Overdue Follow-Ups Cannot Count a Lead With No Date Set

Why this matters

The overdue follow-up count is one of the few numbers in a shop that is supposed to make you uncomfortable, and it is usually the most comfortable number on the page. It reads low, it stays low, and it stays low because the leads nobody ever scheduled a next step for cannot appear on it. The most neglected work in the business is invisible to the one number built to find neglect, and the fastest way to a perfect score is to stop setting dates.

What the count is actually made of

• The numerator is open leads whose recorded next follow-up date is earlier than right now.
• The denominator, whenever a percentage is shown, is open leads that carry a next follow-up date.
• A lead with no date is in neither. It is not overdue, it is not on time, it is not counted. It is not late for an appointment it never had.
• It is a snapshot, not a window. It describes this instant. Read it Monday and Thursday and you have two instants, not a trend, and nothing in it tells you how many leads passed through the overdue state and came back out. A sibling card owns the general problem of a snapshot figure read as though it measured a flow.

Everything below comes out of the third point.

Three states of an open lead, and the number sees one

An open lead is in exactly one of three states: it has a next date still ahead of it, it has a next date that has passed, or it has no next date at all. Only the middle one is the count, and only the first two are anywhere near it. A shop with 214 open leads on a Monday morning:

The number reported upstairs is 14. As a share of the leads carrying a date it is 14 of 96, 14.6 percent. As a share of every open lead it is 14 of 214, 6.5 percent. Both are true and neither is the thing an owner wants to know, which is how many open leads have nobody assigned to a next action. That figure is 14 plus 118, so 132 of 214 open leads, 61.7 percent.

Two numbers, an order of magnitude apart, describing the same Monday.

The companion count, and how to age it

The overdue count does not need fixing. It needs a companion printed beside it, because on its own it is a fact about the date field rather than about the pipeline.

Report open leads with no next date, aged by days since the last logged activity. Aging is what turns an undifferentiated pile into a decision, because a lead that came in yesterday and has no date yet is not the same animal as one that has been silent since spring. The same 118:

Set the working gate at no next date and no activity for more than 14 days. Fourteen days is short enough that a genuinely warm lead has not gone cold inside it and long enough that a lead taken on Friday is not on a list on Monday. Tune it to your own sales cycle; a trade quoting work that gets decided over a season can defensibly run 30.

Against that gate, 90 of the 118 undated leads qualify: 76.3 percent of the undated, and 42.1 percent of all 214 open leads. Set that beside the headline: 90 leads in that state against 14 the reported count can see, a factor of 6.4 on two counts of open leads taken at the same instant.

The 24 leads over 90 days are not a follow-up problem any more. They are a closing problem, and they belong in the next section's rule.

A falling overdue count is ambiguous between work and keystrokes

The overdue count falls in three ways, and only one of them is work:

1. Somebody worked the lead. A contact happened, a next step was agreed, the date moved forward. This is what the number is supposed to measure.
2. Somebody moved the date. No contact, no conversation, the date is now next Tuesday. Costs about two seconds and scores identically to the first.
3. Somebody removed the date. The lead leaves the denominator entirely and the count improves twice over, because it drops from the numerator and shrinks the base.

Nothing in the count separates them. The tell is not the date, it is whether the activity record moved. On paper it looks the same: a rescheduled line with nothing written beside it.

Audit it directly rather than trusting it. Pull 10 leads whose next date moved forward in the last week. Each should carry either a logged contact attempt or a one-line note naming who asked for the later date. Fewer than 8 of the 10 and the count is tracking keystrokes rather than follow-up, and no amount of reporting on top will fix that. This shop ran the check and 9 of the 10 carried one, which clears the gate.

The rule: a next date or a close, no third option

Every open lead carries a next follow-up date, or it gets closed with a reason. There is no third state.

That is the whole fix, and it works because it removes the invisible category rather than trying to measure it. Once no lead can exist without a date, the overdue count becomes what everyone already believed it was: the count of leads whose next action was missed.

The objection is always the same, and it is worth answering with the actual arithmetic. Nobody is asking you to schedule 118 follow-ups. Most of that pile is not warm and never was, and the honest disposition of an old undated lead is a close, not a date. Closing it also produces a reason code, which is the input a lost-reason ranking needs and normally never gets from stalled leads.

Hold two lines when you do it. A close is not a deletion, so the record survives and stays searchable, because a lead closed in March that calls in June should show its history. And do not let "set a date" become a way of avoiding a close: a date more than 90 days out with no customer-stated reason is a close wearing a date.

The same shop, four weeks later

Take the same 214 leads and nothing else, so the comparison is a cohort rather than a mix of old and new.

Of the 118 undated leads:

• 44 were closed as lost with a reason, and 31 of those 44 came from the two oldest aging bands.
• 13 were removed as duplicates or as leads already sitting on the schedule under another record.
• 61 were given a real next step and a date.

That is 57 of the 118 that never needed a date at all, 48.3 percent, so the rule's real cost was 61 dates rather than the 118 it looked like from a distance. Two people cleared it in an afternoon.

The cohort now stands at 157 open leads, 214 less the 44 closed and the 13 removed, and every one of them carries a date. Overdue among them: 21, which is 13.4 percent of the 157.

Read those two moves carefully, because they run in opposite directions and that is the point:

• The reported overdue count went up, from 14 to 21. Nothing got worse. Leads that were invisible became countable, and 7 of them were already late.
• Open leads with nobody assigned to a next action went from 132 to 21, a fall of 111 leads or 84.1 percent, on the same construct at both instants.

A shop that manages to the headline number alone experiences this change as a 50 percent deterioration and reverses it. That is not a hypothetical failure mode; it is the normal outcome of reporting one of these two numbers without the other. Print them together or the improvement reads as damage.

The other blind spot runs forwards

The count only looks backwards. A date in the future scores as on track no matter how far into the future it is, so a pipeline can be parked rather than worked and the number will not flinch. Split the 82 leads that were dated ahead on that first Monday:

Twenty-three of the 82, 28.0 percent, sit more than 30 days out. Most of those are legitimate: a customer who asked to be called before next season, a property manager on a budget cycle. Nine sit beyond 90 days, and when somebody read them, 4 carried a customer-stated date and 5 did not. By the rule in the section above, those 5 are closes wearing a date.

Five leads is a small number and it is the right size for this check, because the point is not the count. It is that pushing a date out is the one move that improves the overdue number and the on-track number at the same time, so it is the move a pipeline drifts toward on its own without anybody deciding to do it. Read the forward distribution once a month and the drift shows up while it is still five leads. In this shop's four-week sweep those 5 are still sitting inside the 157 and still scoring as on track, because the sweep only worked the undated pile.

Two ways to score perfectly without touching a customer

The cheapest way to see what a number is really made of is to find the inputs that break it, and this one has two that take no work at all.

Delete every next date. All 214 leads lose their date. The dated population is zero, the overdue count is zero, and the percentage is undefined or reported as clean. Best possible score. Customers contacted: none. A measurement whose perfect score is reachable by deleting data is measuring the data.

Set every next date to today. All 214 leads are dated, nothing is overdue this afternoon, and tomorrow morning all 214 are. Same work performed, opposite score, one day apart, and both readings are accurate.

Between those two cases sits every real shop, which is why the count moves with recording habits more than it moves with follow-up. That defect has a general form, derived in the card named in the references, and the undated count above is this number's version of the repair it prescribes.

References

• See related: SLA Compliance Only Counts Jobs That Carry a Deadline, which owns the opt-in denominator pattern
• See related: Days Sales Outstanding and the Mismatch Inside It, for a snapshot figure read as though it measured a flow
• See related: Lost Reasons Are Only as Good as the Field Being Filled, for what the close half of the rule produces
• See related: The Follow-Up Automation Worth Setting Up Once, for reducing the number of dates a person has to set by hand