Assume $500 bought 100 clicks. Each dot below represents one click inside a deliberately severe scenario — not a measured visitor and not an industry benchmark.

The curve is visible so you can question it. Replace it with reached-content and call events from your own analytics before making a business decision.

SCENARIO · 100 PAID CLICKS READY
100
ASSUMED TO REACH
0.0s
LOADING
$500
ALREADY SPENT
EACH SQUARE IS ONE PERSON YOU PAID FOR

SEVERE ASSUMPTION

17

of 100 reach the content in this model.

STRONG ASSUMPTION

~90

of 100 reach the content in this model.

What the model does not know

It does not know why anyone left, whether the page was visible, whether the click was qualified, or whether a caller became a customer. Those require instrumentation, not copywriting.

STEP 1100
STEP 287
STEP 368
STEP 447
STEP 624
STEP 817
The useful part is the equation: wasted-click cost = paid clicks that fail your defined reach event × cost per click. The model supplies no real inputs.

What this costs in real money

At $5 a click, 100 clicks cost $500. If your analytics record C calls, cost per call is $500 ÷ C. No call count belongs in that equation until it was observed.

A faster page can improve the chance that content becomes usable; it cannot establish a new call count by itself. Compare equivalent traffic before and after, then keep speed, reach, calls, bookings and revenue as separate measurements.

What was actually measured

In an August 2026 private sample of 100 contractor homepages, 69 were fully weighed. Their median was 2 MB; 30 of the 100 crossed a three-second estimate. This is a descriptive sample, not a representative market benchmark, and it contains no conversion data.

None of those owners knew. Their sites load fine on the office wifi, on a desktop, with everything cached. The customer standing in a driveway on one bar of LTE has a completely different experience — and that customer is the one holding the credit card.

Check yours right now

  • Open your site on your phone with wifi turned off
  • Count out loud until you can tap something
  • If it feels slow, measure it and compare the result with your own reach and call events

That's the whole test. No tools, no consultant, thirty seconds.