WordPress Hosting for Agencies

Live benchmark · Updated 2026-06-08

2026 WordPress Hosting TTFB Benchmark

A machine-readable, weekly-refreshed time-to-first-byte (TTFB) benchmark across 10 major managed WordPress hosts. Curl-based sampling, 5 samples/day per host, median + p90 computed over a rolling 7-day window. Measured from a São Paulo VPS for a realistic "visitor outside North America" view.

↓ Download CSV dataset Methodology

Review Signal's WPPerformance benchmarks stopped updating in 2020 and Hostingstep's annual tests don't publish raw data. This is the first public, weekly-refreshed, machine-readable TTFB benchmark for managed WordPress hosts in the 2022+ era. License CC BY 4.0 — cite this dataset freely.

— WordPress Hosting for Agencies benchmark methodology

Rolling 7-day TTFB ranking

Window: 2026-06-02 → 2026-06-08. Each host sampled 5 times/day (median taken per day) across the window. Lower is better.

# Host Median TTFB p90 TTFB Samples (OK / total) Origins URL
1 pressable 48 ms 52 ms 21 / 21 2 pressable.com
2 nexcess 48 ms 75 ms 21 / 21 2 www.nexcess.net
3 pantheon 51 ms 68 ms 21 / 21 2 pantheon.io
4 cloudways 52 ms 85 ms 21 / 21 2 www.cloudways.com
5 dreamhost 52 ms 92 ms 21 / 21 2 www.dreamhost.com
6 bluehost 65 ms 179 ms 21 / 21 2 www.bluehost.com
7 wordpress-com 88 ms 95 ms 21 / 21 2 wordpress.com
8 flywheel 92 ms 170 ms 21 / 21 2 getflywheel.com
9 kinsta 209 ms 350 ms 21 / 21 2 kinsta.com
10 wp-engine 314 ms 399 ms 21 / 21 2 wpengine.com

Methodology

What we measure: TTFB (time-to-first-byte) — the time from a client sending a GET request to receiving the first byte of the HTTP response body. Measured via curl -w '%{time_starttransfer}', following redirects, with a 15-second timeout cap.

Which URL per host: each host's own public marketing/blog site. These run on the host's own infrastructure and are a reasonable proxy for "what a well-tuned customer site CAN look like" — not necessarily "what the average customer experiences." We publish the exact URL tested per host in the dataset so you can audit the choice.

Sampling: 5 samples per URL per day, median of those 5 stored. Rolling 7-day window computes median and 90th-percentile of the daily medians. Failed samples (timeouts, 5xx, DNS) are excluded before median computation; sample-count columns expose OK vs total so you can detect flakiness.

Origins (multi-region as of 2026-04-24): benchmark now samples from TWO geographic origins — São Paulo (Hostinger VPS) and Canada (bbnlabs5 server) — daily at 06:00 UTC. The Origins column on the table above shows how many regions saw each host in the rolling 7-day window. Multi-region samples expose hosts whose CDN/edge cache strength varies dramatically by visitor geography (we observed Kinsta at 270 ms from São Paulo vs 81 ms from Canada — a 3.3× spread the single-origin benchmark would have hidden). More origins are added as they come online; results aggregate automatically.

What this test does NOT measure: Lighthouse score, LCP, CLS, rendering time, JavaScript execution, TLS handshake (pre-included in TTFB). It does NOT test from multiple geographies simultaneously. It does NOT test a standardized WP-Bench install on each host (that was Review Signal's approach; more rigorous, much more expensive).

Update cadence: daily collection at 06:00 UTC. Rolling 7-day CSV regenerated on every run. Full runner is on GitHub (opening the benchmark pipeline is on the roadmap).

License: CC BY 4.0. Cite as: "2026 WordPress Hosting TTFB Benchmark — 10 Hosts, Weekly Dataset, WordPress Hosting for Agencies, 2026-06-08. Available at https://thebrownbrick.com/2026-wp-hosting-ttfb-benchmark."

Related resources

Per-host deep-dives

Or pick by your agency's home city

Get the monthly roundup

One email per month. New comparisons, price changes, and the rare deal worth knowing about. No spam, unsubscribe in one click.