DIGITARISE

Ecommerce mostly refused to be measured.

We sampled 38 US retail and direct-to-consumer homepages. 17 of them — 45% — returned an error to a plain HTTPS request. That is nearly three times the refusal rate of the hospital and law-firm samples, and it is not random: almost every blocked site is a large established retailer, while almost every site that answered is a digital-native brand.

Read this before the numbers

sampling limit
What this sample can and cannot say

Because the refusals cluster on enterprise retailers — Macy’s, Williams Sonoma, J.Crew, Lululemon, Wayfair, REI, Patagonia and others — the 21 homepages that remain are not a representative sample of “ecommerce”. They are a reasonable sample of digital-native DTC brands, and that is how every figure below is labelled. Treating them as sector-wide would be selection bias, so we do not.

Refusal is itself a measurement

all sectors

How often a sector’s homepages refuse a non-browser client is a real, comparable property — of bot defence, of CDN configuration, and of how reachable that sector is to any automated agent, including the crawlers behind search and AI answers. Select the first metric to see it.

Share of the sampled homepages that returned an error to a plain HTTPS request. A measured property of the sector — and the bound on how representative the rest of these figures can be.

Healthcaren = 3517%
Legaln = 3816%
DTC ecommercen = 3845%

Legal holds the best posture here — 16% against 17% and 45%.

Lower is better · all sectors measured with the same script and method · 2026-07-20

Instrument — every measured sector, one metric at a time.

The DTC brands that answered

n = 21
227 msmedian server response

Half of the 21 measured homepages return their first byte within this time; the slowest takes 1219 ms.

858 KBmedian HTML delivered

Size of the delivered document before any subresource loads. 20 of 21 serve it compressed.

86%ship a known tracker

18 of 21 homepages reference a recognised analytics or advertising tracker in the delivered HTML.

15median third-party domains

The median homepage references this many distinct external domains before any interaction.

67%surface a consent mechanism

Only 14 of 21 expose a detectable consent platform — yet 21 set a cookie before any choice is possible.

86%expose a main landmark

18 of 21 provide a main landmark for assistive technology; 9 ship no first-level heading at all.

Explore every row

live explorer
Median trackers0
≥1 tracker14/29
Any consent signal7/29
  1. 01Northwestern Medicine41/1003rd-party25trackers3pre-cookies2
  2. 02Stanford Health Care53/1003rd-party16trackers2pre-cookies4
  3. 03UPMC54/1003rd-party18trackers1pre-cookies9
  4. 04Banner Health57/1003rd-party16trackers1pre-cookies4
  5. 05Cleveland Clinic59/1003rd-party16trackers1pre-cookies8
  6. 06Ochsner Health61/1003rd-party16trackers2pre-cookies0
  7. 07Boston Children's Hospital61/1003rd-party14trackers0pre-cookies5
  8. 08Nationwide Children's Hospital63/1003rd-party9trackers2pre-cookies4
  9. 09Vanderbilt University Medical Center65/1003rd-party16trackers1pre-cookies0
  10. 10Tampa General Hospital67/1003rd-party13trackers0pre-cookies3
  11. 11Massachusetts General Hospital68/1003rd-party11trackers1pre-cookies2
  12. 12UCLA Health70/1003rd-party11trackers1pre-cookies1
  13. 13Emory Healthcare70/1003rd-party12trackers0pre-cookies2
  14. 14Houston Methodist71/1003rd-party13trackers1pre-cookies4
  15. 15Baptist Health72/1003rd-party9trackers0pre-cookies3
  16. 16Mass General Brigham72/1003rd-party12trackers0pre-cookies1
  17. 17Brigham and Women's Hospital73/1003rd-party10trackers0pre-cookies2
  18. 18Advocate Health73/1003rd-party13trackers0pre-cookies0
  19. 19NYU Langone Health78/1003rd-party9trackers0pre-cookies0
  20. 20Sutter Health79/1003rd-party7trackers0pre-cookies1
  21. 21MD Anderson Cancer Center79/1003rd-party8trackers0pre-cookies0
  22. 22Intermountain Health80/1003rd-party3trackers1pre-cookies1
  23. 23NewYork-Presbyterian82/1003rd-party6trackers0pre-cookies0
  24. 24Cedars-Sinai83/1003rd-party2trackers0pre-cookies2
  25. 25Memorial Sloan Kettering83/1003rd-party10trackers1pre-cookies0
  26. 26Duke Health84/1003rd-party9trackers1pre-cookies0
  27. 27UCSF Health85/1003rd-party8trackers0pre-cookies2
  28. 28Rush University Medical Center85/1003rd-party11trackers0pre-cookies0
  29. 29Penn Medicine86/1003rd-party0trackers0pre-cookies2

29 of 35 US hospitals & health systems homepages · measured 2026-07-20 · score = transparent composite of the raw signals in each row

Switch the dimension to re-rank the sample. Download the dataset (JSON) →

Every refusal, named

17 of 38
Patagonia — HTTP 404 (automated request blocked)REI — connection errorWarby Parker — HTTP 403 (automated request blocked)Chewy — HTTP 429 (automated request blocked)Wayfair — HTTP 429 (automated request blocked)Macy's — HTTP 403 (automated request blocked)Crate & Barrel — HTTP 403 (automated request blocked)West Elm — HTTP 403 (automated request blocked)Williams Sonoma — HTTP 403 (automated request blocked)Anthropologie — HTTP 403 (automated request blocked)Urban Outfitters — HTTP 403 (automated request blocked)Madewell — HTTP 403 (automated request blocked)J.Crew — HTTP 403 (automated request blocked)Lululemon — HTTP 403 (automated request blocked)Fabletics — HTTP 403 (automated request blocked)Hims — HTTP 403 (automated request blocked)Solo Brands — HTTP 403 (automated request blocked)

A sample that lost 45% of its subjects is reported as such, not quietly rescaled.

Read the methodology