Arriving daily is fine. Selling is what costs a newsletter

Alongside the retail edition, today we are publishing the Inbox Index Publisher Edition, because a newsletter is not a retailer and should not be judged like one. Nobody deserves a penalty for arriving daily when daily is what you signed up for. Publishers are scored on their own three dimensions: does it keep its own rhythm, how much does it send, and how often is it selling you something rather than telling you something.

The comparison worth sitting with is two newsletters that arrive at the same pace.

TSN and NYT Wirecutter each arrived seven times a week, and both kept to their own rhythm almost perfectly, TSN on every send and Wirecutter on 98% of them. On volume and rhythm they are nearly identical. TSN sold nothing in any of them, while Wirecutter was selling in 30% of its sends. TSN scores 10 out of 12, and Wirecutter scores 7.

That is the edition working as intended. A newsletter you asked to hear from daily is not marked down for arriving daily, and what moves it down the table is the moment it stops telling you something and starts selling you something.

The top of the table is quieter still. Healthline’s Wellness Wire and TLDR Data score 12, arriving once or twice a week, on time, with nothing to sell. At the bottom, Front Office Sports arrived nearly fourteen times a week on an irregular rhythm and scores 3.

We publish each newsletter’s first week as its own column rather than folding it into the score, because averaging it in would punish brands for a burst that ends, and excluding it silently would hide a real experience: a sender that fires nine emails in week one and then goes quiet has still sent you nine emails.

See the Publisher Edition, numbers and all. It covers the complete quarter, 1 July to 30 September.

An earlier version of this post, written on provisional figures, led with how loudly newsletters greet a new subscriber. Those figures counted several newsletters from the same publisher as one. Scored one newsletter at a time over the full quarter, the finding does not hold, so we have removed it.