CODED × KJO · Day 2

From Data
to Insight

Day two of three. Eight questions in one fixed order, asked of two files from two industries.

Six hours · twelve blocks

Arabella Hotel

Bring your change log

Press → to begin

Yesterday, in three things

You finished holding clean data.

Two dashboards from one file, where the only difference was the prompt.

That board, where fifteen prompts produced four different answers to one question.

The thing in front of you now: a cleaned Work Order Set, with a change log beside it.

That is the bottom rung, and you earned it.

Nothing else on this slide

Industrial Branch handled 257 service requests this year.

Is that good or bad?

One number, out of a clean file, with nothing beside it.

Name the felt problem

A number on its own is not information.

Yesterday you spent an afternoon making a file you can trust. That was real work and it was worth it. This morning I gave you one number out of a clean file, and not one of you could tell me whether it was good news.

Everything you just asked for is the same move: put something beside it.

What you’ll walk out able to do

By the end of today, here’s what changes.

Run the eight steps, in order, on any table from any business.

Add it up honestly: two definitions, the row count, and who changes rank between them.

Put three neighbors beside a number, and say which one changes what somebody does.

Carry two worked examples: two rank flips, one caused by a hole, one by a definition.

The shape of the gap

Four rungs. We climb two today.

A label for the two days, not a theory. Name it, hang the two days on it, move.

Where the week sits on it

Today is the middle two.

Information. Getting a number you can actually defend, out of 2,400 rows rather than 40.

Knowledge. Putting something beside it.

Which rung was yesterday’s answer board on?

Topic 03 · Inside the workbook

An answer in a chat is a screenshot.

1 A chat answer lives in the chat.

2 You retype it into the file you actually send.

3 And a retyped number cannot be traced back.

A column in your sheet is work. You can read the formula, and argue with it.

Live · format as a table, add a column, group by site, then ask what it excluded

Ctrl + T first. Then ask.

A range of cells is not a table, and that is where it fails. Format it first, then put the tool to work on the table itself.

Add a column that flags work orders where the priority is Emergency or Urgent.

Two routes to the same answer

Track A writes into your sheet. Track B answers beside it.

Track A: Copilot in Excel

On the table itself. It writes a column back into your file, with a formula you can read. It needs the data formatted as a table, and it needs a license.

Track B: Copilot Chat, file attached

On a copy it read. Same answers, no license needed. It cannot write into your workbook, so you copy across by hand. That is where transcription errors enter.

Some of you have a button the rest do not. That is a license decision made above this room, not a skill gap. Both routes reach every answer today.

Your turn · 2,419 rows

Work the full Work Order Set.

File: work-orders-full.xlsx

1 Format it as a table. Ctrl + T, first, before anything else.

2 Ask two things. Add a column flagging unplanned work, and tell me the row count and what you excluded.

3 Write both down. Page 2.

Block 03 · your turn

Ctrl + T. Then ask.

The column first. Then the row count and what it excluded.

Workbook · page 2

Format the file as a table before anything else.
Add the column that flags unplanned work.
Write the row count, and what it excluded, on page 2.

12:00

Timing flexes to the room. The clock is a guide, not a gate.

Break

Ten minutes.

10:00

When you come back, one question asked of your file before you open anything.

Need a different length? The ☕ button below runs a custom timer.

Ladder A · file: work-orders-full.xlsx

Eight steps. Always this order.

1DATAHow big is it?

2DATAWhat is missing, and what is impossible?

3INFORMATIONAdd it up.

4INFORMATIONAdd it up honestly.

5KNOWLEDGEBeside the others.

6KNOWLEDGEBeside itself, earlier.

7KNOWLEDGEBeside an outside mark.

8KNOWLEDGESo what?

We type them together. Nobody gets left on a step.

STEP 1DATA

How big is it?

Nothing yet. This is the question you ask before you are allowed an opinion.

How many rows and columns are in this table?
List every column name and what type of value it holds.
What is the earliest and latest date in the file?

STEP 2DATA

What is missing?

Where this file is thin. A file is only as good as its emptiest useful column.

For every column in this table, tell me how many rows have a value
and how many are blank. Show it as a count and a percentage.
Sort it so the emptiest column is first.

CloseDate: 2,079 of 2,419. 340 work orders have no close date.

STEP 3INFORMATION

Add it up.

Up one rung. Information is what a number adds up to. Watch how little it takes.

What is the average time to close a work order, for each site?
Time to close means the number of days between RaisedDate and CloseDate.
Show it as a table, fastest site first.

Step 3 · the result

It reads clean.

Site	Average days to close
Ahmadi Refinery	3.17
Burgan Gathering Center	3.30
Sabriyah Station	3.38
Mina Al-Zour Terminal	3.42
Wafra Station	3.42
North Field Booster	3.70

Every one of those is arithmetically correct.

STEP 4INFORMATION

Add it up honestly.

Same question. Same file. Same tool. Three extra lines.

What is the average time to close a work order, for each site?
Time to close means the number of days between RaisedDate and CloseDate.

Before you answer:
- tell me how many work orders have no CloseDate
- show that count by site, as a percentage of that site’s rows

Then give me two tables:
Table A: the average using only work orders that have a CloseDate.
Table B: the average counting work orders with no CloseDate as still
open, measured from RaisedDate to 30 June 2025, the latest raised
date in the file.

Tell me which sites change rank between the two tables.

Step 4 · the result · the same question, asked properly

Ahmadi was fastest.

Table A uses closed work orders only. Table B counts the open ones.

Name it

A rumor with a decimal point.

1 A number is Information when you know how many rows it used.

2 Ask for the row count on every question, not the hard ones.

3 You cannot tell in advance which one is the hard one.

This one looked easy.

Break

Ten minutes.

10:00

Then we take that number and give it a neighbor.

Need a different length? The ☕ button below runs a custom timer.

Mode change

This block we write the prompts together.

1 Last block I typed and you copied.

2 Now I tell you what each prompt has to do.

3 You tell me how to say it.

Same number. 49.4 days. Three different neighbors.

Callback · the list from ten past nine

You already knew what was missing.

This morning I gave you one number and asked whether it was good. Here is what you asked me for. You did not need me for that.

What you have not had is a name for it, and a habit around it.

The concept

Four things you can put beside a number.

Beside	The question it answers	Step
The others in the same file	Who is out of line with the rest?	5
Itself, earlier	Is this getting better or worse?	6
A target or outside benchmark	Are we where we said we would be?	7
Nothing, and say so	“I have no baseline yet”	anytime

The fourth is always available and almost nobody uses it.

STEP 5KNOWLEDGE

Beside the others.

What does this prompt have to do? Say it in your own words first. Then we type the version the room agrees on.

For each site, show:
- its average days to close
- the average across all six sites
- the difference, in days and as a percentage

Sort worst first.

From 49.4 days to twenty days worse than anyone else.

STEP 6KNOWLEDGE

Beside itself, earlier.

The most dangerous of the four. The room is about to be handed a triumph.

Split the work orders into two periods by RaisedDate:
May to December 2024, and January to June 2025.
Show average days to close for each site in each period,
and the change between them.

The triumph, inspected

Every site improved seventy percent.

1 Nothing improved.

2 A work order raised in June has had weeks to close.

3 One raised last May has had fourteen months.

A period comparison is only fair if both periods had the same chance to finish.

STEP 7KNOWLEDGE

Beside an outside mark.

The honest one. The last line of this prompt is the whole step.

Compare each site’s average days to close, counting open work
orders, against a 30-day target. Show days over or under.
Say plainly that the 30-day figure was supplied by me and is not
derived from this file.

Five sites inside a common mark. One is nineteen days outside it.

STEP 8KNOWLEDGE

Three findings. All true. One number.

Beside	The finding
The others	Ahmadi is twenty days worse than any other site
Itself, earlier	Everything improved seventy percent, and none of it is real
A 30-day mark	Five sites inside it, Ahmadi nineteen days outside

No prompt for this one. It is not the tool’s job. Which finding would change what somebody does on Monday?

Break

Ten minutes.

10:00

Then the same eight steps on a file none of you has opened.

Need a different length? The ☕ button below runs a custom timer.

Ladder B · file: finance-ledger.xlsx · requested by this room

Same eight steps. You drive.

1DATAHow big is it?

2DATAWhat is missing, and what is impossible?

3INFORMATIONAdd it up.

4INFORMATIONAdd it up honestly.

5KNOWLEDGEBeside the others.

6KNOWLEDGEBeside itself, earlier.

7KNOWLEDGEBeside an outside mark.

8KNOWLEDGESo what?

I call the step number. Then I stop talking.

STEP 2DATA

Every column is full. Is it clean?

Steps one and two, on your own. Then wait: somebody will say the file is clean.

Are there any values in this table that should not be possible?
Check for negative numbers in every numeric column and count them.

147 negatives, and every column full. Missing is not the only kind of broken.

STEP 4INFORMATION

Add it up honestly. Again.

The one step with help, because nobody here has met a credit note in a total yet.

Total spend by cost center. Before you answer, tell me how many rows
have a negative amount and what they appear to be.

Then give me two tables:
Table A: the sum of positive amounts only.
Table B: the sum of every amount, so credit notes reduce the total.

Tell me which cost centers change rank between them.

Step 4 · the result · second file, second rank flip

Different cause.

Table A sums positive amounts only. Table B lets credit notes reduce the total.

The trap

Somebody will suggest making them all positive.

The fix	Total	What actually happened
Leave it alone	10,778,660	Nine typing errors still in there
ABS() everything	11,732,772	Overstates by 911,540. Refunds became spend
Delete every negative	11,255,716	Overstates by 434,484. Returns vanish
Fix the nine, keep the rest	10,821,232	Correct

A fix that touches rows you never looked at is a second error, applied at scale.

Step 6, driven by the room

Facilities cut costs by a quarter.

Cost center	Jul to Dec 24	Jan to Jun 25	Change
Facilities	1,246,794	947,028	-24%
IT	951,217	781,970	-18%
Operations	1,189,053	1,029,764	-13%
Maintenance	795,090	906,284	+14%
Logistics	505,783	626,862	+24%
TOTAL	5,599,270	5,179,390	-7%

Take out one credit note and it is ten percent, not twenty-four.

STEP 8KNOWLEDGE

Two files. Two industries. Two rank flips.

1 One was a hole in a column.

2 One was a definition.

3 The same eight steps found both.

You needed to ask how big, what is missing, what does it add up to, and compared with what.

The room · yesterday’s page 8, on the table

Hands up who wrote delete it.

Both answers exist in the room, and that is the point.

The reveal

0KWD

WO-2087 Ahmadi Refinery · 72 hours · a three-day compressor failure

The most expensive single event in fourteen months of data. Everybody who deleted it produced a cleaner chart and removed the most important row in the file.

Its twin in the finance file: TX-9946, a 172,400 KWD credit note doing half of a cost-cutting story on its own.

For any outlier

Three questions.

OUTLIER OR SIGNAL The data cannot tell you which one you are holding. The business can.

1 Is it a data error, or a real event?

2 If it is real, is it a one-off or the first of a pattern?

3 Who in the business would know?

Integration · onto the comparison card

Three sentences turn a finding into something usable.

	The sentence	The test it has to pass
1	What is true. The number, its comparison, its row count	Could somebody reproduce it from your file?
2	What it means. In business words	Does it say anything a manager did not already know?
3	What would break it. The assumption that kills sentence 2	Would you say it in front of the site manager?

Sentence three is the one everybody skips. It is the one that gets you believed.

Close · two rungs today

What you climbed.

Data. Yesterday. A file you can defend, and a record of why.

Information. This morning. Steps 1 to 4, and step 4 was where Ahmadi went from fastest to slowest.

Knowledge. Three neighbors for one number, and the same eight steps on a file you had never opened.

Wisdom. Tomorrow. Somebody has to understand it without you standing next to them.

You now write the row count down without being asked. That is the habit, not the framework.

Carry these to your desk

Three takeaways.

1 Ground the file. Do not paste it. Then ask for the row count every time.

2 Before you ask: which columns does this need, and how many rows have them all? Clean is not complete.

3 A number becomes information when you put something beside it.

Yesterday’s line still holds. The prompt is the product.

The hook into Day 3

Tomorrow is the top rung.

1 Data

2 Information

3 Knowledge

4 Wisdom

Wisdom, which sounds grand and is not. It is one practical question. You have a finding on that card, and it is a good one, tomorrow somebody has to understand it without you standing next to them.

I am going to open tomorrow with a dashboard. Every number on it is correct. And it will not answer the one question the person looking at it actually has.

Break

Lunch.

60:00

That is the day. Lunch runs until three o’clock. Bring your comparison card tomorrow.

Presenter & pointer controls

→SpacePg Dn

Next slide / reveal

←Pg Up

Back

L

Laser pointer (follows the mouse)

B.

Blank the screen

F

Fullscreen

S

Sound on / off (starts off)

T

Start 60-second exercise timer

?

This panel

Esc

Close / resume

Most USB presenter remotes send Page Up / Page Down. They already drive this deck.

From Datato Insight

You finished holding clean data.

Industrial Branch handled 257 service requests this year.

A number on its own is not information.

By the end of today, here’s what changes.

Four rungs. We climb two today.

Today is the middle two.

An answer in a chat is a screenshot.

Ctrl + T first. Then ask.

Track A writes into your sheet. Track B answers beside it.

Work the full Work Order Set.

Ctrl + T. Then ask.

Ten minutes.

Eight steps. Always this order.

How big is it?

What is missing?

Add it up.

It reads clean.

Add it up honestly.

Ahmadi was fastest.

A rumor with a decimal point.

Ten minutes.

This block we write the prompts together.

You already knew what was missing.

Four things you can put beside a number.

Beside the others.

Beside itself, earlier.

Every site improved seventy percent.

Beside an outside mark.

Three findings. All true. One number.

Ten minutes.

Same eight steps. You drive.

Every column is full. Is it clean?

Add it up honestly. Again.

Different cause.

Somebody will suggest making them all positive.

Facilities cut costs by a quarter.

Two files. Two industries. Two rank flips.

Hands up who wrote delete it.

Three questions.

Three sentences turn a finding into something usable.

What you climbed.

Three takeaways.

Tomorrow is the top rung.

Lunch.

Presenter & pointer controls

From Data
to Insight