Scorecards and criteria
The scorecard is the one thing in Qualitify that is entirely yours. It decides what a reviewer is asked, what a score means, and therefore what your team optimises for — which makes it worth more thought than any other setting.
Where it is
QualityWhat counts as a good call
Owners only. Reviewers score against the scorecard and cannot change it, which is what stops the standard drifting quietly as different people apply it.
A new company starts with a reasonable guess about a sales call. It exists so the review queue works on day one, not because it is right for you — replace it.
What a criterion is made of
- The question
- What the reviewer is asked about the call. Written in your team’s own words — a reviewer should not have to interpret it.
- The weight
- How much of the total score this criterion is worth. Weights are relative, so what matters is their proportion to each other.
- Critical failure
- Marks a criterion as one that fails the call outright when it is failed, whatever else went well. Use it for one or two things, never for half the scorecard.
Writing criteria that survive contact with reality
Score what can be heard
The test for every criterion: could two reviewers, listening to the same recording separately, agree? If not, the criterion will produce noise that looks like performance data.
- "Confirmed the delivery address before ending the call" — audible, binary, arguable by nobody.
- "Asked what the customer was using it for" — either they did or they did not.
- "Explained the price including delivery" — checkable against the recording.
- "Agreed a specific next step and a date" — present or absent.
- "Was professional" — two reviewers will score the same call differently all week.
- "Showed empathy" — real, but not scorable from a recording with any consistency.
- "Had good product knowledge" — unmeasurable unless the customer happened to test it.
- "Built rapport" — means something different to everybody who reads it.
Rewrite the vague ones into behaviours
Most unscorable criteria are a real concern described at the wrong level. “Showed empathy” usually means “acknowledged the problem before offering a solution”, which is a thing you can hear. The point is not to drop what you care about but to name the behaviour that demonstrates it.
Keep it short
Four to six criteria. Reviewing properly takes real attention, and a twelve-criterion scorecard gets skimmed — twelve half-considered answers are worth less than five considered ones. If you cannot bear to cut one, that is a sign the scorecard is doing two jobs and you want a second one for a different call type.
Weight by consequence
Even weights are a decision not to decide. If getting the next step agreed is what actually moves a deal, it should be worth more than the greeting. Weighting honestly is also what makes the Where quality is won and lost chart on the overview useful, because it then reflects what you actually care about.
Critical failures
Some mistakes should fail a call no matter how well the rest of it went: promising something you cannot deliver, quoting a price that does not exist, being rude to a customer. Marking a criterion as a critical failure means exactly that.
| Ordinary criterion | Critical failure | |
|---|---|---|
| Effect when failed | Loses its weight from the score | Fails the whole call |
| Use for | The things that make a call better or worse | The one or two that make a call unacceptable |
| How many | Four to six in total | One or two, at most |
Do not mark everything critical
Every criterion being critical means every imperfect call fails, and a scorecard where nobody can pass stops being used. It also destroys your ability to see improvement: if the score is binary, there is no trend to read.
How a score is worked out
A reviewer answers each criterion. The score is the weighted result of those answers, expressed out of ten, with critical failures overriding it.
- Each criterion contributes in proportion to its weight.
- A criterion marked not applicable to that call is left out rather than counted as failed, so the score stays comparable.
- A critical failure fails the call regardless of the rest.
- A call nobody has reviewed has no score. It is not a zero, and it is never averaged as one.
Blank is not zero
Worth knowing because it is the most common misreading of any quality figure. An unreviewed call is absent from the averages entirely. This is why review counts appear next to scores throughout the product — the average is only ever a statement about the calls somebody actually listened to.
Changing the scorecard later
You will, and that is fine — but changing it changes what scores mean, so it is worth doing deliberately rather than continuously.
- Existing reviews
- Keep the scores they were given. A review is a record of a judgement made at a time, and silently recalculating history would make your trends fiction.
- The trend across a change
- Has a seam in it. When you reweight significantly, note the date — a step in the average that week is the scorecard moving, not the team.
- How often
- Settle it in the first month, then leave it alone for a quarter at a time. A scorecard that changes monthly produces no comparable data at all.
What to do in the first month
- Write four criteria, not ten. Review ten calls yourself against them — see Scoring a call.
- Rewrite the ones you found yourself hesitating over. Hesitation means it is not scorable as written.
- Then run a calibration: two reviewers, one call, scored separately. Where they disagree is where your wording is loose. See Coaching and calibration.
- Only then add weights you intend to keep, and leave it alone for a while.