The 23-criteria scorecard nobody fills out
Somebody on your team built a call scorecard. It has sections. Greeting, rapport, discovery, needs analysis, product knowledge, objection handling, trial close, next step, tone, pace, compliance. Each section has three or four items. Total: twenty-something checkboxes.
It got used for about two weeks.
Then it turned into a ritual. The manager listens to the call, gives everything a 4 out of 5, drops in one comment, moves on. The score is 86%. Every rep is between 82% and 90%. Nobody has ever been coached off that sheet.
That's not a discipline problem. That's a design problem. A rubric with 23 items is asking a human to hold 23 judgments in their head while also listening to a conversation. It doesn't happen. What happens instead is halo scoring — the rep sounded good, so everything gets a 4.
The number is five to seven
Here's the working rule: a rubric should have five to seven scored criteria, and you should be able to recite them from memory.
Why that range? Three reasons, and none of them are from a study — they're from the mechanics of how a review actually gets done.
One: scoring time. A seven-item rubric on an eight-minute call takes about four minutes to complete if you know the items cold. A twenty-item rubric takes fifteen, because you're re-reading definitions. Multiply by six reps a week and the long rubric quietly becomes optional.
Two: coaching capacity. You can give a rep one or two things to change this week. Maybe three. A rubric that surfaces nine gaps forces the manager to pick anyway — so the extra criteria added work without adding decisions.
Three: signal dilution. When you average 20 items, a real failure gets buried. A rep who never sets a next step but is warm and knowledgeable scores 85%. A rep who always sets the next step but is a little flat scores 85%. Same number, completely different coaching.
The test: would you change behavior over it?
Before any criterion stays on the sheet, answer this: if a rep scored a 1 on this, would you sit them down?
"Used the customer's name twice" — no. You wouldn't. Cut it.
"Confirmed the vehicle is for them or a family member" — yes, because that changes the whole conversation. Keep it.
Run that test on your existing scorecard. Most sheets lose half their lines in ten minutes.
What usually survives the cut
Different businesses, same bones. A home services call and a B2B inside sales call share more than people think. Here's a version of the five-to-seven that holds up:
- Reason for the call captured in the customer's words. Not a category — their words. "My wife's SUV lease is up in forty days and she hates the third row."
- Timeline and trigger. What's forcing a decision and when.
- Decision structure. Who else is involved, what has to be true for them to say yes.
- Budget or payment reality surfaced. Not a quote — a range that got said out loud.
- A specific, scheduled next step. Day, time, who calls who.
- Objection or hesitation addressed, not deflected. The rep repeated the concern back before answering it.
- Accurate information given. No guessing on price, availability, or lead time.
You could run that on an insurance renewal call, a roofing estimate call, and a software demo booking call. The wording changes; the structure doesn't.
Notice what's missing: rapport, tone, enthusiasm, product knowledge as its own line. Those aren't unimportant — they're inputs. They show up inside the items above. A rep with no rapport doesn't get the real timeline. A rep with no product knowledge gives inaccurate information. Score the outcome, not the vibe.
When more than seven is actually right
Two exceptions.
Compliance. If you're in insurance or lending, there are things that must be said and things that must never be said. Those are not rubric criteria. They're a separate pass/fail checklist. Don't average a disclosure failure into a 100-point score — that's how you end up with "92%, but he didn't give the disclosure." Keep it binary and keep it on its own list.
Role-specific add-ons. A BDC agent booking appointments and a closer running a demo shouldn't share the same seven items. Build one core set of four or five that everybody gets scored on, then add two role-specific items. The BDC gets "offered two specific appointment windows." The closer gets "confirmed the economic buyer was on the call." Core stays stable, edges flex.
Weighting beats adding
The urge to add a criterion usually comes from a real frustration. "Reps aren't confirming the budget" — so somebody adds a line. Then another. The sheet grows.
Better move: weight what you already have.
If next step is the thing that's killing you this quarter, make it 30% of the score instead of one-seventh. A rep who nails everything else and ends with "I'll send you some info" should not score well. Right now, on most sheets, they do.
Weighting also gives you a way to shift emphasis without retraining everyone on a new form. Same seven items, new weights, announced in a fifteen-minute meeting:
"Same scorecard. One change. Next step is now worth 30 points instead of 14. If the call ends without a day and a time, you cannot score above a 70 no matter how good the rest was. That's it. That's the whole change."
How to cut your current rubric this week
Thirty minutes, one sitting.
- Print the rubric. Pull five recent calls — two good, two bad, one you argued about.
- Score all five with the full rubric. Note which items gave you the same score every single time. An item that's always a 4 or always a 5 isn't measuring anything. Cut it.
- For each remaining item, ask the behavior-change test. If a 1 wouldn't trigger a conversation, cut it.
- Whatever's left, if it's more than seven, merge. "Asked about timeline" and "asked about urgency" are one item.
- Write the final list on an index card. If you can't carry it in your head after reading it twice, it's still too long.
Then score the same five calls again with the short version. The scores should spread out — your bad calls should drop into the 50s and 60s. If everything still lands between 80 and 90, your criteria are too easy to pass, not too many.
The real tell
A good rubric produces disagreement.
If you and your sales manager score the same call and land within two points every time, either you're both great or the rubric isn't making anyone decide anything. Usually it's the second one. Narrow criteria with clear thresholds — "a next step with a specific day and time, or no credit" — create arguments. Arguments are where the coaching standard actually gets built.
Twenty-three checkboxes never produce an argument. They produce an 86%.
If you're running these reviews across more than a handful of calls a week, consistent scoring gets hard to hold in your head — that's the point where tooling like MoreSignal earns its keep, applying the same seven items to every call instead of whichever three you remembered that day. But the list comes first. Cut it to seven, weight it, and use it on five calls before you automate anything.