Winning the Open Rate, Losing the Inbox
A new email sequence lifted opens above baseline. Two weeks later sender reputation collapsed, and password-reset emails stopped reaching anyone.
Email deliverability is the one growth metric that punishes you for winning. At one company, the growth team fired a sequence at a list of 50,000 prospects bought from a data broker. The first send opened at twenty-two percent, well above baseline. They celebrated. The second opened at eighteen. The third, eleven. By the fifth send, opens sat under five percent, and support was fielding complaints that password resets and billing notices had stopped reaching customers' inboxes.
The experiment had collapsed the company's sender reputation. High volume into a cold list tripped spam filtering at Google and Microsoft, and once the filters tightened, every message from the primary sending domain got caught, whatever its content. Password resets went to spam. Billing notices stalled. A marketing test nobody had checked against deliverability had broken the transactional email the whole product depended on.
Recovery took months. The team split marketing and transactional onto separate domains, cut volume, mailed only engaged segments, and worked warming protocols to rebuild reputation. Inbox placement crept back, but they'd paid for the lesson with a quarter of degraded operations and an unknown number of customers who churned while their password and billing emails were unreliable.
Deliverability is one of the most under-watched pieces of growth infrastructure anywhere. Teams stare at open rates, clicks, and conversion, all of which are outputs that depend on inbox placement. Placement itself rarely gets watched, and reputation damage stays invisible until it has already cascaded. It's the same trap as any vanity metric that reads green while the thing underneath it rots.
Google, Microsoft, and Yahoo treat sender reputation as a primary filter. They score it on engagement, complaints, bounces, and sending history. Mail engaged users at steady volume and reputation holds. Blast a cold list and it craters, and the damage compounds: once mail lands in spam, fewer people engage, which drags the engagement signal down further.
The diagnostics exist and go unused. Google Postmaster Tools gives domain and IP reputation for free. Microsoft SNDS does the same for Outlook. Whoever owns email should read both weekly. Open rates get inflated by image-loading proxies. Inbox placement, which the diagnostics show, is the truth.
Treat deliverability as a constraint on experiments, not a downstream result. Every change to sending, new list, new sequence, new volume pattern, gets checked against deliverability before it ships. It's the same discipline that keeps a signup test from quietly killing revenue: measure the second-order effect before you celebrate the first.
Splitting marketing from transactional mail is foundational and routinely skipped. Most companies send everything from one domain because it's simpler. Simpler is also fragile. Separate domains, separate IP pools, separate reputation tracks cost a little setup time and buy structural resilience.
The open rate they were proud of is the thing that shut the inbox. Winning the metric was how they lost the channel.