One inbox can be the problem on a solo ad, and the sample is small

Answer One inbox provider can be the problem, and a single account does not make it a law. On August 17, 2026 Andrew Beauchamp described a personal-care sender on a new domain. About 7,000 quiet profiles were held out of alternating campaigns for three weeks. He says the sample is small, and that the drop in Microsoft bounces is the cleaner signal. Do not buy a solo ad because you want his table.

The note that started this: How We Fix Gmail Open Rates, Andrew Beauchamp, August 17, 2026.

The table is still worth reading, because it separates providers. Blended opens hide the mailbox that is punishing the domain. A solo ad is one blast across whatever mix the seller has. If that mix is mostly the mailbox that is already unhappy, you rented the unhappy slice and called it a list.

Beauchamp's update sits at the end of a longer piece about excluding unengaged people at the problem providers instead of shrinking the whole send. The three-week comparison is a second account, a personal-care brand recovering on a new domain, with Gmail and Microsoft lagging. Tighter windows there. About 7,000 profiles excluded per send. Alternating campaigns. Treated versus untreated: Gmail opens 37 to 39 percent versus 44 to 47 percent. Microsoft opens about 26 percent versus 34 to 39 percent. Microsoft bounces 0.47 to 0.62 percent versus 0.07 to 0.14 percent. Revenue per send $1,890 versus $2,535. He writes that revenue per send went up on the treated mail, and that excluding those profiles cost nothing because they were already dead.

What did Andrew Beauchamp report after three weeks?

On one personal-care account, about 7,000 profiles were excluded per send on alternating campaigns. Untreated Gmail opens were 37 to 39 percent versus 44 to 47 percent treated. Microsoft opens were about 26 percent versus 34 to 39 percent. Microsoft bounces were 0.47 to 0.62 percent versus 0.07 to 0.14 percent. Revenue per send was $1,890 versus $2,535.

Keep every noun attached. One brand. New domain. Three weeks. Alternating sends, so the comparison is not a lab with a holdout designed in advance. Opens moved. Bounces moved further, and bounces are harder to fake with a privacy proxy. Revenue per send moved up, which is what persuaded them to keep the exclusion. None of those sentences is a forecast for a stranger's solo ad on a stranger's file.

He also says opens are noisy. Privacy features fire them for some people and hide them for others. That is why he does not want the open column to be the only reason you exclude a person. In the method he describes, a clicker stays, a recent buyer stays, a recent browser stays. The excluded person is quiet on every signal they can see. A solo ad seller who cannot see those signals, and still mails the whole file, is mailing the slice this operator pulled out.

• The numbers belong to one personal-care account.
• The window is three weeks of alternating sends.
• Bounces are the cleaner line in his own words.
• A rented drop cannot rerun his exclusion for you.

What caveat does he put on those numbers?

He writes that the sample is small, that the treated sends came later while the domain was healing, and that some of the open-rate lift may belong to the healing domain. He calls the bounce drop the cleaner signal.

Read that before you screenshot the open column. The treated mail happened after the domain had already started to recover. Time is a confound. He says so. The bounce change is the part he is willing to trust more, because a bounce is a failure, not a tracking pixel. A buyer who quotes 44 to 47 percent as what solo ads do now has deleted the caveat and the sample size in the same breath.

If you sell mail, you may still learn the shape. When one provider lags, tightening only that provider can beat punishing the readers who were fine. That is a segmentation choice on a list you operate. It is not a promise you make to a guest buyer about tomorrow's open rate. Say the caveat if you cite the table at all. Better: cite your own last ten sends, by provider, with complaints.

Can a solo ad buyer demand the same exclusion?

No. You do not control the seller's segments. The useful question is whether they already suppress quiet addresses at the provider that is hurting them, and whether they will show you that provider's complaints. Their case is not your guarantee.

Ask which inboxes take the largest share of the drop. Ask what they do with addresses that have not opened, clicked, or bought. Ask for the complaint rate at the weak provider, not a blended trophy. If the answer is that everyone gets every offer, you are looking at the untreated column on purpose.

Solo ad pricing does not rise because a Portland account made more money per send after a cut. Price the list in front of you. Beauchamp's Monday note is a careful operator showing his homework, including the doubt. Borrow the doubt. Leave the dollars on his page.