demo · dashaproject.uk · UA

LUMÉ — marketing operations & automation

Demonstration portfolio. LUMÉ is a fictional Kyiv network of aesthetic-cosmetology studios used as a model client. Two research projects on real Ukraine-2026 market data, plus a live automation loop you can click through right now.

Research

Analysis on real open sources. LUMÉ is fictional; the market is real.

Unit economics

Network economics: average ticket, LTV, CAC, studio payback, P&L — on real Kyiv prices, salaries, rent and 2026 taxes. Interactive calculator + deep-dive + cheat sheet.

OpenDeep-diveCheat sheet

Competitor map

8–12 real Kyiv cosmetology networks: positioning, price tier, locations, digital presence. Where the niche for LUMÉ is.

OpenDeep-diveCheat sheet

Beauty market radar

Automated collection: ~700 YouTube channels, 125k+ videos, 27k+ search queries, 26/26 Google Trends terms, 1300+ news items. Dashboard + analyst brief refresh daily.

DashboardAnalyst brief
Cron refresh 10:30 daily · UI in Ukrainian

Automations — live

Everything works right now. No login where possible.

Client FAQ bot

Answers on services and prices, captures a booking intent, and hands medical questions to a doctor (never answers them itself). Ukrainian, 24/7.

Open chat
no login

Booking

Cal.com: 5 appointment types, studio hours Mon–Sat 09:00–21:00. Each booking fires into n8n → store + Telegram alert to the admin.

Booking page
no login

Lead quiz form

4 steps, reads UTM from the link, validates the phone. The lead runs the "Lead intake" scenario: attribution, scoring, working-hours SLA, hot lead to the doctor.

Open form
no login · form in Ukrainian

n8n — 6 scenarios

Lead intake, message routing, brief intake, campaign guardrails, ops watchdog, Cal.com booking. All active, all sending Telegram alerts.

How it connects (doc)n8n editor
doc — no login · editor — behind a login (Basic auth lume / LumeDemo2026, then n8n)

Market analyst bot

Each day takes the fresh radar JSON → a brief: market picture, 3 content opportunities with numbers, 2 product signals, 1 "what not to do". Finished text, not a table.

Latest brief
no login

Automation loop — document

What each automation does, best-use scenarios, how they hand work to each other. 3 end-to-end walkthroughs.

Read
in Ukrainian

What to try

  1. FAQ bot. Ask the price of biorevitalisation — it answers. Then ask "is it safe after childbirth" — the bot won't answer, it says it's passing the question to a doctor.
  2. Booking. Pick a slot on the Cal.com page and book under any name. The booking fires into an n8n scenario → store + a Telegram card to the admin (~1 sec).
  3. Lead quiz. Run the quiz — the link already carries UTM. The lead lands in "Lead intake": attribution, scoring, working-hours SLA, a high-value service triggers a separate alert to the doctor.
  4. Radar. Dashboard + analyst brief: where demand is shifting (polynucleotides +49), where the content gaps are, the botox backlash.

The n8n editor (to see the scenarios from the inside) is behind a login; everything else is open.