How the leads get worked

Anyone can send you leads. We can tell you what happened to every one of them.

Most operations cannot answer a simple question: which lead did we lose, and where. Every dial, text, email and appointment outcome in our system is timestamped against the rep who owned it, and none of it is editable by the person being measured. The machine reads its own data overnight — and when something is costing money it names the stage, the rep and the number. We built the software rather than renting one, because every off-the-shelf platform optimizes for storing leads and the job is working them. It ships inside the engagement; you never buy a seat.

app.palmerleadmachine.com/pipeline
Attention — newest first, always SLA clock running
DM
Dana M.
Roof replace · 00:14 since submit · unread
Work now
PS
Paul S.
Lighting · 01:47 since submit · text sent
Work now
RT
Ray T.
Solar · 2 attempts · callback 2:00pm
Scheduled
JL
Jen L.
Bath · 4h 12m untouched · cadence step 2
Slipping

Sorted by arrival, never by a score. The colored edge is a freshness clock, not a guess about who is worth calling — reps do not get to cherry-pick, and the oldest untouched lead cannot hide at the bottom.

In call — Ray T. 04:19 · connected
Showing as
(704) 555-0142
Line quality
Good
Attempt
3rd
Talk time today
41m
Book the appointment without leaving the call
Calendar unfolds from the call bar · disposition pre-fills

Reps call from the browser with a local-presence number, so nobody dials from a personal cell and no homeowner sees an out-of-state area code. If the call fails, the app says why in plain English — and stays silent rather than guessing when it genuinely does not know.

Owner view — last 30 days Financials: owner only
Cost per lead
$23.82
Reached by a rep
55%
Median first touch
2m 54s
Median first dial
8h 48m

The dashboard shows the bad number next to the good one by design. Median first dial is displayed in red because it should be — an owner who cannot see their own response time cannot fix it, and a tool that hides it is working for the wrong person.

Finding · raised by the system, not by a client asking

It found eight hours that nobody had ever measured.

On a live account, the nightly audit put two timestamps side by side that it had been collecting since day one: the moment the automated response went out, and the moment a human first picked up the phone.

Automated response — email + text
2m 54s
First human dial
8h 48m
0 2h4h6h8h10h

Both medians come from our own production database, pulled 2026-07-28. They are two separate medians over the same leads, not a per-lead average — which is why we will not quote you one blended "speed to lead" number. Linear scale, 0 to 10 hours.

  1. 01

    Surfaced

    Nobody asked the question. The audit put the two medians next to each other and flagged the distance between them.

  2. 02

    Located

    Not the targeting. Not the leads. The loss sits in the gap between the machine answering and a person answering.

  3. 03

    Explained

    Nothing was telling a rep that a lead had landed. The queue was correct; nobody was watching it. A leak with a cause is a task — without one it is an argument.

  4. 04

    Owned

    Closing it is the top item on our build list, and the number sits on the client's own dashboard while we do it. They watch it move.

Here is the part that matters to you: you have this number too. Almost no operation in this industry can produce it. A company that cannot measure its own response time cannot coach it, cannot price it, and will keep buying more leads to fix a problem that was never in the leads. The first thing we do is put your real number on a screen — and we would rather show you ours than pretend we do not have one.

The rules the readout runs under
01

Cadences change on evidence, not instinct

When we alter a follow-up sequence, the metric, the window and the decision rule are written down before the change goes live — so nobody gets to reinterpret the result after it lands.

02

An unreadable test is reported as unreadable

If the volume is too small for a difference to be visible, the system says so instead of declaring a winner. It has already told us that about one of our own live tests.

03

Every number has to trace to a measurement

Figures in an automated answer are checked back against the stored data they came from. If one cannot be traced, the answer is discarded rather than patched.

Selective intake · limited roster

Let's find out what your market actually costs.

Fifteen minutes. We look at your trade, your territory and your capacity, and tell you what we think a lead and a booked appointment cost there, and what number we would be willing to put in the agreement. If we are not the right fit, we will say so on the call.

Map my market — 15 min Compare the engagements
No long lock-ins Your ad account, your data We say no when it isn't a fit