Segments
List rebuilt 03:12 · 55,400 people, 18,598 we may write to
A segment here is a rule, not a stored number. Every figure on this page was produced by running the rule over the list when the page was drawn — which is why a clause can be taken apart and counted on its own.
Every cut of the list
A row is a rule. The figures beside it were produced by running that rule over the list a moment ago.
“May be written to” is people with an address the shop is allowed to use, permission on the record and a working mailbox. On the marketplace-only cut it is a rounding error, and that is the truest figure on this page.
Lapsing
Bought more than once, then stopped. Ninety days is about two bags for a household, so this is the point where somebody has started buying coffee somewhere else.
The rule, clause by clause
- 2 orders or more16,082
- last order between 90 and 180 days ago7,980
- All of them together3,084
Each bar is what that clause alone would select from the whole list. The clause closest to the total is the one carrying the segment; the ones far above it are barely narrowing anything.
Shares people with
- Marketplace only1,640 · 53%
- Bought the kit, not the coffee865 · 28%
- Cold brew, last summer620 · 20%
- Top spenders165 · 5%
- Birthday this month164 · 5%
The window closes at 180 days on purpose: past that the message has to be a different message, and it is a different segment.
Who it reaches
The same segment down both channels. Nothing is stored: each ladder is five filters over the 3,084 people the rule selected.
- In the segment3,084
- −Marketplace relay address−1,514
- −Never gave permission−214
- −Took permission back−43
- −Bounced hard before−20
- May be written to1,293
- −In the hold-out group−136
- −Already had 2 this week−9
- Will receive it1,148
223 of the people who fell out could be recovered by the shop — by asking for permission again, by collecting a detail it never asked for, or by not sending something else to them first. The rest could not.
SMS
- In the segment3,084
- −Marketplace order, no number−1,514
- −No mobile number−611
- −Never gave permission−487
- −Took permission back−66
- −Number no longer works−6
- May be written to400
- −In the hold-out group−39
- −Already had 2 this week−2
- Will receive it359
1,100 of the people who fell out could be recovered by the shop — by asking for permission again, by collecting a detail it never asked for, or by not sending something else to them first. The rest could not.
A send’s own ladder goes one step further than these two: it also takes off the people already carrying a message that day, and the group held back on purpose so the result can be measured.
A few of the records
Exactly as the channel hands them over. The shop never sees more than this either.
| Record | Came from | Orders | Spent | Last order |
|---|---|---|---|---|
| s***@gmail.com | Web store order | 18 | ₺35,173 | 102 d |
| k***@outlook.com | Web store order | 5 | ₺13,202 | 114 d |
| w***@icloud.com | Web store order | 2 | ₺4,416 | 173 d |
| n***@gmail.com | Web store order | 3 | ₺3,043 | 101 d |
| Masked · Trendyol relaycannot be written to | Marketplace order | 2 | ₺1,745 | 111 d |
| Masked · Hepsiburada relaycannot be written to | Marketplace order | 5 | ₺5,527 | 107 d |
What the last sends did
- Still grinding? · 11 Sep1,148 reached · 0 orders above the held-back group, ₺0
- BREW10 · 25 Aug1,148 reached · 25 orders above the held-back group, ₺17,497
Where the cuts overlap
People shared by each pair of segments, counted. This is the page's argument: a segment is not a box, and two sends in one week can arrive at the same person.
| Segment | Gone quiet, over six months | Bought the kit, not the coffee | Cold brew, last summer | On the list, never ordered | Regulars | Lapsing | First order, last 30 days | Top spenders | Shared, total |
|---|---|---|---|---|---|---|---|---|---|
| Gone quiet, over six months | — | 5,013 | 5,433 | — | — | — | — | — | 10,446 |
| Bought the kit, not the coffee | 5,013 | — | 392 | — | 882 | 865 | 378 | 401 | 7,931 |
| Cold brew, last summer | 5,433 | 392 | — | — | — | 620 | — | — | 6,445 |
| On the list, never ordered | — | — | — | — | — | — | — | — | 0 |
| Regulars | — | 882 | — | — | — | — | — | 1,006 | 1,888 |
| Lapsing | — | 865 | 620 | — | — | — | — | 165 | 1,650 |
| First order, last 30 days | — | 378 | — | — | — | — | — | — | 378 |
| Top spenders | — | 401 | — | — | 1,006 | 165 | — | — | 1,572 |
| Shared, total | 10,446 | 7,931 | 6,445 | 0 | 1,888 | 1,650 | 378 | 1,572 | 30,310 |
The diagonal is blank: a segment shares all of its people with itself, and printing that would flatten the scale. Where two segments overlap heavily, the calendar takes the second send’s share of them off — which shows up on the send as a named step called “in another send today”.
Sample data. The overlap figures are exact counts of shared records, not estimates: two segments that share people are two sends that can arrive at one person in a day, and the calendar takes the second one off.