The Follow-Up Problem: Why Deals Die in Your Inbox

Ask a solo founder why a deal died and you usually get a story about price or timing. Go back through the thread and it is often simpler: someone said "send me the proposal", and eleven days passed.

Nobody was outsold. The thread went quiet, and there was no mechanism to notice.

How much that costs is measurable at the front of the funnel. The MIT/InsideSales Lead Response Management study analysed three years of data — over 15,000 leads and 100,000 call attempts — and found the odds of qualifying a lead were roughly 21 times higher when contact happened within 5 minutes rather than 30. A separate audit of 2,241 companies found an average first-response time of 42 hours.

Those numbers are about inbound speed, not mid-deal follow-ups, so do not over-read them. The direction is the point: attention decays fast, and the decay does not pause because you got busy.

The two open loops#

Every meaningful conversation leaves behind two things:

• A promise you made. Send pricing, make the intro, share the doc.
• A wait you are in. They said they would come back after their board meeting.

Both are load-bearing. Neither is recorded anywhere by default. They exist in your head, in a Slack thread, in the middle of an email you have already scrolled past.

The failure is not laziness. It is that the loop is created in a conversation and never written down anywhere that will nudge you later.

This is also why open loops feel worse than they are. Masicampo and Baumeister found unfinished goals produce intrusive thoughts and measurably worse performance on unrelated work — and that the interference largely disappears once you write a specific plan, even before the task is done (JPSP, 2011). An open follow-up loop therefore costs twice: the deal at risk, and the attention it takes while it sits there un-planned.

Why a CRM does not fix it#

A CRM is a database of what you remembered to type in. If you had remembered, you would not have dropped the follow-up.

This is why founders end up with an immaculate CRM and still lose deals to silence. The tool records; it does not observe.

What actually fixes it#

Three pieces, in order:

1. Extraction, not entry

The commitment already exists in the email or the call. Something should pull it out — you promised pricing on the 14th; they owe you a decision after the 20th — without you retyping it into a second system. Anything requiring manual entry has the same failure mode as remembering.

2. A staleness signal

Not a task list. A list is a thing you look at when you already remember to. What you want is the inverse: something that notices a thread has been quiet for N days with an open loop attached, and tells you.

The number is not important — five days, ten days. What matters is that silence becomes an event rather than an absence.

3. A draft, not a reminder

"Follow up with Sarah" is a reminder that becomes another chore. What actually gets sent is a draft already written, referencing what you promised, waiting for you to read and hit send. The gap between "I should follow up" and "sent" is where the whole thing collapses.

The mechanical version#

If you want to build this without any new software:

1. Friday, fifteen minutes. Open sent mail from the last three weeks.
2. Write two columns: I owe them and they owe me.
3. Anything older than ten days with no reply gets a two-line nudge, written now, not scheduled.
4. Delete the list. Rebuild it next Friday.

This works. It is also the first thing to go in a busy week, which is exactly why software is the better answer if you can get it.

Where an assistant fits#

This is the part of the job Jarvis is aimed squarely at: it reads your mail on your Mac, extracts what was promised and by whom, notices when a thread has gone quiet, and drafts the nudge for you to approve. Locally, without your correspondence going to a vendor.

Two honest caveats. Extraction is not perfect — it will occasionally read a throwaway "I'll send that over" as a commitment, and you will dismiss it. And it drafts rather than sends, deliberately: an assistant that chases your customers unsupervised is a way to lose them faster.

The pattern behind it#

Inbox triage is about what needs you now. Follow-ups are about what needed you a week ago and still does. The first is a queue problem, the second is a memory problem — which is why doing inbox zero perfectly still leaves deals dying quietly.

If you fix one thing this quarter, fix the second one. It is where the revenue is. The inbox guide covers the first, and the wider admin picture covers where both sit.