Free — From the book Nobody Trains a Horse Once

The Rubric Starter Kit & The Ninety-Day Checklist

Everything here is free and there is nothing to fill in. No email, no form, no follow-up. The book said take them and don’t tell me about it, and that is exactly what this page is.

Go Straight to the Rubric Template ↓
2
Axes to score
never added together
8–12
Items total
more than twelve and nobody uses it
90 days
From an afternoon’s work
to something real
Free
No email required
no follow-up
01

Before You Download Anything, Run the Two-Rooms Exercise

Take two managers. Put them in separate rooms. Ask them each the same question on a sheet of paper: describe what a good customer conversation sounds like here, specific enough that I could tell whether one happened.

Then put the two sheets side by side.

In every company where I have run this, they have never matched. Usually they don’t even overlap in what kind of thing they’re describing. One writes about warmth and listening. The other writes about steps and outcomes. Both are certain they described the same thing, and both are being honest.

Predicting the result is not the same as holding the two sheets of paper side by side. Do it before you build anything — the gap between those two pages is the reason the rest of this page exists.

What the exercise is actually telling you

Adjectives are not standardsBe consultative. Show empathy. Build rapport. All true, none usable — two people will score the same call differently and neither can be shown wrong.
Feedback is aimed at a moving targetIf your two most senior people can’t independently describe quality, every correction given this year was measured against a private, drifting standard.
Coaching becomes personalityYour people aren’t being coached toward a standard. They’re being coached toward whoever caught them.
The dashboards can’t helpClose rate, handle time, CSAT. All real, all telling you that something is wrong, none telling you what.
02

The Rubric Starter Kit

A rubric is not paperwork and it is not judgment. It is the reference value that turns your observations into information — the thing that tells you the dish is pointed at a wall, when every outcome number can only tell you the shot isn’t coming up.

The one test every item has to pass

Would two people watching the same recording agree it happened? If the honest answer is no, rewrite it or throw it out. There is no third option.

Score two things, and never add them together

A

Conversation Quality

The human half. Tone, listening, who had control, how the customer got handled when he turned difficult, whether the rep sounded like a person or a script.

B

Process Flow

The operational half. Did the required steps happen, in a sensible order, all the way through. Was the information gathered. Was a next step set and confirmed.

Why They Stay Apart

Likeable-and-disorganized and machine-with-no-warmth produce the same total on a single-number system. You’d hand them the same feedback. That is how coaching becomes noise.

If what your people do is a procedure rather than a conversation — a closing routine, a safety check, an intake workflow, an equipment inspection — you use the same machinery with one axis instead of two. There is a right sequence, there is evidence each step happened, and there is nothing conversational to grade.

Build it from your best person, not your conference room

  • Pull ten recordings of your best performer and one of your worst
  • Write down what is different — the specifics, in order, in the words your people actually use
  • Expect to find four or five things your best rep has never told anyone about, including things they don’t know they do
  • That is your rubric. It is already proven in your market, with your customers, and it is provably achievable here because somebody on your payroll is doing it

This is also the moment you find out whether your best performer is actually good or just senior. Sometimes it is the second one. Better to learn that now.

The seven rules for writing it

  • Keep it short. Eight to twelve items across both halves. A rubric that needs a reference document open beside it gets skipped under time pressure — which is precisely when it was supposed to be working.
  • Make every item observable. Write what you would see or hear, not what you would conclude. Not built rapport — used the customer’s name in the first two minutes. Not showed empathy — named the customer’s problem back to him before proposing anything.
  • Anchor every level, middle included. The middle level is the one everybody leaves vague, and it is where most scores land.
  • Weight it steeply, don’t spread it. Even weighting tells your people nothing is a priority, and they will optimize for whatever is easiest to hit. Weight by what would still get you the outcome on its own.
  • Use your people’s words. If the floor calls it the callback, the rubric says the callback. In corporate language it becomes somebody else’s document.
  • Caps on the process side only. Never cap conversation quality.
  • Then stop changing it. Revisit once or twice a year, deliberately, in daylight. A standard that changes quarterly isn’t a standard. It’s a mood.

The acceptance test nobody runs

Two people score the same recording, separately, and compare. Within a couple of points, you have an instrument. Twenty apart, you have more adjectives than you thought — and the item they split on is the one to rewrite.

Skip that step and you will have built something that looks like a standard and behaves like an opinion, and you will not find out for a year.

The blank template — select it, copy it, paste it into a spreadsheet

The columns below are separated by tabs. Highlight the whole block, copy, and paste it into Excel or Google Sheets and it lands in proper columns with nothing to clean up. Then fill in the empty cells. (The block scrolls sideways — the columns run wider than the page. Select all of it anyway; it copies whole.)

Section	Item — what you would SEE or HEAR	Level 1 — did not happen	Level 2 — partial	Level 3 — fully met	Weight	Cap?
Conversation Quality
Conversation Quality
Conversation Quality
Conversation Quality
Process Flow
Process Flow
Process Flow
Process Flow
Process Flow
	Weights must total 100 within each section

A worked example, so you can see the difference between an adjective and an observable

This is one company’s rubric for an inside sales call. It is here to show you the shape. Do not adopt it — it is not your business, and a rubric you did not build from your own best performer is somebody else’s opinion with a point value.

Section	Item — what you would SEE or HEAR	Level 1 — did not happen	Level 2 — partial	Level 3 — fully met	Weight	Cap?
Conversation Quality	Used the customer's name within the first two minutes	Never used it	Used it once, late	Used it in the opening and again later	15
Conversation Quality	Named the customer's problem back to him before proposing anything	Proposed first	Restated part of it	Stated the problem in the customer's own terms, then proposed	30
Conversation Quality	Asked about the customer's situation before mentioning price	Led with price	Asked one question, then price	Asked at least two situation questions before price	30
Conversation Quality	Stayed in the conversation when the customer pushed back	Went quiet or got defensive	Answered but changed the subject	Answered the objection directly and asked a follow-up	25
Process Flow	Verified who he was speaking to and their authority to decide	Not done	Asked name only	Confirmed name and decision role	15
Process Flow	Captured all four required intake fields	0–1 fields	2–3 fields	All four captured	20
Process Flow	Stated the coverage or scope in plain language before quoting	Not stated	Stated partially	Stated fully, unprompted	20
Process Flow	Set a specific next step with a day and a time	No next step	"I'll follow up"	Named a day and a time	30	Cap at 2
Process Flow	Confirmed the next step back to the customer out loud	Not done	Implied it	Repeated it and got a yes	15

How to turn a level into a score

Score each item 1, 2 or 3. Points for an item are weight × (level − 1) ÷ 2 — so level 1 earns nothing, level 2 earns half the weight, level 3 earns all of it. Total each section on its own. Two numbers out of a hundred. Never one.

A cap works differently. If a capped item comes in at level 1, the whole process score is held at whatever ceiling you named, no matter what else went right. Caps belong on the process side only.

The calibration test, in five lines

Item	Scorer 1 (1-3)	Scorer 2 (1-3)	Gap

Items where they disagreed:	___
Largest single-item gap:	___
Verdict:  gap of 0 or 1 = INSTRUMENT   ·   gap of 2 = ADJECTIVES

Any item with a gap of 2 is the item carrying an adjective. Rewrite it as something you would see or hear, then run the test again on a different recording.

03

The Ninety-Day Checklist

It takes an afternoon to start and about three months to be real. One thing to do before week one: decide what you are removing. Don’t start until you can name the thing coming off the calendar, out loud, to the people whose calendar it is.

1

Weeks 1–2 — Write the rubric

Ten recordings of your best, one of your worst. Eight to twelve items. Every level anchored in observable behavior. Weighted steeply. Then the step everybody skips: two people score the same recording independently and compare.

2

Weeks 3–4 — Baseline, and say nothing

Score twenty real calls that have already happened. Don’t coach anyone. Don’t tell anyone their score. Don’t announce the program. You are not calibrating people yet — you are calibrating the instrument.

3

Weeks 5–8 — Start the reps

Attach it to the beginning of the work, not to Thursday at two. One scenario, out loud, under pressure, with somebody playing the customer who does not rescue. One objective. One score. One correction. Then run it again.

4

Weeks 9–12 — Cadence and the scoreboard

Now the only thing that matters is whether it happens again without you pushing. Score real calls on the same instrument — the gap between practice score and live score is the most useful number in your business.

Two things that decide whether this survives

  • Go first. Whoever is leading this takes the objections in front of everybody, in the first session, and lets people get them. You do not have permission to ask a person to be publicly bad at their job until you have been publicly bad at yours. Every roleplay program I have watched die skipped this.
  • Run the rubber-stamp test monthly. When did somebody last score badly? If nobody has been in the bottom category for six weeks, you don’t have a standard, you have a ceremony.

Watch for the five deaths

It becomes a meetingDiscussion replaces reps. Nobody is uncomfortable anymore.
It becomes movableRescheduled once, then twice, then it lives on the calendar and nowhere else.
It gets suspended for the busy seasonWhich is the season it was built for.
The manager stops attendingEverybody reads that correctly and immediately.
It becomes a gestureThe one you will not see coming. It still happens every week, and it has stopped meaning anything.

The whole checklist, ready to copy

Select the block, copy it, and paste it wherever your team actually keeps things — a doc, a wiki page, a note stuck to a wall.

THE NINETY-DAY CHECKLIST
From Nobody Trains a Horse Once — daitalink.com/checklist

BEFORE WEEK ONE
[ ] Name the thing coming off the calendar: ______________________

WEEKS 1–2  WRITE THE RUBRIC
[ ] Pull ten recordings of your best performer, one of your worst
[ ] Write down what is different — specifics, in order, in your people's words
[ ] Two sections: your process, plus conversation quality if you train people to talk
[ ] Eight to twelve items total; more than twelve and nobody uses it
[ ] Every level anchored in observable behavior, middle level included
[ ] Weighted steeply, by what would still get you the outcome on its own
[ ] Caps on the process side only, never on conversation quality
[ ] Two people score the same recording independently, and compare

WEEKS 3–4  BASELINE, AND SAY NOTHING
[ ] Score twenty real calls that have already happened
[ ] Don't coach anyone
[ ] Don't tell anyone their score
[ ] Don't announce the program
[ ] Fix items that don't discriminate, caps too harsh, weights that are wrong
[ ] Write down the range, the average, and the SPREAD between best and worst
    Day-0 spread:  best ______  worst ______  gap ______

WEEKS 5–8  START THE REPS
[ ] Attach it to the beginning of the work, not Thursday at two
[ ] One scenario, run out loud, under pressure
[ ] Somebody plays the customer and DOES NOT RESCUE
[ ] One objective per rep, named as a single short phrase
[ ] One score
[ ] One correction, from a priority order you set in advance
[ ] Then run it again
[ ] GO FIRST — whoever leads this takes the objections in session one

WEEKS 9–12  CADENCE AND THE SCOREBOARD
[ ] Score real calls on the same instrument — same anchors, same weights
[ ] Rubber-stamp test, monthly: when did somebody last score badly?
[ ] Watch for the five deaths:
      [ ] It becomes a meeting
      [ ] It becomes movable
      [ ] It gets suspended for the busy season
      [ ] The manager stops attending
      [ ] It becomes a gesture — the one you will not see coming

DAY 90 — VARIANCE NARROWING, NOT AVERAGES RISING
[ ] The gap between top and bottom has closed
[ ] Two managers score the same call within a few points
[ ] Somebody scored badly last month and nothing bad happened to them
[ ] A new hire got to competence faster than the last one
[ ] Practice happened in a week when things were on fire
[ ] You can name what each person is currently working on
    Day-90 spread: best ______  worst ______  gap ______
    Next review date: ____________
No email.
No download.

Everything is on this page, in plain text, because a file you have to fetch is one more thing between you and starting. Highlight it, copy it, put it where your team already looks. Take it, use it, and don’t tell me about it. If you get partway in and hit the wall, you know where to find me — but that is your move to make, not mine.

04

What Day 90 Looks Like

Here is the measure, and it is not the one you would pick.

Variance narrowing, not averages rising. No customer experiences your average. They experience one conversation, with one person, on one afternoon — and if the spread between your best and your worst is enormous, what you are selling is a lottery.

Your average can climb because one strong person got stronger, and that changes nothing about what your customers experience. So the number to watch is the distance between top and bottom. When that closes, your company has a standard. When it closes and stays closed, you have a system.

Other honest signals at day 90

  • Two of your managers can score the same call within a few points
  • Somebody scored badly last month and nothing bad happened to them
  • A new hire got to competence faster than the last one did
  • Practice happened in a week when things were on fire
  • You can name what each person is currently working on

If you are at day 90 and not yet scoring real calls, you have built a training program rather than a system. That is not a failure — it is most of the way. But the last step is the one that closes the loop, and it is the one everybody stops short of.

Write down today’s spread — your best performer and your worst, on whatever measure you actually trust — and put a date on the calendar to look at it again. You cannot show a gap closing from a baseline you never took, and ninety days from now you will not remember where you started. Nobody ever does.

If You Hit the Wall

Most Companies Stall in the Same Place, and It Isn’t Where They Expect.

They write a good standard. Sometimes an excellent one. Then they discover nobody in the building has the hours to score against it every week, forever. That is the wall, and it is the only reason Daitalink exists.

AI Specialist — What They Know

Allie Builds What They Know

Allie identifies knowledge gaps, coaches the right information, and builds the framework for consistent, accurate responses. This is where most training stops.

AI Specialist — What They Do Under Pressure

Ivan Builds What They Do Under Pressure

Ivan creates real conversation pressure — objections, pushback, the hard moments — and scores the response. He turns what your team knows into what they can actually produce when somebody is pushing back.

The Part Nobody Has the Hours For

The Scoring Keeps Happening After You Stop Thinking About It

A person talking to Ivan is having a real conversation. He pushes back, they have to answer him in the moment with nobody rescuing them, and what they actually said gets scored on your rubric. That happens every week — in the busy season, in the week when everything is on fire, when nobody is watching the program anymore. It is the piece that got dropped in every company we left, and it is the only piece that turns a standard into a system. We run it. You don’t manage it, and you don’t rebuild it when the person who set it up moves on.

The Person Who Wrote All of This

Chuck Leblo Built the System Before He Wrote the Book

Twenty-one years closing this gap by hand, inside other people’s companies, at consulting rates. Before that, a Signal Corps sergeant who once found a van transmitting at ninety percent reflected power with every gauge on it reading normal — which is the whole thesis in one story. He founded Daitalink, works as a fractional Chief AI Officer, and lives in Rockwall, Texas.

You do not have to hire anybody to use this page. All of it is in the book, and the templates above are free. I would honestly rather you built your own — you will understand your business better for having written the rubric with your own hand. This is here for the reader who already knows they are going to hit the wall and would rather skip the year.

Take Them. Use Them. Don’t Tell Me About It.

The templates are free and they always will be. If you get partway in and hit the wall — or if you already know you will and would rather skip the year — that is the conversation.

Schedule a Free Diagnostic Discovery Call → See How Daitalink Works Follow Along on LinkedIn

And if you are not at the wall yet — most people reading this are not — there is nothing here to do but the work. The hardcover is built as a team edition, with an inscription page in the front, for the reader who is handing it to people rather than reading it alone.

Always Improving.

The Rubric Starter Kit and the Ninety-Day Checklist are companion materials to Nobody Trains a Horse Once: Why Your Team Drifts, and the System That Catches It by Chuck Leblo (Daitalink, 2026), available in paperback, hardcover, and Kindle. They are provided free of charge, with no registration and no obligation, and may be used and adapted inside your own organization. Nothing on this page is legal, financial, or compliance advice; regulatory requirements change and vary by jurisdiction, so verify current rules with the relevant authority and your own counsel. Daitalink LLC provides fully managed training and conversation oversight systems. All engagements are custom-built per client and results vary based on the nature of the operation. All product names are trademarks of Daitalink LLC.