Systems don't fail when it's calm. They fail at 3am when everybody's tired and something already went wrong.
Your follow-up process was designed on a normal week: full team, normal volume, nobody sick. That's not the week that costs you deals. The week that costs you deals is the one where two people are out, volume spiked, and somebody's kid got sick on Wednesday.
The problem is that a follow-up process doesn't announce when it breaks. Leads don't file complaints. They just buy somewhere else, and the report still looks like a report.
Ask one question about every step in your follow-up: if the person who normally does this is out for a week, what happens?
Walk it step by step. Who assigns new leads. Who works day-two follow-up. Who checks the voicemail box. Who reviews the no-contact list. Who notices when a lead's been sitting for six days.
For each one, the answer is either a name, a role, or nothing. Names are fragile. Roles are better. Nothing is where your leads go to die, and most shops have at least two of them.
Pay particular attention to the steps nobody was ever explicitly assigned. Those got absorbed by whoever cared most, and when that person is out, the step simply doesn't happen and nobody notices for a week.
A campaign hits, or a mailer lands, and lead flow doubles for four days. Follow-up is the first thing that gets deprioritized because new leads feel more urgent than day-five leads. So the backlog grows exactly when it's most expensive.
Test it on paper: at double volume, how many touches does each lead actually get? If the answer drops, you don't have a follow-up process, you have a follow-up process that works at current volume.
Not a planned vacation with a handoff. A Tuesday morning text saying they can't come in. Which steps just stopped, and who found out?
A deal falls apart at closing. A big one. The whole team spends two days on it. Follow-up on everything else goes quiet, and there is no report anywhere that shows you the cost of those two days.
An automation stops firing. A number gets flagged. A form stops posting to the CRM. The dangerous version of this is not the loud failure, it's the quiet one, where everything looks normal and the leads just stop arriving in the place people check.
Ask what would tell you. If the answer is somebody would probably notice, you've found the gap. Set an alert on volume dropping, not just on errors, because a silent failure produces no errors at all.
Pick a normal week and run this on paper in about an hour.
List every follow-up step and name the human who does it.
Cross out your best person. Walk the list again. What breaks?
Double the lead volume in your head. Walk it a third time. What gets skipped first?
Find every step where the answer to who does this is nobody, or it depends.
For each gap, write down one of three answers: a backup human, a written procedure somebody else could follow cold, or something automatic that doesn't need a person at all.
You're not trying to fix everything. You're trying to know where it breaks before a bad week tells you.
You fall to what you practiced. That's true on a fire ground and it's true on a phone floor. When it gets loud, nobody consults the follow-up document. Everybody does what's automatic.
Which means the only follow-up steps you can count on during a bad week are the ones that either don't require a decision, or don't require a person. Everything else is aspirational.
So the useful question isn't whether your process is good. It's how much of it survives a week where nobody has the energy to be good.
Across most shops, the failures cluster in the same places. Check these before anything else.
Day one gets worked because it's on somebody's screen. Day two is the first step that requires somebody to go looking, and it's the first thing dropped when the day gets full. If your process has a genuine drop-off point, it's usually right here.
Leads that never got a conversation don't generate a task, don't age visibly, and don't belong to anybody. They accumulate silently. Ask who reviews that list and how often, and if the answer is nobody in particular, that's your largest single pool of recoverable value sitting untouched.
I'll call you Thursday is a commitment held in one person's head. It survives a normal week. It does not survive a week where that person is out or buried, and the seller experiences it as being blown off, which is worse than never having promised.
If a step begins with someone spots that, it is not a step. It's a hope. Under load, nobody spots anything.
You don't need a manual. You need three short procedures that somebody else could execute cold, without asking a question.
Pick the three steps where absence costs the most, and for each one write down: what triggers it, what exactly to do, and what done looks like. Half a page each. Put them where the team already looks, not in a folder.
The test for whether it's written well enough: hand it to the newest person on your team and have them run it for one day while the usual person watches without helping. Whatever they ask about is the part you didn't actually write down.
Don't try to harden all of it. Find the single step where a week of absence costs you the most deals, and fix that one properly. Then the next.
Most operators already know which one it is. If you had to guess right now which part of your follow-up would go first in a bad week, you'd probably be right. Go check whether you are.
Consent, disclosure, opt-out and calling hours. The four questions that separate a vendor who built this carefully from one who did not, and what a good answer sounds like.
Your average speed to lead is hiding the leads you lose. Here is how to pull it by hour of day and find the window where your response time falls off a cliff.