The Dormant Inbox Problem: Inactive Fans in Your Inbox

Our data shows conversion collapsing between 2am and 6am and peaking in the evening. That single split, measured across 42,852 conversations, is also the explanation for the part of the inbox nobody works: the majority of threads in any working inbox that never move again.

The dormant tail only gets worked with hours a team can spare; spare hours are the ones nobody wants; and the hours nobody wants are the ones the data says are worth least. That is not a discipline problem.

Why does an inbox go mostly dormant?

Arithmetic, not fan behaviour. Threads arrive continuously, attention arrives in shifts, and a shift can only read and answer so many conversations properly. What does not fit does not disappear; it accumulates, faster than any team clears it.

Dormancy is what that gap leaves behind, and it has four distinct causes that get treated as one:

How it went dormant	What the thread looks like now	What it would take to restart
Never answered at first contact	One message from the fan, no reply	An opening as if it were new: nothing was built
Answered once, then dropped in the queue	Short exchange, no conclusion	A reference to the exchange itself, acknowledging the gap
Stalled after a refusal	Real conversation, one clear no	A different subject entirely, not a cheaper version of the same offer
Subscription lapsed	Gone from the feed, thread intact	A separate workflow (see win-back message examples)

Only the middle two hold much value per thread. The first and last are cheap to touch and expensive to work, exactly the distinction a team under pressure never has time to make.

What did we measure, and what does it not show?

The measurement is about hours, not about dormancy. Across 42,852 conversations covering 1.5 million messages and 38,879 sales, we compared conversion by time-of-day band: it collapses in the 2am-6am window and peaks in the evening, the finding developed in what time fans buy.

What that does not establish, and it matters here more than usual:

• It does not measure the tail. We compared hours, not dormant threads against live ones. That most of an inbox never moves is an observation about how inboxes work, not a corpus finding, and we will not dress it up as one.
• It does not say a night sale is impossible. It says the band converts far worse on average, which is a statement about where to spend effort, not a prohibition.
• It does not tell you whether the same tail worked at peak would convert. Nobody ran that comparison, because no team has ever had the hours to run it.
• It is observational. Chatters were not assigned to hours at random, so what the hours are and who is working them are tangled together. The whole boundary is set out in what our data does not show.

The honest summary: the hours finding is solid, the mechanism drawn from it here is an argument: one every agency owner recognises the moment it is stated.

Why does no human team ever get to them?

Because live threads never run out, so the tail is always the thing that can wait. Each decision is individually correct (a conversation in progress has better odds than a thread that went quiet months ago), and the sum of them is an inbox nobody ever opens the back of.

Four approaches, and what each one runs into:

Approach	What it costs on paper	Why it stalls
Work the tail between live threads	Nothing	There is no between: live threads refill faster than they close
Reserve an off-peak slot in the rota	A slot taken from cover	It lands in the spare hours, and the spare hours convert worst
Hire night cover for the backlog	A full shift	You buy the 2am-6am band at full price, the block our data shows collapsing
Broadcast to the whole tail	One send	It moves the tail once, cannot be spent twice, and wastes every reply nobody works

The broadcast row deserves a note, because it is what most teams do. A mass message is the right tool for making a dead pile move at all, and the wrong tool for everything after: the replies land back in the same inbox, behind the same shortage of hours.

What changes when something works the tail continuously?

The economics, not the persuasion. A dormant thread touched by something always on does not take an hour from a converting conversation, and that removed trade-off is the whole difference.

What that changes, in order of importance:

• The tail stops competing with the evening. Peak hours go to threads near a decision; the quiet pile gets touched in hours that were being wasted anyway.
• Continuity replaces campaigns. A tail worked once a quarter is a campaign; a tail touched steadily is a state.
• Restarts get done properly. Reopening a dormant thread puts you back near message one, and our data is unambiguous that an offer placed before the sixth message converts roughly a third worse, the trap anyone clearing a backlog under time pressure falls into. The rule is developed in when to pitch a sale.
• Nothing is deleted to make the queue look manageable. Threads that cost nothing to hold stop being purged for tidiness.

And what must not change, because the same finding cuts both ways:

1. Do not sell in the dead hours. Being able to work the 2am-6am block is not a reason to make an offer in it. Warm the thread, hold the offer for the evening.
2. Do not send the same opener to the whole pile. The four causes need four different first messages, and a fan can tell.
3. Hand over the moment money appears. A thread that comes alive with a real buying signal belongs to a person.
4. Keep the record. A thread touched three times with no reply is a decision to stop, not a reason to touch it a fourth time.

What do you do tomorrow morning?

Start by finding out how big the problem is, because almost nobody knows. Three moves:

1. Count the dormant share. Threads with no message either way for thirty days, against total threads. Write it down; it is usually worse than the guess.
2. Log which hour tail touches happen in. Not how many. Which hour. If the answer is the overnight block, you have found where your reactivation effort is going.
3. Split the pile by the four causes above before writing a message. Two of the four are worth a person; two are worth the cheapest touch you have.

Then set one rule and hold it for a month: no offer goes into a reopened thread before the exchange is running again. It costs nothing to adopt.

How do you check this on your own inbox?

By splitting your own numbers the same way we split ours, which needs no corpus and no tool.

• Conversion by hour band. Take one closed month, bucket sales by the hour they closed, and compare the overnight band against the evening. Use the fan’s local hour, not yours.
• Reactivation by hour. Of the dormant threads you did reopen, compare those opened off-peak with those opened in the evening. If the evening ones do better, your rota is spending reactivation in the wrong place.
• Offer position on restarts. Record how many exchanges preceded the first offer in a reopened thread. It is the number most likely to explain a poor win-back run.
• Hold it for two months. Thread-level data is noisy, and one month of tail work tells you nothing.

If your inbox contradicts the pattern, believe your inbox. What will not change is the arithmetic underneath: threads accumulate whether or not anyone opens them, and the hours left over for them are the ones worth least. The triage deciding what gets opened first is covered in managing fan conversations.