A Cut Above / TONI&GUY Vietnam: where the salon stands

What changed after 2023

Facebook stopped bringing in new clients. The Book Now button on the Facebook page brought about 120 first-time clients a year in 2020 and 2022, then 55 in 2023, 17 in 2024 and 6 in 2025. Its booking volume (about 850 a year in 2020 and 2022) did not simply disappear: loyal clients moved to the Fresha app and the Book Now link, and total online bookings were about the same in 2025 (1,260) as in 2022 (1,297).

Offline bookings fell. Bookings made at the desk, by phone or by message, including walk-ins, dropped from 1,560 in 2022 to 1,035 in 2025, about 525 a year. New clients who first came in offline fell from 215 to 104. This is where the lost visits are (see section 06).

The third bookable stylist left. Karina earned 650–780M ₫ a year in 2020 and 2022. The busiest days have not come back since she left (see section 10).

Prices rose about 30% while visits fell. A cut with Dmytro went from 700k ₫ (2019–2021) to 800k (2022), 1M (Dec 2025), then 1.1M including 10% VAT (from 12 Aug 2026). Average spend per visit rose from 1.17M to 1.51M ₫.

Top services by revenue, all years (million ₫)

Cuts with Dmytro are the engine: 4,878 services and 3.84B ₫, about a third of all revenue. Colour is the second pillar: Root Touch Up, Colour Tone, Creative Colour, Highlights and Balayage together earn roughly 3.6B ₫ at much higher tickets.

An's cuts are priced 30% below Dmytro's and earn far less per hour, but An carries most of the colour work, which is why the two stylists earn almost the same.

570 completed services were recorded at 0 ₫ (add-ons, complimentary fixes, voucher redemptions). Several services have dropped off the menu: brow work, the Trisyscore trichology range, Haircut & Beard Combo (last Dec 2025) and Balayage (last July 2026).

Completed revenue by stylist (million ₫)

2026 is January to September only.

Every previously bookable supporting stylist has left: Karina (Jan 2023), Jodi (May 2024) and Vu Anh (June 2025 to June 2026). Tu, Penny and Jean work in the salon but can't be booked in Fresha, so all recorded work sits with Dmytro and An. That is the main capacity limit and the biggest risk: if either is away, roughly half the takings stop. Section 10 looks at what this means for the 15-a-day target.

Dmytro does more services (1,632 in 2025 vs An's 923); An's average ticket is higher because of colour. Both have lost volume since 2023 at a similar rate, so the decline is about client flow, not one person.

New clients per year (Jan–Sep) and next-year retention

A client counts as new in the year of their first completed visit since January 2019, using the 2019–2022 appointments exports as the baseline. 2026 has the fewest new clients of any year since 2019, fewer even than the 2021 lockdown year.

The share of a year's clients who come back by September of the following year has slipped slowly, from 47% of 2022 clients to 43% of 2025 clients (comparing the same January–September window each time). Among regulars the return rate has held steadier. With fewer new clients arriving, this gradual slip is shrinking the client base, from 609 clients in Jan–Sep 2023 to 458 this year.

The loyal core is very loyal. The top clients have booked 100–180 services each over the period. Their visits look healthy; the leak is in the middle, among clients with 4–10 visits who drift away. The 129 lapsed regulars are people with 4 or more visits whose last visit was between June 2025 and May 2026.

Appointments by channel

Front-desk bookings (“Offline”: desk, phone, messages and walk-ins) are still the largest channel at about 47% of appointments, but they have fallen the most. Facebook booking has almost disappeared (624 appointment lines in 2023, 45 this year), mostly because existing clients now book through the Fresha app instead; Fresha Marketplace bookings rose sixfold from 2022 to 2025. The Book Now link has halved since 2023.

Cancellations have risen to 8.6% of appointments in 2026 from 7.0% in 2025, worth about 220M ₫ of booked revenue each year. No-shows are low at 0.6%. Fresha and Google bookings cancel most often (9%); automated marketing messages produce the most reliable bookings (93% completed) but are used sparingly.

Most bookings are made 1–3 days ahead, so last-minute reminders matter more than long-range promotion.

Four doors, one form. The Facebook and Instagram Book Now buttons, the Book Now link on the website, Reserve with Google and the Fresha Marketplace all open the same Fresha booking form. The channel label records only which door the client came through. The Facebook button went live on 6 December 2019 and has produced 3,517 appointment lines since.

Loyal clients switched doors. Of the 537 clients who used the Facebook button in 2020–2022, 123 still book. Their bookings now come mostly through Fresha Marketplace (38%) and the Book Now link (25%); only 5% still use the Facebook button. Total online bookings held steady.

What was lost was new clients and offline bookings. Facebook stopped bringing in about 115 new clients a year; Fresha Marketplace replaced about half of them. Offline bookings, which visible social media plausibly drives through messages and walk-ins, fell by about a third.

What the numbers show

The first period is the high point in the data. Facebook and Instagram bookings peaked (126 in June 2022, over 100 in most months to December), new clients reached about 34 a month, and new client profiles tagged “Facebook” as their referral source peaked at about 11.5 a month. It is the only stretch where social media clearly brought in new people.

After April 2023 the decline was gradual. Facebook bookings held at 72 in May, then fell over six months to 16 by October 2023, the pattern you would expect when posting stops and an existing audience slowly drifts away. Some of those button bookings moved to the Fresha app rather than being lost, so the clearer signs of social media's effect are new clients and offline bookings, both of which peaked in this period.

The second period produced no measurable lift, and going quiet in 2025 changed little because the channel had already faded.

Why the two periods differed

The first period ran on a plan, direction and content schedule agreed at the start and kept to throughout. The second came during a time of big open decisions for the business, including how to brand the salon and whether to take on TONI&GUY. That made it hard to hold one content direction for long, and four months is short for rebuilding an audience in any case.

The lesson for TONI&GUY is practical: results followed when the plan, the direction and the content production were agreed up front and held steady for long enough to build.

Caveats: the first period overlapped with the post-Covid rebound and a third bookable stylist. Fresha records only the last step of a booking, so clients who saw a post and then booked through Book Now, Fresha or a walk-in are not counted as social. No reach or post data from those periods was available.

What the ads did

They brought in new clients. Over the 36 days of advertising, 31 new clients came in, against 53 in the 86 days before, a rise of about 40% per day. Bookings tagged Google more than doubled (7 to 17 per 30 days) and Facebook bookings doubled from a low base. Sep 2026 had 30 new clients, the best month since August 2023.

They didn't raise total bookings. Bookings stayed at about 227 per 30 days, because Book Now link bookings fell from 43 to 26. Some existing clients probably switched to booking through the new website or Google rather than the old link.

Cost per extra new client is roughly 3M ₫ on total spend, or about 1.7M ₫ counting only the booking-focused campaigns (Google and the Meta booking push). A new client is worth about 2M ₫ in their first year (median; the average is 3.8M ₫), so booking-focused ads roughly pay back within a year before margins.

By campaign

Google Performance Max (EN and VI): 13.1M ₫, 142,651 impressions, 10,357 clicks (7.3% click-through). Website traffic fell from 500–750 visitors a day to about 20 once it paused on 28 September, so nearly all site traffic was paid.

Meta brand awareness: 5.5M ₫ reached 576,000 people about twice each. Engagement: 3.5M ₫ for 2,560 interactions. Booking push: 2.1M ₫ for 8 consultation requests and 2 purchases. Page growth (still running): 3.1M ₫ for 261 Facebook likes and 38 Instagram follows, about 10,000 ₫ per follower.

Tracking gap: Google Analytics records 91 online bookings worth 120M ₫ but credits none to Google ads, and Google Ads counts 360 “conversions” worth only 3.7M ₫. The 91 closely match the 92 Fresha bookings made in the same period through the Book Now link (49), Reserve with Google (33) and the Facebook button (10), so they are all bookings through the salon's own links, not only from the website. Neither figure can be read as bookings from ads. Fixing the existing tracking so completed Fresha bookings are credited to the ad click would solve this.

Evidence that people drop off at the form

Once ads were running, plenty of people reached the Fresha booking page and very few finished. The 91 completed bookings include those from Reserve with Google and the Facebook button, so the website's own completion rate is lower still; bookings from the website are a subset of the 49 Book Now link bookings.

The most likely reason is having to log in or create a Fresha account, which matters most to first-time visitors from ads. The data can't confirm which step loses people: login pages barely appear in Google Analytics (3 views), either because login happens inside the booking page or because few people get that far.

Evidence that the form isn't the whole story

The same form works for people who already know the salon or use Fresha. Fresha Marketplace produced 95 bookings in the same period, the most of any online channel, and about 55% of existing clients' bookings are made online.

Existing Vietnamese clients book online about as often as others. Judging by surname, about 8% of clients are Vietnamese; since 2023 they made 48% of bookings offline, against 44% for other clients. Service names are bilingual, but the form itself is in English only. Fresha translates just the page title into the visitor's device language, which is why Google Analytics shows booking-page views under Vietnamese, Russian, German, French, Chinese and Korean titles. For Vietnamese visitors who haven't booked before, an English-only form that requires an account is a plausible barrier.

Many Vietnamese visitors never reach the form. About 6,000 website users with Vietnamese-language devices produced 979 booking-page views (0.16 each); about 1,570 with English browsers produced 1,535 (about 1 each). Much of the loss happens on the website, or reflects less ready-to-book traffic from the broad Vietnamese Google campaign.

Rain is a modest headwind. Each hour of rain between 10:00 and 19:00 costs about 0.18 client visits. Only very wet days, with 5 or more rainy opening hours, are clearly worse, by about 0.8–0.9 visits or roughly 12–15% depending on the comparison (against a typical day for that weekday and month, or against all other days). The loss falls on clients who booked ahead; same-day bookings barely change and recorded cancellations don't rise, so some clients probably move their appointment rather than cancel.

September is wet every year. September 2026 had 106 rainy opening hours, against 103 in September 2025, and September was also the weakest month of 2024 and 2025. Client visits in September 2026 (140) still beat September 2025 (129). The extra rain compared with August (63 hours) explains about 8 visits, roughly 12M ₫. The weakest week, 15–20 September, was drier than 9–13 September, so something other than rain held it back.

Rainy-day messages to clients booked ahead are a cheap way to reduce this loss.

Cash collected by payment method (million ₫)

2026 is January to September only.

Card payments dropped from 832M ₫ in 2025 to 202M ₫ so far in 2026. TIMO bank transfer became the main method (49% of 2026 takings), then stopped on 9 August when the till switched to a D&C QR code, which has taken 220M ₫ since.

Only one refund in four years (2.5M ₫, May 2024). Tips are recorded as zero on every invoice, so either tips are handled outside the system or they are not being captured.

Vouchers: 40 issued since 2022, almost all redeemed. Two are still live: Cynthia Matu (3.41M ₫ remaining, bought this September) and Phil Kelly (200k ₫). Phil Kelly buys packs of 5–6 vouchers about once or twice a year, which looks like corporate gifting worth a conversation.

Stylist-hours per day: needed for 15 appointments vs available

Assumes opening hours of 10:00–19:00, based on booking start times. Sundays are shorter.

Where the salon is today. In 2026 the salon has averaged 6.3 client visits per open day and has not reached 15 on a single day; the best day had 14. Even in 2023, the strongest year, the average was 8.1 and only 7 days reached 15. The target, 15 client visits a day or about 375 a month, is about 2.4 times current daily volume.

Why two bookable stylists can't get there. Fifteen appointments at 86 minutes each need about 21.5 stylist-hours a day. Dmytro and An have at most 18 hours between them. Fully booked with no gaps they top out at about 12 appointments a day, or about 10 with normal gaps and overruns.

What five bookable stylists would change. With Tu, Penny and Jean bookable, there are about 45 stylist-hours a day, and 15 appointments would fill less than half of them. The juniors could take shorter cuts, kids’ cuts, blow-dries and colour prep, leaving Dmytro and An free for higher-value work.

The history backs this up. From 2020 to 2022 Karina was a third bookable stylist. 2022 had 11 days with 15+ visits and 39 days with 12+. With two bookable stylists, 2025 had 2 days with 15+ and 2026 has none. Even with three stylists the yearly average peaked at 8.1 visits a day, so more demand is needed as well.

Other effects of keeping the juniors hidden

Peak demand is likely being turned away. On their busiest 10% of days, An is booked for 8.5 hours and Dmytro for 7.3, which is effectively full. When Friday and Saturday mornings fill, online clients see no availability instead of a junior with a free slot.

There is no lower-priced option. Every bookable cut starts at 770k ₫ (An) or 1.1M ₫ (Dmytro), including VAT. Junior pricing would give new clients a cheaper way in, at a time when new-client numbers have fallen from 268 in 2023 to 176 this year.

The data hides their work. Work done by the juniors is recorded under Dmytro or An. That overstates the seniors’ output, means junior productivity can't be measured, and means stylist-level pay or commission can't be taken from the system. Tu Nguyen is already a system user with 94 bookings created since 2023, so adding them is a configuration change, not a new setup.

The other half of the problem is demand

On a typical day Dmytro and An are each booked for only about 4.3 hours, roughly half their time. If the juniors’ work is already booked under their names, the salon's real capacity is even less used than these figures show.

Making the juniors bookable removes the hard ceiling and frees peak slots. Filling 15 appointments a day also needs the growth actions in the next section: winning back lapsed regulars, rebooking at checkout and rebuilding online channels.