The cost of a millisecond

The number, and the gap

Take only the publicly auditable, monetised estimates and surprisingly little survives. The cleanest slowdown experiment is Bing: −0.6% revenue per +100 ms. The cleanest peer-reviewed quasi-experiment is Gallino, Karacaoglu & Moreno (2023): ≈ −0.70% conversion per +100 ms at their documented ~3-second baseline. Those two independent designs land almost on top of each other, which is why the best “reconciled effect per 100 ms” for general commercial use is about 0.65% — not 3%, 5% or 8%.

Three findings carry the audit. Speed has a real causal business effect — that part of the folklore is true, and two independent clean designs agree on its size. The effect is local, not globally linear: it depends on the baseline, the page type, the funnel step and the metric, so a single universal millisecond multiplier is a category error. And the most-marketed headlines are about an order of magnitude too high — not because anyone lied, but because they report local cross-sectional slopes, combined metrics and selectively significant results as if they were causal averages.

What the evidence shows, tier by tier

Experiments — the strongest evidence

Real or near-real experiments deliberately slow a random slice of traffic and watch what happens. Google's 2009 slowdown (Brutlag) added 100–400 ms of server latency to web search and lost 0.2%–0.6% of searches per user; the loss grew with exposure and partly persisted after the delay was removed. That's a demand proxy, not revenue, but it has high internal validity. The monetised counterpart is the Bing slowdown A/B (Kohavi et al., reported in HBR and Trustworthy Online Controlled Experiments): −0.6% revenue per +100 ms. It remains the best publicly citable monetised slowdown experiment in the literature.

Quasi-experiments — the best peer-reviewed evidence

Gallino, Karacaoglu & Moreno's “Need for Speed” (Operations Research, 2023) combines fixed effects and generalised synthetic control across seven apparel brands. A 10% slower site cuts conversion by about 2% and sales by about 4.2%; the supplement reports a conversion elasticity of roughly −0.21, which at the paper's documented ~3-second baseline works out to about −0.70% conversion per +100 ms. The effect is concave and uneven along the journey — stronger near checkout — which is the single most important qualifier in this whole literature.

Older econometric work reinforces the non-linearity. Poggi et al. (Information Systems Frontiers, 2014) show, for an online-travel agency, that the right question isn't “what's the universal average effect?” but “at what threshold does the system tip from tolerating into frustrated?” In their setting a tolerance zone runs from about 3 to 11 seconds; past the ~11-second frustration threshold, each additional second costs roughly 3% of sales. That's not a universal retail multiplier — it's a strong argument against linear universal-ROI calculators.

Observational headlines — useful, but not causal averages

The famous big numbers all come from observational cross-sections, and every one of them reads high for the same structural reasons (next section). The folklore tier is weaker still: Amazon's legendary “100 ms = 1% of sales” traces only to a blog post and an old slide — no paper, no dataset, no method — and Google's 2006 “half a second cost 20% of traffic” lives in conference and blog provenance, never an auditable primary. Historically important; not modern evidence anchors.

The reconciled effect — and why it's local

Average the two clean causal designs and you get the ≈0.65% anchor. But the most important word in this whole audit is local. A constant elasticity produces wildly different per-100 ms numbers depending on where you start: the same −0.21 conversion elasticity that yields 0.70%/100 ms at a 3-second baseline yields about 2.1% at one second and only 0.35% at six. There is no universal millisecond multiplier; there is an elasticity that you have to localise.

The anchor has to be bounded in three directions, or it stops being research and becomes a sales calculator:

1. It's baseline-local

   Already-fast pages have a smaller marginal effect; very slow pages — or sensitive checkout/PDP moments — can have a larger one.

2. It's journey-specific

   Deloitte/Google shows larger effects on product-detail and add-to-basket steps; Gallino finds sensitivity concentrated near the transaction. The browse page and the checkout are not the same elasticity.

3. It's metric-specific

   LCP, INP and CLS are not interchangeable. Conflating “server processing delay,” “full page load,” “LCP” and “INP” into one pseudo-metric is how you get a headline that means nothing.

What this changes in practice

The most important consequence isn't “run more tests” — it's test differently. The money probably isn't in a global site-speed average; it's in the few places where speed actually causes abandonment: the product page, the add-to-cart interaction, checkout, and mobile critical flows. So the right artefact isn't a generic performance backlog — it's a revenue-weighted performance backlog, prioritised by where the elasticity is steepest.

Concretely, that means retiring four habits:

• No more generic “7% per 100 ms” multipliers in sales decks — quote the corridor (0.5–0.8%) and localise it.
• No blending of server delay, full page load, LCP and INP into one pseudo-metric.
• No conclusions drawn from raw cross-sections without local-slope and selection context.
• No decision documents that hide null results.

Anything above the corridor needs an explicit local proof — ideally a real holdout experiment, or at least a credibly identified quasi-experiment on exactly the page type and funnel step you intend to monetise. And once you know where to spend the budget, our sister study “The cost of a kilobyte” measures what each fix is worth in milliseconds, and the free performance budget calculator turns an LCP target into the byte budget that hits it. To see this elasticity applied to a real URL — your measured LCP priced as an honest revenue range — run the free Speed → Revenue calculator.

Method

• Discovery, two ways. Direct search for academic and official primary sources, plus recall over curated collections (WPO Stats) used only as a discovery aid — never as a sampling frame, because it selects for positive results.
• Design grading before numbers. Each record is tiered — experiment, quasi-experiment, observational cross-section, before/after case study, or folklore — and tagged with provenance, conflict of interest and auditability, before any effect size is trusted.
• Standardisation only where the design supports it. Monetised effects are placed on a common per-100 ms scale. Where a paper reports a log-log elasticity (e.g. Gallino's −0.21), the per-100 ms figure is the elasticity evaluated at the documented baseline; reported per-100 ms headlines are taken as published. Effects without an absolute millisecond delta (most CWV-era case studies) are listed but explicitly not standardised.
• Triangulation, not pooling. No inverse-variance meta-analysis: most sources publish no standard errors or microdata, so a pooled average would be false precision. The reconciled anchor is the agreement of the two cleanest independent causal designs (Bing, Gallino), not a weighted mean over the whole register.
• Open data. The full register and the standardised file are published below under CC BY 4.0; every row carries its source link so you can re-grade it yourself.

Data & sources

The study register and the standardised per-100 ms file are published under CC BY 4.0 — re-grade the studies, disagree with our tiers, or extend the register with new evidence.

Key sources

• Brutlag, J. (2009). Speed Matters for Google Web Search. Google.
• Kohavi, R., Henne, R. & others (2017). The Surprising Power of Online Experiments. Harvard Business Review.
• Kohavi, R., Tang, D. & Xu, Y. (2020). ‘Speed Matters’, in Trustworthy Online Controlled Experiments. Cambridge University Press.
• Gallino, S., Karacaoglu, N. & Moreno, A. (2023). ‘Need for Speed: The Impact of In-Process Delays on Customer Behavior in Online Retail’. Operations Research, 71(3), 876–894.
• Poggi, N. et al. (2014). ‘A methodology for the evaluation of high response time on E-commerce users and sales’. Information Systems Frontiers, 16(5), 867–885.
• Akamai (2017). State of Online Retail Performance.
• Deloitte (2020). Milliseconds Make Millions. Commissioned by Google.
• Shopify (2026). Store Speed and Conversion: What the Data Shows.

Cite this audit

PageSpeedAudit Research (2026). The cost of a millisecond: an evidence audit of
the pagespeed-to-revenue literature, graded by research design.
https://pagespeedaudit.com/research/the-cost-of-a-millisecond

@misc{pagespeedaudit2026millisecond,
  title  = {The cost of a millisecond: an evidence audit of the
            pagespeed-to-revenue literature},
  author = {{PageSpeedAudit Research}},
  year   = {2026},
  url    = {https://pagespeedaudit.com/research/the-cost-of-a-millisecond},
  note   = {Study register: speed-revenue-study-register.csv, CC BY 4.0}
}