Gelle Pelle Fantesca · MVP of a cosmetic system · July 2020
Six days, five objections, one page
A kitchen appliance for your face at 5,900 ₽ — no category, no reviews, no history. I built everything the buyer saw. Paid traffic was run by colleagues.
6 days
from a blank page to a live landing page, 3—8 July 2020
5 → 5
buyer objections — blocks on the page, one for each
~900
words of copy written by hand — more than two years before ChatGPT
First sales
the funnel worked, the project was handed over to the client
There are no sales figures: spend and revenue stayed with the client and the colleagues who ran the ads. Every number here can be recounted from the copy of the landing page on this site.
ProductA system for making collagen masks at home
Price5,900 ₽ for the system, 1,490 ₽ for 32 capsules
What the product hadA brand and a box
My partPositioning, landing page, copy, content
01 · Product
There was no category — the buyer had nothing to compare it with
Not a cream, not a massager, not a sheet mask. A kitchen appliance for your face from a brand nobody had heard of.
How the system works
Capsulemarine collagen peptide
Basejuice, milk, tea or water
5 minutesone button
Mouldthe mask sets
Onto the face
Why a regular landing page would not work here
A regular product
you know which shelf it belongs on
the page argues with competitors
people compare prices
≠
This device
there is no shelf
nobody to argue with
people do not get what it is and leave before they reach the price
What the product had
A brand and a box
No manufacturer name, no history, no reviews. All of that had to be built before the ads were switched on.
02 · Objections
Five objections, five blocks — in the order they come to mind
Each one did not just dent the sale — it killed it outright. So the page is not a story about the product, it is five answers in a row.
ObjectionBlock on the page
“What even is this?”
→
How to useSix steps from picking ingredients to self-cleaning — against the fear of “I will never figure this out”
“Why 5,900 ₽?”
→
Price comparisonNext to it: a popular brand’s mask at ≈5,000 ₽ and a pack of sheet masks
“Is it collagen or just gelatine?”
→
FAQThis question comes first — otherwise the product looks like making aspic in your kitchen
“What do I pour into it?”
→
Beauty recipesWithout them the device is useless right after the first capsule
“What if it breaks?”
→
Warranty and service12 months, plus a register of service centres in 22 cities behind a link in the footer
The order is taken from the copy of the landing page itself, not drawn up after the fact.
03 · Six days
Six days from a blank page — and the markup shows it
Tilda stamps the creation time into every block and every form field. This is not reconstructed from memory.
Timestamps in the copy of the landing page
142 of 148
fall within 3–8 July 2020. The other six date from 2017 and 2018: library blocks that came with the template.
Fri, 3 JulyStarted in the evening
The weekendMost of the work
Wed, 8 JulyFinal edits
Before ChatGPT
2 yrs 5 mo
between the landing page and the release of ChatGPT on 30 November 2022. Copy, content selection and layout — all by hand. Not to say it used to be harder: just one date next to another.
50
blocks on the page — from the first screen to the footer
10
screens: from “About the product” and the instructions to the FAQ, recipes and contacts
22
cities in the service centre register behind a link in the footer
The whole landing page, scrolling inside as you scroll. Captured from the copy — open it: the page is live, there is just nowhere left to send an order.
04 · The copy
The work was done by three blocks other landing pages do not have
Every page has a first screen, benefits and a price. Here conversion rested on the FAQ, the recipes and the price comparison.
FAQ: the most awkward question first
1 of 8
“Are collagen and gelatine the same thing?” Anyone who has already thought “aspic” will not read on to question seven. Here the FAQ is not a reference section — it is part of the funnel.
Recipes: what to pour in
avocado oil
aloe juice
honey and cucumber
lavender oil
Before the purchase they answer “what would I even do with it”. After it, they give a reason to come back to the site.
The price in someone else’s money
5,900 ₽ — the system and 32 capsules, that is 32 masks
≈5,000 ₽ — one mask from a popular brand, 75 ml
20 sheet masks in a pack
The price was not defended — the frame of reference was changed. The buyer does the maths from there.
05 · Content
The first screen stopped promising and started explaining
First came a prototype with the structure, colours and photo palette. Next to the published page you can see where the plan met reality.
The published landing page · the prototype with a photo palette of juices, oils and herbs · the Instagram profile.
First-screen headline
A long-term investment in beauty
→Folk recipes in a new format — the best alternative to a salon treatment
A product with no category has nothing to back up the word “investment”. You explain an unfamiliar thing with two familiar ones it fits between.
Fifth menu item
Reviews→Instagram
There were no reviews and nowhere for them to come from. A live feed — unboxings, pinned stories — does not replace them, but it shows that someone stands behind the box.
06 · Ending
The test answered the question before any batch, warehouse or logistics
The ads brought people in; the decision to buy was made on the page. The first sales came in — I handed over the materials and the working funnel and left the project.
3—8 July 2020The landing page built and published
Summer 2020First sales, the project handed over
2021—2023The domain lives on in the web archive
TodayNo domain, but there is a copy
The product was later shut down. I do not know why: by then I was no longer on the project.
This is not a page about a product, it is five answers in a row. Conversion came from words and trust, not pictures.
What I see in this page today
Two questions about one thing
“When to start using collagen care” appears in the FAQ twice. Day six, the proofreading was not what it should be — better I say it myself.
A feed instead of reviews
The right move is to collect honest reviews from the first buyers and put them back on the page. The project never got there: I left before that.
No impact figures
Nobody measured the page’s conversion separately from the ads. The case shows not an effect but a way to take apart a product with no category.
My part
Positioning
Objection mapping, page structure, prototype, brand colours and photo palette.
Every number is taken from the copy of the landing page on this site — portfolio/gp. The six days, 148 timestamps and 50 blocks were counted from the Tilda markup; the 22 cities from the service centre register; the roughly 900 words from the page text without the register. The prototype, mockup and Instagram frame come from my portfolio. ChatGPT was released on 30 November 2022. The gellepelle.ru domain is down; web archive snapshots run from December 2021 to February 2023. There is no advertising data in this case: the ads were run by colleagues.
