Analytics for Mirrors
Last Updated: 14/09/2026
The Mirror Analytics section measures what happens in front of your ARShades Mirror stations: how many sessions each station generates, how far visitors get through the try-on journey, where they give up, what they try on, and how many photos they take away with them.
It is the store-floor counterpart of Analytics for VTO services: the same questions, asked of a physical device in a physical location rather than of a website.
How to Access the Section
From the left-hand menu, select Analytics › Mirrors.
Mirrors is added to the Analytics menu only when at least one Mirror station is associated with your teams. If you have no Mirror, the entry is not there. Administrators see every station; everyone else sees only the stations their teams own — both in the data and in the device filter.
Filters
All the figures on the page obey the filter bar at the top.
| Filter | What it does |
|---|---|
| Date range | The period under analysis. Opens on the last 30 days; the calendar offers quick presets: Today, Yesterday, Last 7 days, Last 30 days, This month, Last month, Last 3 months. |
| Mirror device | One or more specific stations. Empty means All mirrors. Stations are listed by their store name, followed by the city or group when available. |
| Mirror type | Restricts the analysis to one kind of station — for example Photobooth or Portrait. Empty means All types. |
| Reset | Returns the three filters to their defaults in one click. |
| Report | Opens the Excel report generator. See Downloadable reports. |
Overview
Six headline figures, each of which can be opened as a daily chart over the selected period.
| Card | What it counts |
|---|---|
| SESSIONS | Total Mirror sessions in the period. |
| MIRRORS | Distinct stations that produced at least one session. |
| PHOTOS | Photos captured during those sessions. |
| SHARED | Photos the visitor actually took away (shared to their phone). |
| COMPLETION % | Share of sessions that reached their natural end, rather than being abandoned or closed by staff. |
| AVG SESSION | Average session duration, in minutes and seconds. |
Click a card to swap the chart underneath it: the chart always shows the selected metric day by day, so a spike or a dead afternoon is visible at a glance.
User Journey
User Journey Funnel
The funnel counts, for each screen of the try-on flow, how many sessions reached it. It covers seventeen steps across the three paths a visitor can take:
- Standard flow — Gender Selection, Frame Type Selection, Brand Selection, Glasses Product Selection, 3D Model Viewer, VTO Portrait, VTO Photobooth, Favorite Styles Result.
- Showroom flow — the Showroom variants of the same first steps, plus Showroom Page, VTO Quick Showroom, VTO Style Selection Showroom and Finish No Photos Showroom.
- Completion — Photo Acquisition and Finish Get Photos Showroom. Reaching either of these is what counts as a finished journey.
A session appears in a step if it opened that screen at least once. A visitor who goes back and forth is counted once per step, not once per visit.
Mirror Type and Language
Two small distribution charts sit beside the funnel:
- Mirror Type — how the sessions split across the kinds of station you own.
- Language — which interface language visitors used. On a multilingual location this is the quickest read on who is walking in.
Demographics
When a station recognises the person in front of it, the recognition is recorded anonymously — a set of estimated attributes, never an identity.
| Figure | Meaning |
|---|---|
| People recognized | How many people were detected across all sessions in the period. One session can contain more than one. |
| Wearing glasses | How many of them were already wearing glasses. |
| Wearing glasses rate | The same figure as a percentage of people recognised. |
| Avg per session | Average number of people detected per session. |
Three breakdowns follow, each as a share of the people recognised: Age Range, Gender and Ethnicity. Attributes the station could not estimate are grouped under Unknown rather than dropped, so the percentages always describe the whole population detected.
Abandonment
Every session ends in exactly one of three ways, and the three always add up to the total:
| Outcome | What happened |
|---|---|
| Completed | The session reached its natural end. |
| Terminated | Staff closed the session from the station. |
| Abandoned | Everything else — the visitor walked away without finishing. |
Six figures summarise the period:
| Figure | Meaning |
|---|---|
| COMPLETED | Number of completed sessions. |
| ABANDON % | Share of sessions that did not complete — abandoned and terminated together. |
| ABANDONED | Sessions the visitor left unfinished. |
| TERMINATED | Sessions closed by staff. |
| BOUNCE % | Share of sessions shorter than 10 seconds: someone touched the station and left immediately. |
| AVG ABANDON TIME | How long an unfinished session lasted on average — how much patience visitors had before giving up. |
Where people drop out
A table walks the journey step by step:
- Step — the screen.
- Sessions — sessions that reached this step.
- Reach % — that figure as a share of all sessions.
- Drop-off % — sessions lost between the previous step and this one.
The largest Drop-off % is the single most useful number on this page: it names the screen that is costing you visitors.
Abandonment by station
A second table repeats the outcome split per station — Store, Type, Sessions, Completed, Terminated, Abandoned — so you can tell a problem with the flow (bad everywhere) from a problem with one location (bad in one shop only).
Photos
Nine counters describe what happened to the pictures taken at the station:
| Figure | Meaning |
|---|---|
| Photos captured | Photos taken across all sessions. |
| Photos shared | Photos the visitor sent to themselves. |
| Share rate | Shared as a percentage of captured. |
| Deleted | Sessions in which the visitor deleted their photos. |
| Retakes | How many times visitors retook a shot. |
| Avg retakes / session | Retakes divided by the sessions that involved the photo flow. |
| QR scanned | Sessions in which the visitor scanned the QR code to collect their photos. |
| Showroom confirmed | Sessions in which the visitor confirmed the Showroom step. |
| Privacy accepted | Sessions in which the visitor accepted the privacy notice. |
The View captured photos button opens the gallery of the photos taken in the selected period, with the number of available photos in the label. Each entry shows the station, the date and time, and the model and colour tried. A search box filters by mirror, session, model, brand or colour. The button is disabled when the period contains no photos.
A visitor who takes a photo has tried something on. A visitor who shares it has taken your product home on their phone. When the two numbers drift apart, look at the QR step.
Device Breakdown
One row per station, for comparing locations directly: Store, Type, Sessions, Avg Session, Ended, Terminated, Completion %, Photos, Shared and Last Session.
Last Session is worth a habit: a station whose last session is several days old is usually a station that is switched off, not a station nobody likes.
Products
Three tables, selectable one at a time, each showing the number of entries in its label:
- Top VTO Products — the most tried models: Product, Sessions, VTO views, 3D views, Variants viewed.
- Top Variants — the same ranking at colour level: Product, Frame color, Lens color, Sessions, VTO views, 3D views.
- Favourites — the models visitors explicitly marked as favourites: Product, Favourites, Variants.
The distinction matters: views measure curiosity, favourites measure intent. A model that is tried by everybody and favourited by nobody is doing a job for your window display, not for your order book.
Downloadable Reports
The Report button opens Generate Mirror Excel Report, which produces an Excel workbook in two shapes:
| Report | What it contains |
|---|---|
| VTO Explorer (per store) | Device activity together with the full Mirror analytics — sessions, journey, abandonment, people recognition, products, photos. Selecting several stations adds one sheet per station plus aggregate sheets. |
| Photobooth / Portrait (per location) | Photobooth and Portrait sessions grouped by location, one sheet per city. |
Both reports ask for their own date range, and offer Select all / Deselect all on the station list. Generation runs through visible stages — loading the sessions, resolving the variant codes, building the workbook — and can take a while on a wide date range; leave the dialog open until it reports that the report has been generated.
Reading the Data
- The page always describes the filtered set. Every card, chart and table on the page reflects the date range and the station filters. Before comparing two periods, check that the station selection has not changed as well.
- Completion is not the only measure of a good session. A visitor who tries four models and walks away without a photo has still engaged with your products; that session counts as abandoned. Read COMPLETION % beside AVG SESSION rather than on its own.
- Bounce sessions are noise you want to see. Sessions under ten seconds are usually passers-by brushing the screen. A high BOUNCE % is not a failure of the experience — it is a fact about where the station is placed.