How to estimate conversion loss from site speed

You have a load time. You want a dollar figure. Here's the honest path between them — the one that survives a skeptic asking "where did that number come from?"

Step 1 — get a measured load time

Not a guess, not a screenshot impression. Use real lab/field data (Google's Lighthouse / PageSpeed powers this): Largest Contentful Paint, Time to Interactive, and Cumulative Layout Shift together describe how quickly — and how stably — the page becomes usable. Mobile is what matters; that's where the traffic and the impatience are.

Step 2 — map speed to a conversion rate

You don't need a bespoke study. Use a defensible pair of anchors:

• A fast site (~2s) converts at roughly your industry's healthy benchmark.
• A slow site (~5s+) sits meaningfully below it.

The gap between those two rates is your estimated speed-driven conversion loss. Keep the anchors conservative and state them out loud — that's what makes the estimate defensible instead of magical.

Step 3 — turn the rate gap into lost actions

monthly visits × (healthy CVR − current CVR) = lost conversions/month

Example: 2,000 visits, healthy CVR 4%, current (slow) CVR 2.5% → 2,000 × 1.5% = 30 lost conversions/month.

Step 4 — turn lost actions into lost revenue

lost conversions × close rate × value of one win

If 60% of those become customers at $250 each: 30 × 0.6 × $250 ≈ $4,500/month.

Step 5 — state it as a range, with the inputs

Never a single number. Present it as $3,500–$5,500/month, and list the four assumptions (visits, CVR anchors, close rate, win value). A skeptic who can see and adjust your inputs will trust the output. A skeptic handed one confident figure with hidden math will (correctly) dismiss it.

The honesty checks

• Don't double-count across multiple issues — the same visitor can't be lost twice. Take the biggest issue in full and a fraction of the rest.
• Don't invent a loss for a brochure site that can't convert at all — measure the opportunity of adding a conversion path instead.
• Don't hide the assumptions. The visible math is the credibility.

Let the tool do the arithmetic

Every step here is mechanical once you have the measured signals — which is exactly what an audit automates: it pulls the real speed data, applies your industry's benchmarks, and prints the conversion-loss range with the assumptions attached.

Run a free audit and get this calculation done on your real numbers in about two minutes.