Home / Case studies / Sales enablement

[ Case study ]

Every discovery call scored the same way, by something that never gets tired

How we turned a written sales playbook into a scoring system that reads call transcripts, marks them against eighteen criteria, and cites the section of the playbook it is marking against.

Pronto Marketing · Managed website services

36

point rubric, 18 criteria at 0 to 2

3

gaps that showed up in almost every call

11+

client accounts running the same system

10

months in continuous use

Before

Call reviews that happened when a manager had a spare hour, which was never.

The system

A playbook, a rubric derived from it, and an agent that scores transcripts against both.

After

Every call scored and returned as a memo with citations and specific replacement language.

The problem

Sales coaching fails for a boring reason. Reviewing a call properly takes about as long as the call, the person qualified to do it is the person with the least free time, and by the time feedback arrives the rep has run four more calls the same way.

  • Call reviews depended on a manager finding an hour they didn’t have
  • Feedback arrived too late to change the next call
  • Two reviewers scored the same call differently
  • No record that made improvement, or its absence, visible over time
  • A written playbook nobody referenced during or after a call

Deliverable one

The playbook came first, on purpose

You can’t score against a standard that doesn’t exist. Before any scoring was built we wrote the discovery call playbook the scoring would reference: five sections, thirteen numbered subsections, and a time budget for each segment of the call.

Introduction2 to 3 min
Rapport2 to 3 min
Qualification5 to 10 min
Current state10 to 15 min
Next steps5 to 7 min
Time budgetsEach call segment given a target duration, which makes drift measurable.
Rapport and setupDeep discoveryQualificationCall controlValue building
Scoring categoriesEighteen criteria at zero, one or two points, grouped into categories.

Deliverable two

The memo, which is the actual product

1Ingest calltranscript2Score against18 criteria3Cite playbooksections4Write coachinglines5File memo andtable
A scored call memo with names redacted
A real memo, names removedScore out of 36, a written summary of what the call did and didn’t do, then strengths and priority improvements, each citing the playbook section it marks against.

Every memo has the same three parts: a header with the score and a summary, a section of strengths and priority improvements each citing the playbook subsection it draws on, and a set of overall recommendations.

The detail that made reps actually use it is small. Where a criterion scored zero, the system doesn’t just mark it missing. It writes the sentence the rep could have said. “Next time you could say: what would make this a clear win for your team?” is coaching. “Discovery incomplete” is a grade.

What the scores actually showed

This is the part most case studies would leave out. Across the scored calls the numbers didn’t climb. They ran between 22 and 29 out of 36, averaging around 24, with the high point arriving early and the line staying flat afterward.

What the system did deliver was a diagnosis that held up under repetition. The same three gaps appeared call after call: buying authority never confirmed, budget never raised, competing options never asked about. One or two calls showing that is an anecdote. Ten calls showing it is a finding, and it points at the playbook and the compensation design rather than at the rep.

  • Consistency, because the same rubric ran on every call regardless of who was listening
  • Speed, since a memo arrives after the call rather than after the quarter
  • A record, which is what let the flat trend be visible at all
  • Portability, and the same system now runs across more than eleven separate client accounts

A defect worth naming: the rubric is regenerated per call, and one later scoring table came back with nineteen criteria against a thirty six point scale. Pinning the criteria list rather than regenerating it is the fix, and it is the kind of drift that only surfaces when you keep the receipts.

What changed

The system ran on this account for ten months, then outgrew the original engagement. It now runs across more than eleven client accounts, which wasn’t the plan when it was built for one team’s discovery calls.

The honest summary is that scoring calls doesn’t by itself make reps better. It makes what is wrong impossible to argue with, which is the necessary first step and the one most teams skip.

Built on

Google WorkspaceCall transcriptsCustom rubric

If you can’t see the calls, you can’t fix them

A playbook nobody references and reviews nobody has time for is the default state. It doesn’t have to be.