Cross-Platform Measurement: Why the Numbers Never Match

You run two advertising channels.

The first says: I brought you forty sales. The second says: I brought you thirty-five.

And your sales for the month: fifty.

The rule this course rests on

Every platform claims the same sale — and the sum of their claims exceeds your sales.

Read it twice.

And it is not lying. Each platform tells the truth from its own angle: **"this person saw my advert

then bought" ← and they may have seen two adverts, searched, asked a friend, then bought**.

And the mistake is not in the platforms ← but in adding their numbers together.

Three reasons the numbers never match

1. The window ← **each platform credits itself with what happened within a certain period after the

interaction ← and the periods differ** (lesson four).

2. The model ← **who gets the credit when more than one source is involved: the first, the last, or all

of them? ← and each platform answers in its own favour** (lesson five).

3. Double counting ← the same event sent twice and counted twice ← **and this is a technical error

that gets fixed** (lesson three).

And all three together mean: the numbers can never match ← **and the goal is not for them to match,

but to know by how much they differ and why**.

Four questions this course answers

1. What exactly do I count? ← (lesson two).

2. Why is my number on the platform different from my number in my store? ← (lessons three and four).

3. Which channel actually deserves the credit? ← (lesson five).

4. What do I do about what no platform sees? ← (lessons six and seven).

And try them now: add up what your channels claim this month, and compare it against your sales ←

and the difference is why you are here.

What this course is not

It is not about measuring one channel — that is in "Running Google Ads Yourself", lesson six, **which

covers defining a conversion and the four ways figures deceive you inside a single channel. This is about

the contradiction between two channels or more.**

Nor is it about your own data — your questions, your inventory and your comparisons are in "Analysing

Your Business Data", and we will use its rules in lesson seven.

Nor is it about ownership and permissions — that is the "Business and Ad Accounts" course, **and the

measurement tool is an owned asset like any other**.

And I will not explain the interface buttons or name a tool. **Tools change their names and screens every

few months ← and what does not change is that all measurement rests on four things: what you count, how

the event arrives, which window it is credited to, and which model distributes the credit**.

And I will not mention a single figure of my own — nor a "usual" gap rate ← **these differ by field,

channel and market, and a published figure becomes a reference the reader measures themselves against on no

basis**.

When this course does not concern you

Three cases, and I will say them before you continue:

1. You have only one channel ← so there is no contradiction to resolve ← **read lesson six of

"Running Google Ads Yourself" and it is enough**.

2. Your sales are few enough that you know where each one came from ← so ask rather than measure ←

and asking is more accurate than any tool at small counts.

3. You do not spend on more than one channel ← and do not intend to ← **so this course solves a

problem you do not have**.

And the second is the one people resist: measurement is a tool for large counts — **and at small

ones, the direct question is more honest and cheaper**.

What you need to start

• At least two advertising channels — or a channel and a free source
• Your actual sales figure for each month ← from your store or your book
• One definition of what you count as a conversion — and you will write it in lesson two
• Half an hour a month

Action steps

1. Write the rule: the sum of platform claims exceeds your sales.
2. Add up what your channels claim this month and write the total.
3. Write your actual sales beside it.
4. Compute the difference — and do not fix it yet.
5. Stop adding platform numbers together in any report.
6. Go honestly through "when this does not concern you", starting with the second.
7. Do not compare yourself against any published rate.
8. Write the four questions and record where you hesitated.