Results
1 Aug – 31 Aug · vs previous period: 1 Jul – 31 Jul
Every send in the period, beside the group that was sent nothing. The two never appear as one figure: what the people who received the messages ordered is counted apart from what the hold-out ordered, and the difference between them — measured once, over the period, not once per send — is the only thing on this page called revenue.
1 Aug – 31 Aug · what the messages added
₺92,807
Written to
0.026
orders each · 7,132 people
The hold-out
0.008
orders each · 775 people
The difference
0.019
132 orders in all
One comparison, over the whole period: everybody who received a message against the 775 people in the list’s hold-out group, who were sent nothing by anybody. Across the period’s 4 sends, 7,563 messages went out, 189 orders followed within the week and the till took ₺132,780 — but most of those people were coming back with or without the message, which is precisely what the number at the top takes out.
775 people were held back and placed 6 orders between them, which is enough to set a rate against. Holding 10 people in every hundred out of everything is what this costs, and it is a decision a person made. 28 people left the list over the same period.
BREW10 reminder
SMSTo Regulars · Fri 28 Aug
Received it
- People
- 509
- Ordered within a week
- 214.1%
- Revenue
- ₺15,120
- Unsubscribed
- 20.39%
In the hold-out · sent nothing
- People
- 40
- Ordered within a week
- 12.5%
The list’s own hold-out — 10 in every hundred, sent nothing by anybody — as it falls inside this segment.
₺10,511 of the period’s lift over the hold-out is allocated to this send — ₺21 for every person it reached. Allocated, not measured: the measurement is the block at the top of the page, and this is its share by orders. The ₺15,120 on the left includes everybody who was coming back anyway, which is what the hold-out is there to take out.
No open rate exists on SMS · 52 clicked
BREW10
E-mailTo Lapsing · Tue 25 Aug
Received it
- People
- 1,148
- Ordered within a week
- 363.1%
- Revenue
- ₺24,840
- Unsubscribed
- 70.61%
In the hold-out · sent nothing
- People
- 136
- Ordered within a week
- 10.7%
The list’s own hold-out — 10 in every hundred, sent nothing by anybody — as it falls inside this segment.
₺17,497 of the period’s lift over the hold-out is allocated to this send — ₺15 for every person it reached. Allocated, not measured: the measurement is the block at the top of the page, and this is its share by orders. The ₺24,840 on the left includes everybody who was coming back anyway, which is what the hold-out is there to take out.
325 opened · 82 clicked
What comes off the roaster
E-mailTo On the list, never ordered · Fri 21 Aug
Received it
- People
- 4,301
- Ordered within a week
- 390.9%
- Revenue
- ₺20,280
- Unsubscribed
- 170.40%
In the hold-out · sent nothing
- People
- 468
- Ordered within a week
- 10.2%
The list’s own hold-out — 10 in every hundred, sent nothing by anybody — as it falls inside this segment.
₺19,031 of the period’s lift over the hold-out is allocated to this send — ₺4 for every person it reached. Allocated, not measured: the measurement is the block at the top of the page, and this is its share by orders. The ₺20,280 on the left includes everybody who was coming back anyway, which is what the hold-out is there to take out.
1,409 opened · 173 clicked
Guatemala is back
E-mailTo Regulars · Mon 17 Aug
Received it
- People
- 1,605
- Ordered within a week
- 935.8%
- Revenue
- ₺72,540
- Unsubscribed
- 20.12%
In the hold-out · sent nothing
- People
- 161
- Ordered within a week
- 31.9%
The list’s own hold-out — 10 in every hundred, sent nothing by anybody — as it falls inside this segment.
₺45,768 of the period’s lift over the hold-out is allocated to this send — ₺29 for every person it reached. Allocated, not measured: the measurement is the block at the top of the page, and this is its share by orders. The ₺72,540 on the left includes everybody who was coming back anyway, which is what the hold-out is there to take out.
656 opened · 137 clicked
Every send in the period
The blocks above, in one table. Nothing here is reachable only by reading the prose.
| Send | Segment | Sent to | Orders | Revenue | In the hold-out | Their orders | Allocated |
|---|---|---|---|---|---|---|---|
| BREW10 reminderSMSFri 28 Aug | Regulars | 509 | 21 | ₺15,120 | 40 | 1 | ₺10,511 |
| BREW10E-mailTue 25 Aug | Lapsing | 1,148 | 36 | ₺24,840 | 136 | 1 | ₺17,497 |
| What comes off the roasterE-mailFri 21 Aug | On the list, never ordered | 4,301 | 39 | ₺20,280 | 468 | 1 | ₺19,031 |
| Guatemala is backE-mailMon 17 Aug | Regulars | 1,605 | 93 | ₺72,540 | 161 | 3 | ₺45,768 |
| Total | 4 sends | 7,563 | 189 | ₺132,780 | — | — | ₺92,807 |
The last column splits one measurement between the sends, by their own orders, so its total is the figure at the top of the page and not a sum of eight separate answers. The two columns beside it are left empty rather than added: the same person is in the hold-out for every send they would have been in, so adding those counts together would count them several times over.
What this page will not do
Four refusals. They are the reason the figure at the top can be believed.
It will not measure a send against a handful of people
A hold-out cut out of one segment's recipients is fifty people, and fifty people who placed two orders cannot tell one order rate from another. So the group is the whole list's: 10 in every hundred, sent nothing by anybody, compared over the period rather than send by send. A single send gets a share of that figure, worked out from its own orders, and the column says allocated because that is what it is. The cost is real — those are orders the shop does not take — and a small shop has to hold back more of its list, not less, for the comparison to carry.
It will not credit one order to two sends
An order inside the window after two messages is counted once, against the message that went out last. Adding it to both is how a marketing report comes to claim more revenue than the shop took.
It will not average a measurement with an estimate
The order book is a measurement. A platform's modelled attribution is an estimate. Where the two disagree the difference is named and both are shown, exactly as the advertising desk does it; neither is quietly blended into a third number that is true of nothing.
Opens are not the result
An open is a loaded image, which half of the mail clients in the world load for their own reasons and the other half block. It is on this page in small type because a fall in it can still mean something; it is never the headline, and no send has ever been judged on it here.
Sample data. Every figure comes from the reach ladder and the hold-out group the list carries; nothing on this page is a rate somebody typed.