's words. Features nobody asked about stay short or are skipped. |
| Reaction check | Did they check for a reaction after each main section? | A question such as "How does that compare to how you do it today?" after each section, not only at the end. |
| Questions | Did they clarify a question before answering it? | The seller asks what is behind the question, answers it, and checks the answer landed. |
| Next step with stakeholders | Did they agree a next step that includes the people who were not on the call? | A dated meeting with the missing stakeholders, such as the economic buyer or IT. |
Pricing and negotiation scorecard
On a pricing call, the risk is giving away margin to end an awkward silence. Score whether the seller kept the conversation on value and on the decision.
| Section | Step | What met looks like |
|---|---|---|
| Value first | Did they restate the value before naming a price? | The seller recaps the problem and what it costs the buyer, in the buyer's words, before the number. |
| Price stated | Did they state the price clearly and then stop talking? | The number is said once, without apology, and the seller waits for the buyer to respond. |
| Price objection | Did they find out what the price is being compared to? | The seller asks "Compared to what?" or similar, and learns whether it is budget, a competitor or doing nothing. |
| Decision process | Did they confirm how the decision gets made from here? | Who signs, what review or paperwork comes next, and a target date. |
| Trade, not discount | Did they ask for something in return for any concession? | Any change to price comes with a change in terms, such as a longer contract or an earlier signature. |
| Next step | Did they agree a clear next step with a date? | A date for the decision or the next call, and who will be on it. |
Follow-up call scorecard
A second call should move the deal forward, not repeat the first one. Use the discovery scorecard from the first call to decide what this one needs to fix.
| Section | Step | What met looks like |
|---|---|---|
| Recap | Did they recap the last call in the first few minutes? | A short summary of the problem, the agreed next step and any open questions, confirmed by the buyer. |
| What changed | Did they ask what changed since the last call? | The seller asks about new people, new dates or new priorities, and gets an answer. |
| Commitments | Did they check the actions agreed last time? | The seller asks about each action by name, such as "Did you get a chance to share the proposal with Lisa?" |
| One MEDDICC gap | Did they close one gap from the first call? | The seller picks one step that was a partial or a miss, such as Economic buyer, and gets it to met. |
| Next step | Did they agree a dated next step that moves the deal forward? | A meeting with a date and a purpose that is more than "check in". |
SPIN, GPCT and Challenger scorecards
If your team sells with a methodology, build the scorecard from its parts.
- SPIN: one step each for situation, problem, implication and need-payoff questions. See SPIN selling explained for the questions behind each step.
- GPCT: one step each for the buyer's goals, plans, challenges and timeline. See how to use the GPCT sales methodology.
- Challenger: steps for teaching a new insight, tailoring it to the buyer and taking control of the next step. See the Challenger sale.
Turning scores into coaching
Scores only matter if they change what happens in the next call. Here is a simple routine for a manager.
Look at trends, not single calls
One bad call tells you little. A person who has missed the same step on six of their last eight calls has a habit. Look at each person's results over a few weeks, and look at each step across the team.
Two views are useful:
- Per person: which steps does this person hit, and which do they skip? This tells you what to coach in their next 1:1.
- Per step: which steps does the whole team miss? This tells you what to fix in training, not in individual feedback.
Pick one step to coach
A seller can work on one thing at a time. If they miss four steps, pick the one with the most weight, or the one that will make the next one easier. Economic buyer, for example, usually comes before Decision process.
Say it plainly: "This month we are working on asking who signs off. Everything else is fine for now."
Use the call moment in the 1:1
Do not describe the miss in general terms. Open the call at the place where it happened, play 30 seconds, and ask the seller what they would say differently.
The same works for a seller who answers objections before asking a question: play the moment, then work through how to handle that objection together.
Hearing yourself say "so, who else is involved?" and then move on is more useful than any feedback summary. Write down the timestamp when you score each step, so you can find the moment in seconds.
Practice it with a colleague
Feedback without practice rarely changes behavior. After the 1:1, have the seller practice the step before their next real call. Ask a colleague to play a buyer who dodges the question, "I'd have to check on that", and try the question three or four times.
Then check the next few calls. If the step moves from miss to partial to met, you coached the right thing. If not, the step wording or your coaching needs to change.
Score calls yourself this week
You do not need software to start. Copy the template into a spreadsheet with one row per call and one column per step. Mark each cell met, partial or miss.
- Pick five recorded calls from the last two weeks, a mix of people and outcomes.
- Score each one against the template.
- Ask a second manager to score two of the same calls without seeing your marks. Compare, and rewrite any step where you disagree.
Here is how three calls might look:
| Call | Metrics | Economic buyer | Decision criteria | Decision process | Identify pain | Champion | Competition | Next step set |
|---|---|---|---|---|---|---|---|---|
| Corvane, Maya | Met | Miss | Partial | Miss | Met | Partial | Miss | Met |
| Halcyon Labs, Dan | Partial | Miss | Met | Partial | Met | Miss | Met | Miss |
| Northwind Freight, Priya | Met | Miss | Met | Miss | Partial | Miss | Partial | Met |
Read down the columns first. Here Economic buyer is missed on all three calls, so that is the step to coach this week.
How to run a calibration session
Calibration is how a scorecard becomes fair. If two managers read a step differently, the same call gets two different scores, and the team notices fast. A calibration session fixes the wording, not the people.
It is the second-manager check from the list above, turned into a routine. Invite everyone who scores calls: managers, team leads and whoever runs enablement.
A 45-minute agenda
The 45 minutes assume people score the call before the session. For a short call, such as a cold call, score it together in the first 10 minutes instead.
- Pick one call (before the session). Choose a recent, routine call of the type the scorecard covers. Avoid a famous win or loss, since everyone already has an opinion on it. Share the recording and the scorecard a day ahead.
- Everyone scores alone (before the session). Each person marks met, partial or miss on every step, with one line of evidence: a timestamp or a quote. Nobody compares notes until everyone is done.
- Compare step by step (5 minutes). Put all the scores side by side in one sheet, one column per person. Mark the steps where everyone agrees and skip them.
- Discuss only the disagreements (20 minutes). For each one, each person shares their evidence. Then ask one question: do we disagree about what happened, or about what the step means? If it is what happened, replay the moment. If it is what the step means, the wording is the problem.
- Rewrite the step (10 minutes). Change the question or the met line until everyone would have scored it the same way. Add an example from this call to the met line, since examples settle more arguments than adjectives.
- Re-score with the new wording (10 minutes). Score the disputed steps again. If people still disagree, flag the step and bring a new call to the next session.
Here is what step 3 might look like for three managers on one discovery call:
| Step | Ana | Tom | Lee |
|---|---|---|---|
| Metrics | Met | Met | Met |
| Economic buyer | Partial | Miss | Partial |
| Decision process | Miss | Miss | Miss |
| Champion | Met | Miss | Partial |
| Next step set | Met | Met | Met |
Metrics, Decision process and Next step set need no discussion. Economic buyer is close. Champion is the one to spend time on: a met next to a miss means the three of them read the step differently.
How often to calibrate
Calibrate every week for the first month, while the wording is new. After that, once a month is enough. Add a session whenever you change a step, add a new scorecard or bring in a new manager.
Keep a short log of every rewrite and the reason for it. New managers read it before they score their first call.
What agreement looks like
You do not need perfect agreement. Aim for the same result on most steps, and no step where one person says met and another says miss.
A met next to a partial is a small gap, and it often comes down to judgment. A met next to a miss means the step is unclear. Rewrite it before anyone scores with it again.
Rolling it out without people gaming it
How you introduce scoring decides whether people trust it. If the team thinks scores are a performance review in disguise, they will steer calls toward the scorecard, or stop recording the hard ones.
A few habits keep scoring about coaching:
- Say what it is for. Scores pick what to coach next. They are not for pay, rankings or performance reviews.
- Share the scorecard before the first score. Nobody should learn what a step means from a miss. Send the steps, the met lines and an example call a week ahead, and invite edits. People who helped write a step argue less with the result.
- Score managers' own calls too. If you sell or join calls, put your calls through the same scorecard. Share your results first, including a miss.
- Keep scores out of pay and rankings at the start. No leaderboard and no link to commission. Once a number decides money, people work on the number instead of the call. Wait until the steps are stable and the team trusts the wording before you discuss any other use.
- Celebrate a step moving from miss to met. Say it in the team meeting: "Dan set a dated next step on every call this week." That does more for the team's trust in scoring than any rule.
Watch for box-ticking. If someone asks "Who signs off?" in the last minute of every call and moves on, the step is met but the point is missed. Rewrite the met line to include what the seller did with the answer.
A message to send your team
Here is a short message a manager can send a week before the first score. Change the call type and the dates to fit.
Subject: Call scoring starts next Monday
Hi team,
From next Monday, I will score a few of our discovery calls each week against a short scorecard. It is attached, with a line on what "met" looks like for each step.
This is for coaching. Scores won't be used for pay, rankings or reviews. In each 1:1, we will use them to pick one thing to work on.
I'm scoring my own calls too, and I will share mine first.
If a step is unclear or doesn't fit how you sell, tell me by Friday and we will change it.
Thanks,
Ana
Reading the results
After a few weeks you will have a grid: one row per person, one column per step. This is the view the coaching routine above starts from.
Here is a team of four after two weeks of discovery calls. Each cell shows how many of that person's calls met the step.
| Person | Calls | Metrics | Economic buyer | Identify pain | Champion | Next step set |
|---|---|---|---|---|---|---|
| Maya | 8 | 6 of 8 | 2 of 8 | 7 of 8 | 4 of 8 | 7 of 8 |
| Dan | 7 | 3 of 7 | 1 of 7 | 6 of 7 | 2 of 7 | 3 of 7 |
| Priya | 9 | 7 of 9 | 3 of 9 | 8 of 9 | 6 of 9 | 8 of 9 |
| Sam | 4 | 3 of 4 | 0 of 4 | 3 of 4 | 2 of 4 | 4 of 4 |
Team-wide gaps vs individual gaps
Read the columns first. Economic buyer is low for everyone, so it is a team-wide gap. That points to training or to the step itself, not to one person. Fix it once for the whole team: a short session on asking who signs off, one shared question to use, and a practice round in pairs.
Then read the rows. Dan's Next step set is low while the rest of the team is strong. That is an individual gap, and it goes into Dan's next 1:1. Dan is also low on Metrics, but Next step set is Critical, so coach that first.
Look for strengths too. Priya meets Champion more often than anyone else. Use one of her calls as the example in the team session.
The grid above counts met only. Keep the partial and miss counts in your sheet as well. A step that is mostly partial means the person knows to ask but stops early. A step that is mostly miss means they do not ask at all, and it needs different coaching.
How many calls you need
One call is an anecdote. Wait for at least 5 to 8 calls of the same type per person before you treat a row as a pattern. Below that, one unusual call, such as a buyer who ran the meeting or a call cut short, can swing a step from strong to weak.
Sam's row has four calls, so treat it as a hint and keep scoring. Do not mix call types in one row either. A discovery step scored on a demo will look like a miss when it was never the goal of the call.
Score trends over a month
Compare the same grid week by week. What matters is direction: is the step you coached moving from miss to partial to met?
Track the coached step for each person on its own row, week 1 to week 4. If it moves, keep going. If it stays flat for three or four weeks, change the coaching or the step wording. A team-wide step that improves after a training session tells you the session worked.
Common mistakes
Most scorecards fail in the same few ways.
Too many steps. A scorecard with 20 steps produces 20 findings per call and no priority. Cut it to eight or fewer. If you want more detail, build a second scorecard for a different call type.
Scoring without coaching. If scores go into a dashboard and nobody discusses them, sellers learn to ignore them. Tie every score to a conversation. One step, one 1:1, one practice session.
No calibration between managers. If two managers read "partial" differently, scores mean different things by team. Run the calibration session above, weekly for the first month and then monthly. Wherever managers differ, rewrite the step until they agree.
Scoring only bad calls. If you only score calls after a deal is lost, scoring becomes a post-mortem. Sellers also never see what good looks like. Score the routine calls and the wins too, so the scorecard shows what works.
Treating the score as a ranking. A single number across people invites arguments about the number. Use scores to decide what to coach, and let the step-level results lead the conversation.
A one-month plan
Here is how the pieces above fit into four weeks, for one call type and one team.
| Week | What to do | Done when |
|---|---|---|
| Week 1: build and test the scorecard | Pick one call type, write 5 to 8 steps with met lines, and check them against the good and bad steps table. Score five calls yourself. At the end of the week, send the team message. | You have a draft scorecard that has been tested on real calls. |
| Week 2: score and calibrate | Score two or three calls per person. Run your first calibration session with every manager who scores. | Each step has been tested by more than one person and rewritten where you disagreed. |
| Week 3: first coaching 1:1s | Pick one step per person. Play the call moment in the 1:1 and set up a practice round with a colleague. Run a second calibration session. | Every person knows the one step they are working on this month. |
| Week 4: review patterns and adjust the steps | Read the person and step grid for team-wide and individual gaps. Rewrite or drop any step nobody can score consistently. | You know the team-wide gap, each person's next step, and you have version two of the scorecard. |
By the end of the month, most people will have just enough calls to start reading their row. Keep the grid going into month two before you draw firm conclusions. Once the first scorecard works, copy the method to your next call type.
Where to start
Pick one call type and one scorecard. Write five to eight yes/no steps, say what met and partial look like, and weight the few that matter most. Check each step against the good and bad steps table, then score a handful of calls by hand to test the wording with a second manager.
Next, follow the one-month plan. Tell the team what scoring is for before the first score, calibrate until the steps read the same to every manager, and coach one step per person. When discovery works, use the templates above for cold calls, demos, pricing calls and follow-ups.
Then decide how you will cover the rest of the calls. If a few reviews a week is all your team can do today, see what share of your calls get scored and what it would take to cover the rest.