Credits and billing
What the learning month costs
The first month costs more per result achieved, and that is not a fault: it is what you pay once.
Looking at the first month's usage, the usual reaction is to think this will be expensive. It rarely is: what you are seeing is not the cost of working, it is the cost of learning to work — and that is paid once.
Where the extra early spend comes from
| Cause | Why it happens | When it disappears |
|---|---|---|
| Requests that had to be redone | The message was not clear first time | As soon as the wording is fixed |
| Test sends to colleagues | You need to see it from the other side | In the first week |
| The same thing requested twice | Two people did not know the other was asking | Once you split who asks for what |
| Documents uploaded and re-uploaded | You learn quickly which format works | Almost immediately |
Important
Worth knowing before drawing conclusions from month one: **that usage is not spread evenly, it clusters in the first few days**. Divide the month's spend by four weeks and project a year and the figure is an exaggeration — and it is exactly the calculation taken to management when the first invoice lands. The number to plan with is week four's, not the month's average.
How to shorten that month without skipping the learning
- 1
Run the full circuit once, on a small real case
One complete run teaches more than ten half ones.
- 2
Before a large batch, send to two
If the wording or the template is wrong, it shows there.
- 3
Split who asks for what from the start
Asked twice is the silliest spend of all.
- 4
And fix the message the moment anyone asks
Every doubt resolved is saved on all later sends.
Watch out
There is an early cost that appears in no account and is the dearest of all: **the things done outside the system because the process was not ready yet**. Documents requested by separate email, papers left on a desk, a batch handled by hand «just this once». That consumes nothing and costs a great deal, because it has to be redone or you live with a hole in the record.
Which number to budget with
Worth knowing
Each organisation can have its own consumption policy, so your account's detail overrides any general estimate. What does not change is that the first month is no basis for projecting anything.
›How long until it settles?
Usually the second full cycle of the process, not the second month.
›Should spend be capped while learning?
Yes, and it is easy: decide who can launch large sends.
›What if month two is still high?
Then it is no longer learning: look at what is being repeated.
A real case
The situation
A company projects the year by dividing its first month's spend by four.
What you do
- Uses week four as the reference and checks it against the year's busiest month
What you get
The budget stops being inflated by the first days' test sends.
The situation
The first month consumes more than expected.
What you do
- Counts the learning month separately
What you get
The year's forecast is not distorted.
The situation
Actions are repeated out of unfamiliarity.
What you do
- Checks consumption by user and trains where needed
What you get
The learning curve shortens.
The situation
Testing happens in production and consumes.
What you do
- Tests with a small, real case
What you get
Learning costs little.
The situation
The tool is dismissed over the first month's cost.
What you do
- Compares with the second month's consumption
What you get
The decision uses the settled figure.
The situation
Nobody knows how long the initial overspend will last.
What you do
- Compares consumption month by month
What you get
You see when it settles.
This article answers
- we spent a lot in the first month
- how much does a pilot consume
- cost of learning to use it
- budgeting the first months