How do you tell a lost call from one that just booked a few days later?

The mistake is measuring the outcome at the wrong moment

A call ends. At that instant it is either booked or not booked, and that is a clean fact. But the business question is whether the opportunity was lost, and that cannot be known at the end of the call. The customer may call back, may fill out a form that night, may be entered manually by a dispatcher the next morning under a slightly different name.

If you count every unbooked call as lost, you will overstate losses. If you wait long enough to catch every late booking, your report is never current. The compromise is an explicit window, chosen once and applied consistently.

How the match is made, in order of reliability

• Caller phone number to customer phone number. Strongest signal, but it fails when the caller used a spouse's phone or the office entered a different contact number.
• Service address. Slower to match because it is often captured later, but it resolves household cases that phone matching misses.
• Name plus postal code. A weak fallback. Useful for surfacing candidates for a human to confirm, not for automatic credit.
• Job creation timestamp inside the window. Combined with any of the above, this is what converts a match into an attributed booking.

Choosing the window

Emergency trades book fast; the overwhelming majority of same-intent bookings happen within a day or two. Replacement and installation work behaves differently, because the homeowner is comparing options and may take weeks. Using one window for both distorts the slower category badly.

A workable approach is a short window for demand-type calls and a longer one for quoted or estimate-type calls, with the window recorded next to the metric. Wiring that logic into the operational system is normal integration work, and it is where a lot of the value in field service system integration actually sits.

What to publish

Publish two numbers side by side: booked on the call, and booked within the window. The gap between them is itself informative. A large gap means your team is generating interest but not closing on the phone, which is a coachable pattern. A small gap with a large unbooked remainder means the calls really are being lost.

Also keep an unmatched bucket rather than assigning it. Some customers cannot be matched at all, and a report that admits that is far more durable than one that quietly rounds the difference into whichever column looks better. Feeding both figures into revenue reporting keeps the marketing conversation honest.