Ella Britten-Jones
Research, writing, product and film — Sydney
Nearly four years at MECCA taught me to ask better questions — and notice the problems people had learned to work around.
VESSEL in development — an illustrative composition of the hanging system with its containers, sketches, specification drawings and material swatches.
Working at MECCA taught me how much you can learn by genuinely listening to people.
I loved asking questions, listening for what someone actually needed and finding the right solution with them. Over time, the same frustrations and workarounds began appearing in completely different conversations.
People were already finding their own workarounds. I started wondering whether there was a better one.
VESSEL V1 — PRE-LAUNCH
A hanging travel skincare system with a container matched to each way a product is actually applied.
A real physical product I conceived and designed, developed directly with a manufacturer in China. Physical samples have been manufactured, received, tested and iterated.
Matched refillable containers, decanting tools and an interior mirror complete the system. Carry-on capacities are printed directly onto each container.
What travelling with your existing skincare usually looks like.
Pump / cleanser · Dropper / serum + facial oil · Jar / moisturiser · Tube / SPF · Atomiser / fragrance
I started using AI less as an answer machine and more as something to argue with. It gave me a fast way to challenge assumptions, expose gaps in my reasoning and test whether the idea still held up when I pushed against it.
I tested the competitive landscape, customer behaviour, positioning and economics against evidence strong enough to change my mind. Sometimes, it did.
Pre-launch model
Pre-launch assumptions, not performance data.
~AUD $28~AUD $53
My first landed-cost estimate was around AUD $28 a unit. Proper benchmarking put it closer to $53 — and showed I had left fulfilment out of the model entirely. I kept the worse number and rebuilt around it.
My core hypothesisRanked fifth
Ranked against the strength of the evidence, the problem VESSEL is built around — products losing the format they are meant to be applied in — came fifth. I left it where the evidence put it.
VESSEL’s own core hypothesis ranked fifth — and I left it there
I ranked each frustration by the strength of its evidence. The problem VESSEL is built around — products losing the format they’re meant to be applied in — came fifth, with the least proof. I left it where the evidence put it.
Swipe to compare all columns →
| # | Frustration | Intensity | Evidence |
|---|---|---|---|
| 1 | Leaking products damaging clothes and luggage | High | Strong — most cited problem across sources |
| 2 | Security confiscation of over-limit products | High | Strong — frequently cited, financial and emotional cost |
| 3 | Routine disruption — skipping or switching products | High | Strong — Mintel & Euromonitor 2025 |
| 4 | Pre-trip packing stress | Med-high | Qualitative — recurring in consumer journals |
| 5 | Application method compromise — the format problem | Med-high | Qualitative — VESSEL’s core hypothesis, unproven at scale |
| 6 | Product waste from partially used minis | Medium | Qualitative — consistent but background |
| 7 | Bathroom clutter and disorganisation | Medium | Qualitative — functional, not emotional |
| 8 | Travel minis unavailable for some premium products | Medium | Qualitative — category-specific |
Source — VESSEL Mini Tax Report, Section 03.
No contacts, no introductions, no idea how any of it worked. I started reaching out to manufacturers, learning what questions to ask and getting better at separating someone who could make a bag from someone who could actually make VESSEL.
I narrowed the search to three manufacturers and assessed them against the same requirements. Margaret in Fujian ultimately stood out because she engaged with the specification itself — questioning dimensions, tolerances and construction rather than simply sending back a quote.
A render is a wish. A specification is something a stranger can build from.
Selected manufacturing specification
A specification has to say what happens when a requirement cannot be met.
Function-critical pocket widths hold to ±1 mm across the full panel, everything else to ±3 mm, and nothing about the materials or construction changes without written approval.
A cross-section translating the locked product principle into forces, load paths and measurable geometry: elastic compression aligns each container against the back wall while the base ledge carries its weight.
A sequenced factory instruction showing how the structural board is enclosed, where the pocket panel is attached and which shortcuts would compromise alignment, rigidity or function.
Selected pages from a 36-page manufacturing specification and the sourcing record behind it. Tap any document to read it at full size.
The bag and refillable containers were developed through separate manufacturing paths. I started with the bag so I could test the complete system using off-the-shelf containers — refining the layout, storage and pocket dimensions before committing to custom containers designed precisely to fit them.
Some of it was wrong.
Parts of the sample had been manufactured outside specification. Pocket dimensions were off and some of the test containers didn't fit properly. Physical testing also exposed details that couldn't be fully understood on paper.
I documented every deviation, photographed the evidence and changed the way tolerances were communicated — separating the dimensions that controlled whether VESSEL actually worked from those where the manufacturer could have more room.
Once VESSEL was physical, people could respond to something real. I used quick Instagram polls with friends to ask what they packed, how they packed it and what they wanted kept together or separate. Some answers shaped V1; others became questions to carry forward.
The first sample made me rethink storage. Rather than making VESSEL bigger, I designed a detachable hygiene pouch: divided storage for toothbrush and toothpaste on one side, and a razor on the other. I put it back to the research group; the directional response informed its inclusion in V1.
Instagram polls run with a small, self-selected group of friends. Directional response, not validation of demand.
I developed the campaign narrative, art-directed its visual world, used AI-assisted production tools and made the final creative and editorial decisions.
Campaign film — final edit in progress
V1 developed through specification, sampling, physical testing and iteration.
Refining the system through further sampling, testing and manufacturing development.
Keep testing the premise while V1 is still in sampling, then finalise the bag and hygiene pouch and develop the fitted container set with a specialist supplier.
Two products can look alike without sharing a manufacturing origin. SOURCE IT treats resemblance as a lead, not a conclusion, and keeps confidence and supporting evidence visible.
Working prototype using demo data. Live search is not connected; commercial launch, sourcing accuracy, customers and supplier partnerships have not been validated.
Read the SOURCE IT case study →
Working face to face with customers taught me to notice what people do, what they work around and what they do not always think to say out loud.
When a pattern stays with me, I want to understand it properly, challenge my first answer and make something that can be tested. AI made me faster at all of it; it never decided what was worth building or when something wasn’t good enough.
The cost model I corrected upward, the hypothesis I left in fifth place and the sample that came back wrong are the work. When the story belongs to someone else, the habit is the same: find the version that is true, show the working, don’t overclaim the rest.
Tell me what you’re working on.