Funnel research

Prepared for

The evidence

Part two of three. Sources, counts and confidence.

Nine research files, 982 kilobytes of raw material, gathered on 8 September 2026. This page says what was read, how it was checked, and how far each figure can be leaned on. It also lists the six claims that were found during research and then dropped, because a report that only shows what survived is not an evidence report.

The PAHI product page at the fold, 8 September 2026. No review count, no shipping promise, no guarantee. The review block sits 3,468 pixels further down.
01

What was read

37,727

Reviews, posts and comments in the three largest corpora alone

A2, A3 and A4

~700

Live competitor ad cards, read in the Meta Ad Library

21 searches, 11 brands

1,011

Source URLs cited across the nine research files

700 distinct pages, across 70 hosts

32

Screenshots taken while walking competitor funnels

Six brands, end to end

700 distinct source pages is what the nine files cite between them, spread across 70 hosts. The raw figure is 1,011 URLs, but review feeds are cited one page of results at a time, so collapsing those back to the document they page through is the honest count. Both numbers can be recounted from the files, and the report's own test suite recounts them on every run.

Nine workstreams. Every quote count is the researcher's own, machine-checked.
Workstream Corpus Quotes URLs How it was checked
A0
PAHI's own reviews 		32 rows in the Judge.me export 18 genuine reviews quoted in full n/a, export Sender domains and originating locations, review by review
A1
PAHI off-site footprint, two customer videos 		All reachable PAHI pages, profiles and listings 75 quotes 82 Rendered pages or raw HTML only. Search summaries rejected as sources
A2
Orthofeet, Propet, Dr. Comfort, Apex 		12,010 items. 7,143 brand and retail reviews, 4,867 Reddit 308 fragments checked 209 Machine compare against the stored corpus. 0 failures. 10 re-fetched live
A3
Kizik, Skechers Slip-ins, Vionic, OOFOS, HOKA 		12,109 items across six sources 419 quoted strings 122 Quotes emitted from the stored corpus by script, never retyped. 0 failed verbatim
A4
Hoka, Brooks, New Balance, ASICS, Ryka and the cross-shopped set 		13,608 items. 10,198 Zappos, 671 Amazon, 2,739 Reddit 562 quote strings 193 Byte-for-byte against the harvest, 0 misses, plus 5 live spot-checks
A5
Competitor advertising, Meta Ad Library 		About 700 live ad cards, 21 searches, 12 landing pages 343 candidates, 331 verbatim-matched 187 247 cited Library IDs all present in the captured data. 12 residuals each explained
A6
Caregiver and patient communities 		About 3,000 comments. 60 Reddit threads, 11 AgingCare threads, 5 videos 443 quote fragments 187 443 of 443 matched back to the raw harvest
B1
Competitor funnels, walked in a real browser 		6 of 11 brands walked end to end. 32 screenshots Copy read off the rendered page live walk Counts read off the rendered page. Fields not displayed written as not shown
B2
Category benchmarks 		18 primary publishers reached and read Figures, not quotes 31 Derived figures recomputed. Sources graded panel, survey, aggregate, vendor

Four of the nine workstreams ran a machine verifier over their own finished file before delivering it. Between them those checks caught a substituted pronoun, two dropped leading words, five capitalisation drifts inside quote marks, two paraphrases wrongly sitting in quote marks, one invented example testimonial, and one line that came from a search-engine summary rather than a page anyone had read. All were removed or corrected before delivery. The same discipline is applied to this report: every quote on these pages is re-checked against its named research file each time the report is built.

02

PAHI's own reviews, split

The Judge.me export holds 32 published rows. They are not all customers, and the split matters more than the total, because 13 of them have to come down and one has to be resolved before any traffic is bought.

18

Genuine customers. US locations, real domains, specific detail

All five stars

13

Written by the outgoing development agency

Agency sender domains, Indian originating locations

1

Written by an employee, undisclosed

Company email domain

12

Separate colourway products the 18 genuine reviews are stranded across

One to three each

Why the product page still says be the first to write a review

The 18 genuine reviews sit on 12 of 26 colourway products. The other 14 hold none, and the product page audited on 8 September was one of those. The catalogue problem causes the reviews problem. Merging 26 products into 3, with colour as a variant, lands 18 reviews on 3 pages and requires no new content from anyone.

A further 7 review records point at product handles that were renamed. They are orphaned. All 7 are agency reviews, so nothing genuine was lost, but it confirms that review records do not follow a rename. That matters before the merge, not after.

The two that must never become copy

Two published reviews describe not falling. They may remain displayed as customer reviews. They must never become a headline, an advertisement, or brand copy, because they state fall prevention and PAHI's own disclaimer forbids it. This one is included here so the boundary is concrete rather than abstract.

I have been falling recently from my battle with Parkinson’s. I ordered the shoes and have worn them for 2weeks and have not had a fall yet. The recommend this shoe fits anyone who is having balance problem.
Customer with Parkinson’s · Judge.me, pahiusa.com · five stars. Displayable as a review, never usable as a claim

The lines that can be used describe how it feels rather than what it prevents. Both of these are already public, and both sit on the safe side of the line.

I have MS and these are great for balance and stability. I have been wearing them everywhere and I don't have to worry with every step. The purple color is fabulous. I would highly recommend!!
Jennifer F., customer with MS · Judge.me, pahiusa.com · five stars
I was able to walk straight rather than walking like I I had consumed too much alcohol.
Customer with MS, on video · YouTube, machine-generated captions · 00:30

On the video quote. It is taken from a machine-generated caption track, which is why it reads with a doubled word. The words are the customer's. Name spellings from those caption tracks are unreliable and are not used anywhere.

Fit is the question people actually ask

Before buying, the questions PAHI receives in public are about width and sizing. After buying, one customer reveals guidance that exists in support conversations and appears nowhere on the site.

How wide do they come?
Susan Glover McGinnis, prospect · Facebook, Alec’s Shoe Store
I will look into this, especially since my wide feet overpronate.
Mark Gonzales, prospect · LinkedIn
I ordered a 1/2 size down as recommended and they fit great.
Jennifer F., on the B-Steady Slide · Judge.me, pahiusa.com · five stars
03

The competitor set

Six brands walked end to end in a real browser on 8 September, home page to first checkout screen, no purchase completed anywhere. Five more were queued and not walked, so this is a bounded sample and is described as one. Element positions were measured against the same fold on every site.

Brand Sitewide offer Return window Reviews on the page walked Fit tooling
PAHI None shown anywhere on the storefront 30 days from delivery, reason-gated 0 on the product page walked Size guide button only
Orthofeet Free Shipping, Returns and 60-Day Wear Test, in the top bar of every page 60-day wear test 17,585, averaging 4.2, above the price Fit: True to size, plus a size and fit chart
Kizik Risk free trial, free shipping and easy returns, top bar 30 days, not worn outside 8,701, above the price Three-point fit meter plus review-derived width and size distributions
Vionic Free shipping over $100 30 days, branded as a 30-Day Challenge 603, averaging 4.4, above the price Fit guide plus a Bazaarvoice size rating
Dr. Comfort Free shipping over $75 and free returns, plus a 15% signup coupon 30 days from ship date, unworn, plus a 6-month warranty None on the product page walked Size chart plus a Shoe Advisor quiz
Propet An extra 10% off, in a popup that fires on load and again at checkout 14 days, unworn only 125, averaging 4.8, above the price Size chart link

What to look at. No review count, no shipping promise, no guarantee. Size and width sit below a size guide link with no fit signal beside them.

pahiusa.com, walked 2026-09-08

What to look at. Free Shipping, Returns and 60-Day Wear Test in the top bar. A 4.2 rating and 17,585 reviews above the price. Fit: True to size beside the size grid.

orthofeet.com, walked 2026-09-08

Every brand walked puts its review count above the price. PAHI's review block sits 3,468 pixels down the page and says be the first to write a review. Shipping goes in the top bar on all six, three of them behind a threshold PAHI does not need. PAHI ships everything free and says so only in the cart drawer, after the decision has been made.

What roughly 700 competitor ads do not contain

The ad sweep covered 11 brands and 10 discovery keywords in the Meta Ad Library. Of the cited advertisements, 247 Library IDs were checked back against the captured data and all 247 were present. Three absences run through the whole sample.

  • No demonstration: no before and after, no side by side, nothing showing the product doing what it claims.
  • No fit quiz: not one, in about 700 cards, where the landing pages are product page, collection, advertorial or lead form.
  • Nothing about the mechanism: every brand sells foam, arch support or hands-free entry, and none of them sells the shape of the base.

These are absences within a sample, not proof of absence in the category. They are stated that way deliberately. The strongest observed offer construct in the sweep is a 60-day wear test, which appears in more live creatives than any other offer line, and which PAHI cannot run because it refuses worn returns. Discounts of 45% to 70% cluster on dropship advertisers.

04

Category benchmarks

There is no publisher of footwear-only funnel benchmarks. Where footwear is not broken out, the row uses Apparel and Accessories and says so. The grade in the last column is what stands behind the number, not how much we like it.

Metric Figure Source and sample Grade
Order value, Apparel and Accessories, Meta-driven orders $86.27 median Triple Whale, 40,000+ brands, Aug 2025 to Jul 2026 Panel
PAHI's B-Active Mesh against that median 2.09x $180 against $86.27. Arithmetic, recomputed Derived
Conversion rate, all paid channels, Apparel and Accessories 1.80% median Triple Whale, 53,000+ brands Panel
Add-to-cart rate, Apparel 6.33% to 7.12% Triple Whale Panel
Cart abandonment, all ecommerce 70.22% Baymard Institute, pooled from 50 studies Aggregate
Return reasons in footwear that are fit or size 61% Loop Returns, 23.4m returns, 4,000+ Shopify merchants Vendor
Shopify merchants now charging a return fee 65.2%, averaging $9.04 Loop Returns, Nov 2024 to Oct 2025 Vendor
Purchase likelihood at five reviews against none 270%, and 380% on higher-priced items Spiegel Research Center, 2017. No sample published Survey
Meta CPM premium, Health and Wellness against Apparel +64.5%, with ROAS 35.7% lower Triple Whale. Used here as a proxy, not as a footwear figure Panel
Online return rate, all US ecommerce 19.3% of online sales NRF and Happy Returns, 2,006 consumers and 358 professionals Survey

Panel means the publisher measured it on their own platform and states a sample. Survey means they fielded one and state a sample. Aggregate means they pooled other people's studies and say which. Vendor means a self-published benchmark with no disclosed method. Derived means we did the arithmetic, and the arithmetic was recomputed.

05

How far to lean on each figure

If a number in the strategy is challenged in the room, this is the answer.

Measured directly

From PAHI's own systems and pages, on 8 September 2026.

  • 18 genuine customer reviews, all five stars, and their distribution across 12 colourway products. Judge.me export.
  • 13 reviews written by the outgoing agency and one undisclosed employee review. Sender domains and originating locations.
  • 26 products for 3 styles, 551 variants, one sentence of copy per product.
  • Analytics and advertising tags firing only after consent is granted.
  • Two customer video testimonials, transcribed with timestamps.
  • The returns portal. The policy page itself loads and renders. Its own call to action leads to an empty page.
Verified independently

Confirmed against a second source that publishes it.

  • A retailer listing the same shoe at $175 against PAHI's $180. PAHI has since asked that retailer to take the listing down.
Confirmed by PAHI, not checked by us

Answered by PAHI on 8 September. Strong, and the artwork is what to ask for next.

  • Three independent credentials. The APMA Seal of Acceptance, on all products. ASTM slip resistance, which PAHI permits stated as ASTM approved. Innovation of the Year 2026, from the Eldercare Innovation Awards. Together they are the strongest proof the brand holds that does not depend on collecting more reviews first, and none of them needs a person to front it.
  • A patented outrigger design, and laboratory testing that exists and is being sent. The patent is written as patented and never with a number, which is PAHI's own house rule. The laboratory data has not been received, so nothing in this report rests on it. When it lands it is the strongest asset the brand owns.
  • Worn shoes cannot be returned, against PAHI's own criteria for what counts as worn. That is why no wear test appears anywhere in this report.
  • An anchor price of $225 against the $180 sold today. From HM3's strategy deck. Every price figure in this report is the $180 read off the live product page.

This research did not reach the credentials itself, because the sweep covered reviews, advertising, funnels and published benchmarks rather than certification registers. They are recorded here as PAHI's own answers, which is a different thing from a figure we measured, and the distinction is kept because the rest of this page keeps it.

Counted by us

Medium to high confidence. Bounded by the sample we read.

  • About 700 competitor advertisement cards reviewed. The absence claims drawn from them are absences within that sample rather than proof of absence in the category.
  • No competitor ran a fit finder or quiz within that sample.
  • Word frequencies across the corpora, including 102 uses of balance against 26 of one leg, and zero uses of outrigger, catamaran or base of support in 13,608 items.
Published research, directional

Real sources with real limits. Named, so the limits travel with the number.

  • Reviews. The purchase-likelihood gap between no reviews and some reviews, and the plateau at five. Spiegel Research Center. Strong direction, weak precision: the study is from 2017 and publishes no sample size. It remains the study the industry cites because it is the only one housed in a university, and the action it implies costs almost nothing.
  • Returns. Fit and size driving 61% of footwear returns. Loop Returns, published 26 February 2026. Vendor published, sample not disclosed. Directionally consistent with fit being the most common pre-purchase question in PAHI's own corpus.
  • Order value and acquisition cost. Triple Whale panel data. Vendor published. Used here only for the ratio argument, which holds even if the absolute figures move substantially.
Our own constraint, not a benchmark

The 12,000 visitors per variant figure comes from our standard experiment methodology, not from an outside source. It is the sample needed to detect a 20% improvement on a 3% baseline at ordinary confidence. It is arithmetic, and it is the reason month one is sequenced rather than tested.

06

What was rejected

Six claims were found during research and then set aside because they could not be confirmed at the named source. None of them appears anywhere in the strategy. They are listed because knowing what was thrown away is how you judge what was kept.

Attributed, not confirmed
Footwear gross return rate of 19.06% across 336 brands
Attributed to: A Loop Returns benchmark report
Cited by a third-party blog. Not found on any Loop page we could read.
Checkout redesign yields up to 35.26% conversion uplift
Attributed to: Baymard Institute
Repeated across the industry. Absent from the Baymard page we read.
About 66% of adults measure an E width against a D-width standard, from 50m+ scans
Attributed to: Aetrex, via Footwear News and WWD
The article is paywalled. The quote was visible only in a search index, never on the page. If it holds it is the single most useful category fact for PAHI, which is why it is worth buying rather than guessing.
Global diabetic footwear market of $8.96bn in 2024, growing 7.7% a year
Attributed to: Grand View Research
Not confirmed on the Grand View page we read.
Meta removed sensitive interest targeting categories in January 2022
Attributed to: Meta
Second-hand only. Not found in Meta's own change logs.
Meta removed detailed-targeting exclusions on 31 March 2025
Attributed to: Meta
Second-hand only. Not found in Meta's own change logs.

A separate list, longer than this one, records the figures that do not exist at all from any credible publisher: a footwear-only conversion rate, a footwear order value benchmark, a US footwear return rate with a stated sample, a footwear repeat purchase rate, a wide-width market size. What circulates in their place is SEO content restating ranges that trace to nobody. If one of those becomes load-bearing, it has to be bought or measured rather than quoted.

One correction to an earlier working note belongs here, and it is included because a report that quietly drops a corrected claim is harder to trust than one that shows the correction. It was recorded that PAHI's footer returns link led to a page that could not be found. Walked live on 8 September, that page loads and renders correctly. What is broken is one step further in: the page's own Start a Return or Exchange button leads to an empty page with no portal behind it. The fix is an application, not a URL.

What to look at. This is where the documented return path ends. The page responds, and there is nothing on it. An unsatisfactory returns policy is the seventh most common reason people give for abandoning a checkout, named by 13 percent of them. This is what PAHI's return promise resolves to today.

pahiusa.com/apps/returns, walked 2026-09-08

07

What could not be reached

Stated rather than papered over, because each of these is a known edge on the sample.