Report scope. This report presents the validated manual-audit findings together with supporting PageSpeed indicators for mobile and desktop performance, automated mobile accessibility, Largest Contentful Paint (LCP) and First Contentful Paint (FCP). The patient-journey audit remains the primary analysis.
Declaration of interest. iGlowly develops patient-information software for aesthetic-clinic websites, including concern-led navigation and question-first support. This creates a commercial interest in some of the topics examined. The study used documented audit rules, consistent task conditions, explicit denominators and transparent derived measures. A de-identified dataset with clinic and domain identifiers removed, together with the codebook and calculation tables, accompanies the report so the published aggregate results can be reproduced.
Data and reproducibility
The de-identified record-level data, coding rules and published calculation tables are available for review and recalculation:
This study examined 200 patient-facing aesthetic-clinic websites to understand what a prospective patient could find, understand and do before contacting a clinic. Each task was completed from the website alone, outside normal clinic hours and within a three-minute limit. The audit did not assess clinical quality, rank clinics or test what happened after an enquiry was submitted.
The unit of analysis was the website or domain—not necessarily a unique legal business. All 200 canonical domains were unique, but separate patient-facing domains operated by related businesses or sister brands were retained because each presented a distinct public journey. Findings therefore describe the websites audited.
The central finding is not that clinic websites lacked ways to make contact. Almost all did. The more important gap appeared earlier in the journey: information, navigation and action were not always connected. A visitor could often find a booking button, form or phone number before finding a complete route from their concern to relevant options, the audited information about the selected service, or a clearly linked practitioner.
This pattern differed by clinic type. For non-surgical clinics, the strongest gap was concern-led navigation. Most listed treatments, but relatively few helped a visitor translate a concern such as wrinkles or facial ageing into relevant options. Surgical and mixed websites were generally better aligned with a consultation-first pathway, but often made the consultation request easier to find than the consultation price or process.
The supporting technical analysis identified a pronounced mobile constraint. Desktop performance was higher for 183/199 comparable websites (92.0%), while only 12/200 websites reached a green mobile performance score of 90–100. In PageSpeed mobile lab testing, 168/200 websites (84.0%) recorded LCP in the poor band above four seconds and 141/199 (70.9%) recorded FCP in the poor band above three seconds. These are lab-test indicators, not exact patient waiting times.
Findings at a glance
Patient-journey measure
All websites n=200
Surgical + mixed n=80
Non-surgical n=120
Concern-led route found
93 (46.5%)
63 (78.8%)
30 (25.0%)
All four core service-information elements
77 (38.5%)
48 (60.0%)
29 (24.2%)
Neither risks nor suitability information found
96 (48.0%)
19 (23.8%)
77 (64.2%)
Any selected-page call to action
170 (85.0%)
70 (87.5%)
100 (83.3%)
Treatment price visible directly on website
99 (49.5%)
24 (30.0%)
75 (62.5%)
Price visible only after entering booking flow
30 (15.0%)
0 (0.0%)
30 (25.0%)
No treatment price found
71 (35.5%)
56 (70.0%)
15 (12.5%)
Consultation-request pathway
99 (49.5%)
67 (83.8%)
32 (26.7%)
Named + service-linked practitioner + credential shown
107 (53.5%)
59 (73.8%)
48 (40.0%)
Any contact channel
196 (98.0%)
80 (100.0%)
116 (96.7%)
Strict question-first task possible without identification
6 (3.0%)
1 (1.2%)
5 (4.2%)
The four core service-information elements were: an explanation, downtime or recovery, suitability or contraindication information, and any risk or side-effect wording. The risk measure recorded presence only; it did not assess completeness or clinical adequacy.
Supporting technical findings at a glance
Technical indicator
Result
Mobile performance score 90–100
12/200 (6.0%)
Desktop performance higher than mobile
183/199 (92.0%)
Mobile LCP above 4 seconds
168/200 (84.0%)
Mobile FCP above 3 seconds
141/199 (70.9%)
Automated mobile accessibility score 90–100
91/200 (45.5%)
Eight principal findings
Treatment catalogues were common, but concern-led guidance was not.
Treatments or procedures were listed by 189/200 websites audited. Only 93/200 helped the visitor move from the tested concern to relevant options. The contrast was greatest among non-surgical websites: 109/120 listed treatments, while 30/120 provided a concern-led route.
The largest website profile was action-enabled but information-incomplete.
A selected-page call to action was found on 170/200 websites, but only 68/200 combined that action with all four core information elements. Another 102/200 offered an action without the complete core set.
For nearly half, neither risk nor suitability information was found for the selected service.
No risk or side-effect wording and no suitability or contraindication information were found on 96/200 websites (48.0%). The pattern appeared on 77/120 non-surgical websites (64.2%) and 19/80 surgical or mixed websites (23.8%). These were presence-only measures and do not assess the clinical adequacy of information that was shown.
Non-surgical sites often exposed price and booking before presenting a complete information pathway.
A treatment price was found for 105/120 non-surgical websites: 75 showed it directly and 30 revealed it only after the visitor entered a booking flow. An online-booking pathway appeared on 83/120, while 29/120 contained all four core information elements for the selected service.
Surgical and mixed websites generally used the appropriate consultation-first model, but consultation explanation was less consistent.
A consultation-request pathway was available on 67/80 surgical or mixed websites. Only 38/80 showed either a consultation price or a description of the first-consultation process, and 34/80 combined consultation information with a request pathway.
Provider visibility weakened as the visitor’s question became more specific.
Using all 200 websites as the headline denominator, 171 named at least one practitioner, 131 connected an exact practitioner to the selected service, and 107 completed the full chain of name, service linkage and a visible professional registration or qualifying credential.
Contact routes were common; a route was not the same as an answer.
At least one phone, email, form, WhatsApp or chat channel was found on 196/200 websites. These channels showed that a visitor could initiate contact, not that the clinic or website would answer immediately outside opening hours. Only 6/200 allowed the strict task of asking a question without first giving identifying contact details.
Most websites fell into the poor mobile loading bands in PageSpeed lab testing.
Mobile LCP was above four seconds for 168/200 websites (84.0%), while mobile FCP was above three seconds for 141/199 websites with a valid FCP result (70.9%). Only 12/200 websites (6.0%) achieved a green overall mobile performance score. These supporting indicators are lab snapshots and do not measure each visitor’s exact waiting time.
What the findings mean
The evidence points to a continuity problem rather than a simple shortage of content or contact options. Many websites contained strong individual components—a treatment menu, a price, a practitioner page, a booking tool or a contact form—but did not always connect those components into one coherent pre-enquiry journey. A form that accepts a message at midnight is available after hours; it is not, by itself, an answer after hours.
The clearest improvement opportunity is therefore not “add more buttons.” It is to help a visitor:
recognise their concern;
identify relevant options;
understand the audited practical, suitability and risk-related information about the selected service;
see who provides it and what credential is shown;
understand likely price or consultation expectations; and
take a next step appropriate to the type of service.
1. Study question and scope
The study asked a deliberately practical question:
What can a prospective patient find, understand and do on an aesthetic clinic’s public website before making contact, particularly when staff cannot be assumed to be available?
This is narrower than an assessment of the clinic itself. The audit unit was the patient-facing website or domain. The final dataset contained 200 unique canonical domains. It was not designed to establish whether every domain represented a different legal business: related businesses, sister brands or the same practitioner could operate separate patient-facing websites. Those domains remained separate audit units because a visitor encountered each as a distinct website journey.
The study was passive and website-only. No account was created, no personal details were entered, no clinic was contacted and no form, booking, WhatsApp message or enquiry was submitted. Each information-finding task had a three-minute limit. The timer continued through slow loading, cookie banners, pop-ups, navigation and page transitions because those features formed part of the visitor’s experience.
The results describe the websites audited, not all aesthetic clinics in the UK and not necessarily 200 independent businesses. The sample was sourced from Google Maps listings using Apify, filtered for relevance and checked for a live URL before website assessment. It is not presented as statistically representative.
Sample profile
Clinic category
Websites
Share
Surgical
50
25.0%
Mixed
30
15.0%
Surgical + mixed
80
40.0%
Non-surgical
120
60.0%
Total
200
100.0%
The main comparison throughout the report is between 80 surgical or mixed websites and 120 non-surgical websites. Surgical and mixed clinics were combined because their principal pathway measures were sufficiently similar for this descriptive analysis, particularly price visibility, consultation-request pathways and selected-page actions. A sensitivity table separating them is included in the appendix.
Geographic profile
Recorded location
Clinics
Share
London
76
38.0%
Manchester
57
28.5%
Liverpool
42
21.0%
Other recorded locations
21
10.5%
City not stated
4
2.0%
London, Manchester and Liverpool accounted for 175/200 websites audited. This geographic concentration is an important limit on interpretation and one reason not to extrapolate the findings to all clinics nationally.
Scenario context
One concern or procedure scenario was selected for each clinic according to the services presented. Non-surgical websites were usually tested from a concern such as wrinkles, facial ageing or sagging skin. Surgical and mixed websites were often tested using a known procedure or procedure-led scenario such as breast augmentation, rhinoplasty or blepharoplasty.
This difference is central to interpreting concern-led navigation. A person already researching rhinoplasty may reasonably begin with a procedure name. A person concerned about wrinkles may instead need the website to translate the concern into possible treatments, technologies or approaches. For that reason, the non-surgical result is the more informative test of concern-to-option guidance.
2. From a concern to relevant options
Almost every website offered a catalogue of treatments or procedures. Far fewer offered a clear route for a visitor who did not already know which treatment name to choose.
“Concern-led route found” does not mean that a treatment menu existed. It means the site helped the visitor move from the selected concern or scenario to relevant treatment or procedure options within the audit conditions. A flat menu of treatment names, brands or devices without a concern-to-option route could therefore be recorded as not found.
The non-surgical navigation gap
The non-surgical result is particularly clear:
109/120 websites listed treatments;
30/120 connected the tested concern to relevant options; and
81/120 listed treatments without providing that concern-led route.
Among non-surgical websites that listed treatments, 81/109 (74.3%) did not connect the tested concern to possible solutions within the time limit.
Once a concern-led route existed, it usually continued effectively. Of the 30 non-surgical websites with such a route, all 30 provided a dedicated service page and 29 also presented a selected-page call to action. The principal break was therefore at the start of the journey: helping the visitor identify which options were relevant.
A connected-journey signal
The 30 non-surgical websites with a concern-led route were also more likely than the 90 without one to contain several other connected-journey features.
Measure within non-surgical clinics
Concern route n=30
No concern route n=90
Dedicated selected-service page
30 (100.0%)
58 (64.4%)
All four core information elements
11 (36.7%)
18 (20.0%)
Price found
26 (86.7%)
79 (87.8%)
Online-booking pathway
23 (76.7%)
60 (66.7%)
Selected-page action
29 (96.7%)
71 (78.9%)
Exact practitioner identifiable
22 (73.3%)
50 (55.6%)
Full practitioner chain
19 (63.3%)
29 (32.2%)
Price visibility was almost identical in the two groups. The more noticeable differences involved continuity: a dedicated page, a page-level action and a practitioner connected to the service.
These are descriptive associations. The audit does not establish that concern-led navigation caused the other features, or that one site design produces better patient or commercial outcomes.
Figure 1. Non-surgical websites (n=120). Related but independent measures, not cumulative funnel stages.
3. Understanding the selected treatment or procedure
A basic explanation was usually visible on the selected page. Information supporting a more balanced and practical understanding appeared less consistently.
Selected-service information
All websites n=200
Surgical + mixed n=80
Non-surgical n=120
Treatment explanation
168 (84.0%)
78 (97.5%)
90 (75.0%)
Downtime or recovery
144 (72.0%)
75 (93.8%)
69 (57.5%)
Suitability or contraindications
100 (50.0%)
59 (73.8%)
41 (34.2%)
Risks or side effects mentioned
81 (40.5%)
50 (62.5%)
31 (25.8%)
Neither risks nor suitability information found
96 (48.0%)
19 (23.8%)
77 (64.2%)
Before/after or result examples
91 (45.5%)
49 (61.2%)
42 (35.0%)
All four core information elements
77 (38.5%)
48 (60.0%)
29 (24.2%)
All five elements, including results
54 (27.0%)
36 (45.0%)
18 (15.0%)
The four core elements were explanation, downtime or recovery, suitability or contraindications, and the presence of any risk or side-effect wording. Result examples were analysed separately because they are not essential to every service and the audit did not assess their authenticity, representativeness or quality.
The most direct gap was the combination of the two decision-relevant fields: on 96/200 websites (48.0%), neither suitability or contraindication information nor risk or side-effect wording was found for the selected service. This rose to 77/120 (64.2%) among non-surgical websites, compared with 19/80 (23.8%) among surgical or mixed websites.
Information thinned as the visitor’s questions became more decision-relevant
The cumulative information pattern shows where the selected-page journey narrowed:
Cumulative information visible
All websites
Surgical + mixed
Non-surgical
Explanation
168 (84.0%)
78 (97.5%)
90 (75.0%)
Explanation + downtime
143 (71.5%)
75 (93.8%)
68 (56.7%)
+ suitability
100 (50.0%)
59 (73.8%)
41 (34.2%)
+ risks
77 (38.5%)
48 (60.0%)
29 (24.2%)
+ result examples
54 (27.0%)
36 (45.0%)
18 (15.0%)
Because this table is cumulative, its second step contains 143 websites with both an explanation and downtime information, rather than all 144 websites where downtime information was found.
The audit found 87 websites with an explanation but no risk or side-effect wording. This represented 87/168 (51.8%) of websites containing an explanation. Among non-surgical websites with an explanation, 59/90 (65.6%) had no risk wording and 49/90 (54.4%) had no suitability information.
The surgical and mixed group was more information-rich on these measures, but the complete core set was still found on 48/80 websites rather than across the group.
Important boundary on the risk measure
“Risks or side effects mentioned” was a presence-only field. A “yes” means that some relevant wording was visible for the selected treatment or procedure. It does not mean that the information was complete, balanced, procedure-specific, clinically adequate or sufficient for consent. Conversely, “not found” means only that the wording was not visible within the audit conditions.
Figure 2. Cumulative information present on the selected service page. Each step is a subset of the one above.
Key statistic. Presence-only website measure; “not found” applies only within the audit conditions.
4. Information and action did not always develop together
Calls to action were not the main overall shortage. A page-level action appeared on 170/200 websites. The more common pattern was an action without the complete core information set.
Website profile
All websites
Surgical + mixed
Non-surgical
Complete core information + selected-page action
68 (34.0%)
42 (52.5%)
26 (21.7%)
Selected-page action without complete core information
102 (51.0%)
28 (35.0%)
74 (61.7%)
Complete core information but no selected-page action
9 (4.5%)
6 (7.5%)
3 (2.5%)
Neither
21 (10.5%)
4 (5.0%)
17 (14.2%)
Among the 170 websites with a selected-page action, 102 (60.0%) did not contain all four core information elements. The action-without-complete-information profile was the single largest group, accounting for 51.0% of all websites audited and 61.7% of non-surgical websites.
Non-surgical websites were frequently transaction-ready before they were information-complete
Within the non-surgical group:
105/120 exposed a treatment price either on the website or inside a booking flow;
83/120 offered an online-booking pathway;
100/120 offered at least one selected-page action; and
29/120 contained all four core information elements.
Among the 105 non-surgical websites where a price was found, 83 (79.0%) did not contain the complete core information set. Among the 83 with an online-booking pathway, 67 (80.7%) did not contain that set.
These findings do not assess whether the websites had enough information for a clinical decision or consent. They show that commercial and enquiry functions were often more consistently visible than the four audited information elements.
A complete journey remained uncommon
Only 46/200 websites combined a concern route, all four core information elements and a selected-page action. Adding price visibility reduced that number to 24/200. In the non-surgical group, 11/120 combined the concern route, core information and action; 8/120 also exposed a price.
This is a deliberately demanding combined measure. Its value is not as a score, but as a way to identify where individually available components failed to form an uninterrupted pre-enquiry pathway.
Figure 3. Website profiles by clinic type. “Information” means all four core elements; “action” means any selected-page call to action.
5. Price visibility and consultation clarity
Treatment-price visibility differed sharply by clinic type. This should be interpreted alongside the different pathways observed: non-surgical websites frequently presented exact-price booking, while surgical care begins with consultation.
Treatment pricing
Price visibility
All websites n=200
Surgical + mixed n=80
Non-surgical n=120
Visible directly on website
99 (49.5%)
24 (30.0%)
75 (62.5%)
Visible only after entering booking flow
30 (15.0%)
0 (0.0%)
30 (25.0%)
Found in either location
129 (64.5%)
24 (30.0%)
105 (87.5%)
Not found
71 (35.5%)
56 (70.0%)
15 (12.5%)
The 64.5% “found in either location” figure should not be shortened to “64.5% showed prices on their websites.” A treatment price was visible directly on 99/200 websites (49.5%); on another 30/200 (15.0%), it became visible only after the visitor entered a booking or service-selection flow. No treatment price was found on 71/200 (35.5%).
All 30 booking-flow-only prices occurred on non-surgical websites. Those prices were available to a visitor who began selecting a service, but not directly visible to someone comparing options on the public treatment pages.
Of the 129 prices found, 113 (87.6%) were exact, nine were presented as “from” and seven as ranges. Non-surgical pricing was overwhelmingly exact: 100/105 prices found. Surgical and mixed prices were more varied, with 13 exact prices, six “from” prices and five ranges.
The absence of a surgical procedure price should not be interpreted in isolation as a website failure. A more relevant question is whether the site explains the consultation that precedes a treatment decision.
Consultation information
Consultation measure
All websites n=200
Surgical + mixed n=80
Non-surgical n=120
Consultation price found
62 (31.0%)
19 (23.8%)
43 (35.8%)
First-consultation process found
48 (24.0%)
33 (41.2%)
15 (12.5%)
Either price or process found
83 (41.5%)
38 (47.5%)
45 (37.5%)
Neither found
117 (58.5%)
42 (52.5%)
75 (62.5%)
Consultation-request pathway
99 (49.5%)
67 (83.8%)
32 (26.7%)
Consultation information + request pathway
48 (24.0%)
34 (42.5%)
14 (11.7%)
The surgical and mixed websites generally reflected the consultation-first model: 67/80 offered a consultation-request pathway, while 15/80 offered an online-booking pathway.
The gap was in explanation. Among the 67 surgical or mixed websites with a consultation-request pathway, 33 (49.3%) did not show either a consultation price or a description of the first-consultation process. Across the whole surgical and mixed group, 34/80 combined consultation information with the request pathway.
For a prospective surgical patient, an immediate treatment-booking button is neither necessary nor necessarily appropriate. Useful website clarity instead includes what the first consultation is for, who it is with, whether there is a charge, what happens afterwards and how to request it.
Figure 4. Commercial pathways by clinic type. Direct website pricing and booking-flow-only pricing are shown separately.
6. Who provides the selected service?
Practitioner information was common, but three distinct questions produced different answers:
Is anyone named on the website?
Can the visitor connect an exact person to the selected service?
Is a professional registration or qualifying credential visible for that practitioner under the audit rule?
The headline table uses all 200 websites as its denominator. The conditional table that follows then examines only the 171 websites that named a practitioner.
Provider-visibility measure
All websites n=200
Surgical + mixed n=80
Non-surgical n=120
Practitioner information present
175 (87.5%)
77 (96.2%)
98 (81.7%)
At least one practitioner named
171 (85.5%)
76 (95.0%)
95 (79.2%)
Exact practitioner identifiable for selected service
131 (65.5%)
59 (73.8%)
72 (60.0%)
Registration or qualifying credential shown
142 (71.0%)
76 (95.0%)
66 (55.0%)
Named practitioner, credential not shown
29 (14.5%)
0 (0.0%)
29 (24.2%)
Named practitioner not linked to selected service
40 (20.0%)
17 (21.2%)
23 (19.2%)
Name + service link + credential shown
107 (53.5%)
59 (73.8%)
48 (40.0%)
Among the 171 websites that named at least one practitioner:
Provider profile
All named n=171
Surgical + mixed n=76
Non-surgical n=95
Service-linked + credential shown
107 (62.6%)
59 (77.6%)
48 (50.5%)
Service-linked, credential not shown
24 (14.0%)
0 (0.0%)
24 (25.3%)
Credential shown, not linked to service
35 (20.5%)
17 (22.4%)
18 (18.9%)
Neither service-linked nor credential shown
5 (2.9%)
0 (0.0%)
5 (5.3%)
In other words, a credential was shown on 142/200 websites (71.0%) in the headline analysis and on 142/171 practitioner-naming websites (83.0%) in the conditional analysis. The numerator is the same; only the question and denominator change.
The figures show why “a practitioner page exists” is not the same as “the visitor can identify who provides this service.” A generic team page can name several people without making the treatment-to-practitioner connection clear.
Two action-related findings reinforce this:
32/200 websites offered an online-booking pathway without making the exact practitioner for the selected service identifiable.
On 16/200 websites, a visitor could reach an online-booking pathway without finding a named practitioner anywhere on the site. All 16 were non-surgical.
Important boundary on credential reporting
“Credential shown” is a presence-only website measure using the agreed audit rule. “Not shown” means the visitor could not see a qualifying professional registration or credential within the audit conditions. It does not mean that the practitioner lacked one, and the study did not independently verify any practitioner’s professional status.
Figure 5. Provider visibility, all websites (n=200). “Not shown” does not mean a practitioner lacked a qualifying credential.
7. Planning a visit
Addresses were usually visible. Opening times and practical travel information were much less consistent.
Practical information
All websites n=200
Surgical + mixed n=80
Non-surgical n=120
General clinic location present
188 (94.0%)
80 (100.0%)
108 (90.0%)
Address or location answer found in strict task
182 (91.0%)
77 (96.2%)
105 (87.5%)
Opening hours found
89 (44.5%)
44 (55.0%)
45 (37.5%)
Saturday-opening answer found
81 (40.5%)
41 (51.2%)
40 (33.3%)
Parking information found
24 (12.0%)
10 (12.5%)
14 (11.7%)
Public-transport information found
23 (11.5%)
11 (13.8%)
12 (10.0%)
Parking or public-transport information found
31 (15.5%)
13 (16.2%)
18 (15.0%)
Finance or payment-plan information found
27 (13.5%)
23 (28.8%)
4 (3.3%)
Only 21/200 websites allowed the visitor to find all three of the following: an address or location, an answer about Saturday opening, and either parking or public-transport information.
The practical-information fields describe findability, not whether every clinic should offer Saturday opening, parking, public transport or finance. A “yes” to Saturday opening meant that the website answered the question, whether the answer was open or closed.
A small but revealing location detail
Twelve websites did not present a general clinic location during the initial clinic-understanding task, and four of those did not state even a city. All 12 were non-surgical websites. In the separate, stricter address/location task, an answer was not found for 18/200 websites.
The two measures should not be conflated: the first records whether general clinic location was present during initial orientation; the second reflects a dedicated information-finding task.
Figure 6. Visit-planning information, all websites (n=200).
8. Contact routes, answers and after-hours capability
The audit found many ways to initiate contact. It did not test whether those channels provided an answer while the clinic was closed. Email, forms and WhatsApp can accept a message at any hour; a phone number can remain visible at any hour. In each case, the response may still depend on staff becoming available.
This report therefore distinguishes three layers:
Capability layer
What the audit recorded
What the result does not prove
Information available now
Treatment pages, prices, consultation information, hours and practical details
That the information was complete or clinically adequate
Action available now
Booking, consultation request, callback, WhatsApp, form or phone route
That a human would respond immediately
Answer available now
An automated widget responding to the standard question
That every treatment question could be answered accurately or safely
Contact channels: common, but primarily routes to staff
Contact capability
All websites n=200
Surgical + mixed n=80
Non-surgical n=120
Email
184 (92.0%)
79 (98.8%)
105 (87.5%)
Phone
180 (90.0%)
79 (98.8%)
101 (84.2%)
Contact form
171 (85.5%)
74 (92.5%)
97 (80.8%)
WhatsApp
63 (31.5%)
22 (27.5%)
41 (34.2%)
Website chat
18 (9.0%)
5 (6.2%)
13 (10.8%)
Any contact channel
196 (98.0%)
80 (100.0%)
116 (96.7%)
Neither chat nor WhatsApp
122 (61.0%)
54 (67.5%)
68 (56.7%)
Strict question-first task possible without identification
6 (3.0%)
1 (1.2%)
5 (4.2%)
WhatsApp and phone counted as present if they were findable anywhere on the website within the audit time. They did not need to appear as floating buttons or in a persistent header or footer.
The strict anonymous-question field asked whether a visitor could begin with a basic question without first providing a name, email address, phone number or other identifying contact detail. Only 6/200 websites allowed that task. Of the 196 websites with at least one contact channel, 190 (96.9%) did not.
This measure is not only about privacy. It distinguishes a question-first journey from a lead-capture-first journey. A visitor may be researching recovery, suitability, price, location or possible options without being ready to become an identifiable lead. Requiring contact details before the first question changes the function from immediate information support to data capture and later follow-up.
The audit did not send emails, forms or WhatsApp messages and therefore cannot say that clinics responded slowly. It can say something narrower and more defensible: the presence of those channels did not demonstrate that the website itself could resolve a visitor’s question outside clinic hours.
Global and selected-page next steps
The most common site-wide next steps broadly matched the clinic’s service model.
Global next step
All websites n=200
Surgical + mixed n=80
Non-surgical n=120
Online booking
89 (44.5%)
10 (12.5%)
79 (65.8%)
Consultation-request form
92 (46.0%)
64 (80.0%)
28 (23.3%)
Callback request
18 (9.0%)
11 (13.8%)
7 (5.8%)
WhatsApp
56 (28.0%)
21 (26.2%)
35 (29.2%)
Phone
134 (67.0%)
65 (81.2%)
69 (57.5%)
Any asynchronous global next step
172 (86.0%)
70 (87.5%)
102 (85.0%)
At selected-service-page level, the actions were:
Selected-page action
All websites n=200
Surgical + mixed n=80
Non-surgical n=120
Online booking
82 (41.0%)
11 (13.8%)
71 (59.2%)
Consultation request
73 (36.5%)
54 (67.5%)
19 (15.8%)
Callback request
18 (9.0%)
11 (13.8%)
7 (5.8%)
Contact form
67 (33.5%)
34 (42.5%)
33 (27.5%)
Phone
74 (37.0%)
40 (50.0%)
34 (28.3%)
Any selected-page action
170 (85.0%)
70 (87.5%)
100 (83.3%)
The action categories overlap: one selected page could contain several.
Across the two layers, 156/200 websites had both a global asynchronous next step and a selected-page action. Fourteen had neither. The remaining 30 exposed an action at one level but not the other.
Placement still mattered for specific actions. Online booking appeared both globally and on the selected page for 73 websites, but only globally for 16 and only on the selected page for nine. Consultation requests appeared at both levels for 66 websites, only globally for 26 and only on the selected page for seven. A phone action was available globally for 134 websites but appeared on the selected service page for 74.
These actions matter after hours because they let a visitor move forward without waiting to place a call. They should not be described as 24/7 answers: booking confirms an action, while a form, callback request or WhatsApp message may still begin a delayed human conversation.
Chat: a small secondary analysis
Chat appeared on 18/200 websites, and not every widget could be identified as automated. Friction could be assessed passively for 13: seven allowed a first message without details, while six required a name, five an email address and four a phone number. These requirements overlap.
Only 10 widgets met the automated-test condition. When asked the standard Saturday-opening question, one answered correctly, five asked for personal data, two gave no response, one answered incorrectly and one responded without providing the requested answer.
The conditional sample is too small to estimate the performance of aesthetic-clinic chat technology generally. Within the tested sample, however, collecting details was more common than answering the operational question. Chat should therefore remain a supporting report box, not a population-level headline.
Figure 7. Main panel: all websites (n=200). Conditional chat panel: automated widgets tested (n=10), shown separately.
9. Two different website models, two different improvement priorities
The clinic-type split should not be read as a league table. The two groups serve different pathways, and the data show that their websites emphasised different functions.
Journey measure
Surgical + mixed n=80
Non-surgical n=120
Concern-led route found
63 (78.8%)
30 (25.0%)
All four core information elements
48 (60.0%)
29 (24.2%)
Neither risks nor suitability information found
19 (23.8%)
77 (64.2%)
Treatment price found in either location
24 (30.0%)
105 (87.5%)
Online-booking pathway
15 (18.8%)
83 (69.2%)
Consultation-request pathway
67 (83.8%)
32 (26.7%)
Any selected-page action
70 (87.5%)
100 (83.3%)
Full practitioner chain
59 (73.8%)
48 (40.0%)
The non-surgical price figure comprises 75 prices visible directly on the website and 30 visible only after entering a booking flow. All 24 surgical or mixed prices were visible directly.
Non-surgical websites: improve orientation before transaction
Non-surgical websites were commonly built around treatment lists, prices and booking. The main opportunity is to support the visitor who begins with a concern rather than knowledge of a product, brand, device or procedure name.
The strongest non-surgical priorities are:
concern-led navigation;
consistent suitability and risk wording;
practitioner-to-service linkage;
visible credentials under the agreed rule; and
keeping prices visible without requiring entry into a booking flow.
Surgical and mixed websites: explain the consultation that already structures the journey
Surgical and mixed websites generally offered the expected consultation-request route and contained more of the audited selected-procedure information. Immediate treatment booking is not the relevant benchmark for surgery.
The strongest priorities are:
explain what the first consultation involves;
show the consultation price or state clearly how it is determined;
identify who is likely to conduct it;
keep the consultation action on the selected procedure page; and
preserve the stronger treatment-information and provider-visibility patterns already present.
10. Recommendations
The recommendations below follow directly from the observed gaps. They are website-content and journey-design recommendations, not assessments of clinical practice.
Three minutes was the maximum for each information-finding task—not a usability target. The allowance was deliberately generous. A prospective visitor may stop searching much sooner, so key information should be obvious within seconds rather than merely discoverable before the timer expires.
1. Build a genuine concern-to-option route for non-surgical visitors
A “concerns” page should do more than repeat treatment names. It should:
describe the concern in language a prospective patient may use;
explain that different causes or presentations may require different approaches;
show the relevant treatment or consultation options without implying that every visitor is suitable;
link each option to a dedicated page; and
provide a clear next step for professional assessment.
This is the single clearest opportunity in the dataset because treatment lists were already widespread. The missing component was often the bridge between the visitor’s language and the clinic’s service taxonomy.
2. Use a standard information structure on every treatment or procedure page
Each selected-service page should make the audited information categories easy to locate:
what the treatment or procedure is;
the concern or indication it may address;
expected recovery or downtime;
who may or may not be suitable;
risks and side effects;
who provides the service;
price or consultation expectations; and
the appropriate next step.
The recommendation is not to make every page longer. A consistent hierarchy, concise summaries and optional detail can make balanced information easier to scan.
3. Put the right action at the end of the information journey
The dominant problem was not a lack of calls to action. It was that the action could appear without the complete audited information or away from the page where the visitor made sense of the service.
The action should match the service model:
non-surgical: book, request an assessment or ask a question;
surgical: request a consultation rather than imply immediate treatment booking;
uncertain visitor: offer a low-commitment route to ask which consultation or service is appropriate.
The same primary action should be available globally and on the relevant service page, with wording that sets an accurate expectation about what happens next.
4. Separate price transparency from the act of booking
Where a non-surgical service uses standard pricing, make the price visible on the treatment page or a clearly linked price page. A price hidden inside a booking flow is technically available, but less useful to someone comparing options.
For services where a single treatment price is not meaningful, explain why and replace ambiguity with useful context: consultation fee, starting price, range, factors that affect price, or the stage at which a personalised quotation is provided.
5. Explain the first consultation, especially for surgery
A consultation-request form should be accompanied by a short explanation covering:
the purpose of the first consultation;
who it is with;
whether there is a fee;
what information or assessment it includes;
whether treatment can occur on the same day, where relevant; and
what the visitor can expect after submitting the request.
This would close the principal gap found in the surgical and mixed journey without pushing those websites toward inappropriate direct treatment booking.
6. Connect each service to an identifiable practitioner
Team pages and credentials are most useful when connected to the service the visitor is researching. Each treatment or procedure page should identify the relevant practitioner or practitioner group and link to a profile showing the professional registration or qualifying credential used by the clinic.
Where the actual practitioner is selected later, the site can state that clearly and explain how the match is made. The aim is not to imply certainty where it does not exist, but to avoid leaving the provider relationship ambiguous.
7. Publish basic visit-planning information in one predictable place
At minimum, a clinic website should make its city or location, full address, opening hours and Saturday status easy to find. Where relevant, add parking and public-transport guidance. These details should sit on a clear location or contact page and be consistent with site-wide footers and external listings.
8. Distinguish contactability from immediate answers—and recognise the limits of static pages
Forms, email, phone, callback requests and WhatsApp allow a visitor to initiate contact, but not necessarily receive an answer—particularly outside clinic hours, when this audit was conducted. Improving treatment, consultation, pricing, suitability, risk and practical information on static pages is the essential first step. Yet even within the audit’s three-minute search window, important information was frequently not found.
Static pages cannot anticipate every question or ensure that each visitor finds the specific detail they need. A question-and-answer layer can address this remaining gap by surfacing relevant, clinic-approved information when staff are unavailable, without requiring personal details before they are necessary.
If automated, it should identify itself clearly, operate within defined medical limits and provide a route to staff when it cannot answer safely. It should not rely only on incomplete marketing pages, or it may reproduce the same gaps found in this audit. The study did not test response speed, comprehension, conversion or any particular automation model.
11. Conclusion
The websites audited were rarely without a contact route. Most offered multiple ways to initiate contact, and most selected-service pages included some form of next step. But a route that receives a message is not necessarily an answer, particularly outside clinic hours. The main weakness was continuity and support before that next step.
For non-surgical visitors, treatment catalogues did not consistently answer the first question: “Which options are relevant to my concern?” For surgical and mixed visitors, consultation forms did not consistently answer: “What will the first consultation involve?” Across both groups, explanations, prices, practitioner pages and calls to action often existed as separate assets without forming one complete patient-facing route.
The most useful standard for improvement is therefore not the number of pages, buttons or channels a site contains. It is whether a visitor can move coherently from concern, to understanding, to an appropriately informed next step—and whether a question can be answered before the visitor is required to become a lead.
12. Technical performance and accessibility
Five Google PageSpeed/Lighthouse outputs were analysed as supporting technical indicators: mobile and desktop performance scores, the automated mobile accessibility score, mobile Largest Contentful Paint (LCP) and mobile First Contentful Paint (FCP). The technical analysis is descriptive and does not rank individual websites.
The clearest result was the difference between mobile and desktop performance. Among 199 websites with comparable valid scores, desktop performance was higher for 183 (92.0%), the same for five (2.5%) and lower for 11 (5.5%). The median desktop-minus-mobile difference was 17 points. The median mobile score was 59, compared with 78 on desktop.
Supporting indicator
All websites
Surgical + mixed
Non-surgical
Median mobile performance score
59 (n=200)
58 (n=80)
59 (n=120)
Mobile performance score 90–100
12/200 (6.0%)
4/80 (5.0%)
8/120 (6.7%)
Median desktop performance score
78 (n=199)
76 (n=79)
80 (n=120)
Desktop performance score 90–100
53/199 (26.6%)
13/79 (16.5%)
40/120 (33.3%)
Median automated mobile accessibility score
88 (n=200)
85 (n=80)
90 (n=120)
Automated accessibility score 90–100
91/200 (45.5%)
26/80 (32.5%)
65/120 (54.2%)
Median mobile LCP
7.86 s (n=200)
8.89 s (n=80)
7.04 s (n=120)
Mobile LCP above 4 seconds
168/200 (84.0%)
66/80 (82.5%)
102/120 (85.0%)
Median mobile FCP
3.63 s (n=199)
3.77 s (n=79)
3.35 s (n=120)
Mobile FCP above 3 seconds
141/199 (70.9%)
58/79 (73.4%)
83/120 (69.2%)
Using Lighthouse performance-score bands, 56/200 mobile results (28.0%) were poor, 132/200 (66.0%) needed improvement and 12/200 (6.0%) were good. Desktop results were stronger: 10/199 (5.0%) were poor, 136/199 (68.3%) needed improvement and 53/199 (26.6%) were good.
Loading milestones showed the more substantial constraint. Mobile LCP was good for 6/200 websites (3.0%), needed improvement for 26/200 (13.0%) and was poor for 168/200 (84.0%). Mobile FCP was good for 13/199 websites (6.5%), needed improvement for 45/199 (22.6%) and was poor for 141/199 (70.9%). Seventy-five websites recorded LCP above ten seconds; one recorded LCP below one second and five recorded FCP below one second.
The automated mobile accessibility score had a median of 88. Using the displayed score bands, 91/200 websites (45.5%) scored 90–100, 109/200 (54.5%) scored 50–89 and none scored below 50. This score reflects only the Lighthouse checks that can be automated. It is not a complete accessibility assessment and does not establish legal or standards compliance.
Interpretation boundary. PageSpeed results are single-run lab snapshots and can vary with test conditions, page changes and measurement timing. LCP and FCP are technical paint milestones, not literal measures of how long every patient waited or whether a page was usable. One PageSpeed performance error encoded as zero was excluded from the desktop analysis, and one unavailable FCP result remained not applicable; all other valid technical values for those websites were retained.
Sponsors and partners
To be added. This section will acknowledge the distribution and research partners supporting this study, with their logos, links and any partner comment. Inclusion as a partner does not imply that a partner influenced the audit method, the data, or the findings, which remain the responsibility of iGlowly.
Methodology
Funding and competing interests
The study was funded and conducted by iGlowly. iGlowly develops patient-information software for aesthetic-clinic websites, including concern-led navigation and question-first support. This creates a commercial interest in some of the capabilities examined.
No clinic paid to participate, no individual clinic is ranked or named, and the audit did not evaluate iGlowly or compare software products. The final coding rules were documented and applied across all 200 records, ambiguous entries were rechecked, and all published combined measures were derived from the recorded audit fields.
The website coding was not independently duplicated. To support scrutiny, the de-identified record-level dataset, audit codebook and calculation tables accompany the report. These materials allow the published aggregate calculations to be reproduced while preserving the study’s non-ranking approach.
Sampling and audit unit
Clinic websites were sourced from UK Google Maps business listings using Apify. Listings were filtered to remove non-relevant businesses, including dentists, hairdressers and standalone laser-hair-removal outlets. URLs were validated as live before assessment so that clinic inclusion did not depend on the apparent content or quality of the website.
The audit unit was the patient-facing website or domain. The 200 canonical domains were unique; exact duplicate domains were removed. Separate patient-facing domains could remain as separate audit units where they appeared to share a practitioner, related business or sister brand. The study did not attempt to resolve legal ownership across domains and does not claim 200 independent businesses.
The final sample comprised 200 domains: 50 surgical, 30 mixed and 120 non-surgical. Audits were completed during July 2026.
Passive, outside-hours audit
Each website was assessed as an ordinary visitor would use it, beginning at the homepage and navigating normally. The audit took place outside normal weekday clinic hours: on Saturdays and Sundays, and on weekdays before 08:00 or after 22:30.
No account was created, no personal details were entered and no clinic was contacted. No form, booking, WhatsApp message or enquiry was submitted. Chat was assessed only under the passive and limited-interaction rules described below.
Three-minute information-finding tasks
Each information-finding task had a three-minute limit. The timer began when the relevant starting page started to load and continued through slow loading, pop-ups, cookie banners, complex navigation and page transitions. If the requested information was not found within that period, it was recorded as not found.
The same rule was applied to every website. “Not found” therefore means not visible within the audit conditions, not necessarily absent from every page of the website.
Scenario selection
One scenario was selected for each clinic according to the services shown on the website. Non-surgical scenarios generally began with a concern, while surgical and mixed scenarios more often used a known procedure or procedure-led concern. A selected treatment or procedure page was then assessed using a fixed set of fields.
Direct and derived measures
Most fields were coded as yes, no or not applicable under predefined rules. The report presents direct field counts and combined measures derived from those fields.
Key combined measures were defined as follows:
Measure
Definition
Any price found
Price visible on the website or, where no site price was visible, after entering the first stage of a booking flow
Four core information elements
Explanation and downtime/recovery and suitability/contraindications and any risk/side-effect wording
Any selected-page action
Online booking, consultation request, callback request, contact form or phone action on the selected page
Any asynchronous global next step
Global online booking, consultation request, callback request or WhatsApp
Consultation information
Consultation price or first-consultation process found
Consultation-request pathway
Global consultation request or selected-page consultation request
Online-booking pathway
Global online booking or selected-page online-booking action
Full practitioner chain
Practitioner named and exact practitioner identifiable for selected service and professional registration or qualifying credential shown
Strict anonymous question
A visitor could begin a basic question without first providing identifying contact details
Neither risks nor suitability found
Risk/side-effect wording not found and suitability/contraindication information not found for the selected service
Specific coding boundaries
Concern-led navigation. A treatment list alone did not qualify. The site had to help the visitor move from the selected concern or scenario to relevant options within the time limit.
Risks and side effects. This was a presence-only field. It did not assess completeness, balance, clinical adequacy or sufficiency for consent.
Practitioner and credential. “Practitioner named” recorded whether the site named at least one person providing the selected service. “Professional registration or qualifying credential shown” followed the agreed audit rule and recorded website visibility only. If no practitioner was named, the credential field was not applicable. A credential not shown on the website does not mean that the practitioner lacked one.
Pricing. The public website was checked first. A booking flow was checked only if no treatment price was visible on the site. Consultation price was assessed on the website, not inside the booking flow.
Surgical booking. The absence of immediate surgical online booking was not coded or interpreted as a failure. Consultation information and consultation-request pathways were treated as the more relevant measures.
Contact presence. Phone or WhatsApp counted as present if findable anywhere on the site within the audit time; it did not need to be a floating or persistent element.
Anonymous enquiry. The presence of a contact channel did not by itself qualify. The visitor had to be able to begin a basic question without giving a name, email, phone number or other identifying contact detail. The standardised field was used to distinguish question-first access from lead-capture-first access; it was not treated only as a privacy score.
Chat. Pre-chat friction was recorded only where it could be determined passively. A widget was treated as automated only where it explicitly identified itself as AI, automated or a chatbot, or where its behaviour clearly appeared automated. The standard Saturday-opening question was tested only in those cases.
After-hours capability. This was derived from existing information and next-step fields. It did not imply that web pages themselves “closed.” Information pages and automated answers could support immediate self-service; booking and request routes could let the visitor act; forms, email, callback and WhatsApp could accept a message but did not prove an immediate human response.
Analysis
The analysis is descriptive. It reports counts, percentages and transparent combined measures for:
all 200 websites audited;
80 surgical or mixed websites; and
120 non-surgical websites.
No inferential claim is made about all clinics in the UK. Cross-tabulations describe features that coexisted in the audited websites; they do not establish causal relationships.
Supporting technical testing
The final technical dataset retained five Google PageSpeed/Lighthouse outputs for the audited domains: mobile and desktop performance scores, mobile automated accessibility score, mobile LCP in milliseconds and mobile FCP in milliseconds. Results were summarised for all websites, surgical + mixed websites and non-surgical websites using valid metric-specific denominators.
Performance-score bands followed Lighthouse conventions: 0–49 poor, 50–89 needs improvement and 90–100 good. LCP was grouped as good at 2.5 seconds or below, needs improvement above 2.5 to 4 seconds, and poor above 4 seconds. FCP was grouped as good at 1.8 seconds or below, needs improvement above 1.8 to 3 seconds, and poor above 3 seconds.
Two metric-specific exceptions were documented during QA. One desktop performance request returned an error that the collection script encoded as zero; this result was treated as missing, giving n=199 for desktop performance. One website was unavailable for the final FCP test, giving n=199 for FCP. Other valid technical results for both websites were retained, and the raw source file was not altered.
The sample was sourced from filtered Google Maps listings and is geographically concentrated. It is not statistically representative of all aesthetic clinics in the UK.
The unit was the patient-facing domain. All domains were unique, but the sample was not necessarily 200 independent legal businesses or practitioners.
The audit reflects a July 2026 website snapshot. Websites can change.
The three-minute rule prioritised practical findability. Information may have existed elsewhere but remained undiscovered within the protocol.
A single scenario and selected service were used for each website. Other pages on the same site may differ.
The audit assessed website presence and navigation, not the clinical completeness, accuracy or quality of the information.
Practitioner credentials were recorded from the website and were not independently verified.
The study did not contact clinics, submit enquiries, measure response time, assess conversion or estimate revenue effects.
Chat findings are based on very small conditional samples and should remain secondary.
PageSpeed results are single-run lab-test snapshots and may vary over time or under different test conditions. Performance, LCP and FCP are supporting indicators rather than precise measures of patient waiting time or overall website quality; the automated accessibility score is not a complete accessibility assessment.
Appendix A. Surgical and mixed sensitivity check
The combined surgical + mixed group was used for the principal analysis. The following split shows that the two categories were similar on several central pathway measures.
Measure
Surgical n=50
Mixed n=30
Concern-led route
42 (84.0%)
21 (70.0%)
All four core information elements
31 (62.0%)
17 (56.7%)
Any treatment price found
15 (30.0%)
9 (30.0%)
Consultation-request pathway
42 (84.0%)
25 (83.3%)
Consultation information
27 (54.0%)
11 (36.7%)
Any selected-page action
43 (86.0%)
27 (90.0%)
Full practitioner chain
39 (78.0%)
20 (66.7%)
The split should be retained in the analytical appendix, particularly for consultation information. The combined group remains useful in the main report because it preserves a clear comparison with non-surgical pathways without claiming that surgical and mixed clinics are identical.
Appendix B. Chart production brief
Charts should show counts as well as percentages wherever space permits. Website denominators must remain visible: all websites n=200, surgical + mixed n=80 and non-surgical n=120. Avoid an overall clinic score, rankings, unlabeled traffic-light judgements or graphics that imply statistical representativeness. Red–amber–green technical bands may be used only with the published Lighthouse and web.dev thresholds stated alongside them.
Key-stat callout. Neither risks nor suitability found
Large-stat callout with clinic-type comparison
All websites 96/200 (48.0%); surgical + mixed 19/80 (23.8%); non-surgical 77/120 (64.2%). Use “not found within the audit conditions” and keep the presence-only caveat adjacent
Named 171/200 (85.5%); exact service-linked practitioner 131/200 (65.5%); full chain 107/200 (53.5%)
6. Visit-planning information
Horizontal grouped bars
Address/location 182/200 (91.0%); opening hours 89/200 (44.5%); Saturday answer 81/200 (40.5%); parking or public transport 31/200 (15.5%)
7. A route is not an answer
Main two-bar contrast plus conditional annotation
Any contact channel 196/200 (98.0%); strict anonymous-question task 6/200 (3.0%). Annotate separately: chat present 18/200; automated widgets tested n=10; correct answer 1/10. Do not graph 1/10 against the n=200 values as if denominators matched
8. Supporting mobile technical indicators
Three 100% stacked red–amber–green bars
Mobile LCP: good 6/200 (3.0%), needs improvement 26/200 (13.0%), poor 168/200 (84.0%). Mobile FCP: good 13/199 (6.5%), needs improvement 45/199 (22.6%), poor 141/199 (70.9%). Automated accessibility: 91/200 (45.5%) scored 90–100, 109/200 (54.5%) scored 50–89 and 0/200 scored below 50. Keep the lab-test and automated-assessment caveats adjacent.
Technical reporting note
The supporting technical chapter reports mobile and desktop performance, automated mobile accessibility, mobile LCP and mobile FCP using distributional summaries rather than clinic rankings. Performance and paint timings are treated directionally, not as literal patient wait times. The manual patient-journey outcomes remain the report’s primary findings.
----
Minimum citation guardrails: say “websites” or “clinics audited,” retain “found” or “not found within the audit conditions,” and do not extrapolate the sample to all UK clinics. The audit assessed website visibility and patient-journey continuity, not clinical quality, practitioner qualifications or causation.
