A customer chooses a photograph of a rain-soaked garden. That is an opening for a conversation, not proof that they will enjoy a fragrance described as a rain-soaked garden. Perhaps they like the plants, the quiet or a memory the image evokes.
For an overseas fragrance brand working with Chinese manufacturing and content partners, the useful question is whether visual content can help customers choose what to sample. That calls for more than a beautiful campaign: it needs an explainable suggestion and a way for the actual scent experience to disagree.
A discovery interface is not a measurement of smell
In its June 17, 2026 announcement, The Estée Lauder Companies described Jo Malone London’s Scent Scanner on Pinterest, rolling out in the U.S. and France. With permission, a user could connect a board, whose visual cues informed recommendations from the brand’s fragrances. We checked the announcement on October 17, not the live service or its current availability.
The business-design implication is that content can become an input to a selection journey. The announcement does not establish independently verified recommendation accuracy, sampling satisfaction or incremental sales. It is not evidence that aesthetic preference is a general-purpose measurement of scent preference.
There is a distinct production task alongside this. Nightflare Creative’s LVNEA case page describes building fragrance identity through still photography, botanical styling and art direction. The undated case description establishes the work it presents, not a recommendation system or commercial outcome.
An identity campaign asks what the brand feels like. A discovery journey also asks why a customer might try one candidate before another. Both can use compelling imagery, but the second needs product mapping, explanations and a real next step.
ONCE’s Two-List Fragrance Discovery method
This is a proposed prototype method, not an ONCE-validated model. Keep one record of chosen visuals and the customer’s reasons. Keep a separate record of what they actually sampled and preferred. Between them, retain the suggested candidates and the rationale for suggesting them.
Begin with a small set of brand-owned or appropriately licensed images or short clips. Explore the spaces, materials or situations the brand genuinely wants to distinguish. Ask people to explain their choices in their own words. A preference for a timber-lined room might concern color or furniture; do not automatically label it a preference for woody fragrance.
Next, someone who understands the products can suggest a limited set of candidates using approved product descriptions. Explain the connection as a suggestion, identify uncertainty and allow the customer to change direction—or receive no recommendation. Decorative citrus or flowers do not establish ingredients or benefits.
Then let actual sampling challenge the suggestion. Through a feasible store, brand-experience or sample-delivery arrangement, record what was smelled, the subsequent preference and the reason. Do not mark an unfinished sampling journey as satisfaction. Product owners should determine suitable sample preparation and use; this article does not prescribe a sensory-study protocol.
Consider a hypothetical home-fragrance brand. A participant chooses a warm study image but prefers a different candidate after smelling the samples. Preserve that disagreement rather than reporting a successful “warm-study conversion.” Ask whether the original attraction concerned the room, lighting or fragrance description. The answer may suggest a change to the visual, explanation or candidate. One case establishes neither algorithm failure nor consumer error.
Do not rush the two lists into a match-rate claim. Visual choice, agreement to sample, completed sampling, post-sampling preference and purchase are separate events. Record only what happened. Demonstrating superiority over ordinary fragrance selection would require a suitable comparison, not a collection of agreeable adjectives.
Buy content for the actual point of uncertainty
If the task is to establish an identity, distinctive photography or a hero film may be the right scope. There is no need to add a recommendation feature the team cannot maintain.
If shoppers like the atmosphere but cannot choose between products, commission content that makes approved distinctions understandable. Show why particular candidates are worth exploring without promising that an image reproduces the smell. The visual can open the journey; the product description must stand on its own.
If shoppers have already shortlisted products but cannot sample them, more atmospheric films may not remove the obstacle. Establish a sampling route the business can actually fulfill, then make content explaining access and next steps. Do not depict an unbuilt delivery capability as a global service.
For international production planning, bring those decisions to the Chinese content partner before creating a large asset list. Specify whether each image invites an aesthetic choice, explains a product or helps someone proceed to a real experience. Shared footage does not make those tasks interchangeable.
Prototype the service before commissioning the system
The first version need not connect social accounts or reproduce Scent Scanner’s technology. Approved brand content, manual recommendations and a small, complete sampling-and-follow-up journey can reveal where explanations and operations need work. Label it exploratory; do not promise automatic identification of personal scent preferences.
Scope visual content and choice prompts separately from product mapping, sampling operations, follow-up records and any later software implementation. Name who approves descriptions, maintains the candidate range and handles unavailable products or undelivered samples. A production company should not silently inherit perfumery, retail operations or cross-border fulfillment.
Use information sparingly. Explain the purpose, obtain appropriate permission and agree access, retention and deletion arrangements. A prototype does not require someone’s full saved-board history, account password or health and psychological information. Do not repurpose their choices for model training without authorization. These are service-design safeguards, not legal clearance for a particular deployment.
For ONCE, the opportunity to investigate is ongoing discovery content informed by permitted, actual feedback. The useful context is why someone chose an image, what was suggested and how the physical experience changed the conversation. It is not an existing proprietary recommendation engine, a customer-data asset we claim to own or a demonstrated margin gain.
When discussing fragrance content production with ONCE, start with approved descriptions, a maintainable candidate range, a feasible sampling route and the questions customers struggle to answer. Give people reasons to try—and room to dislike—before deciding how much content and technology to buy.
Source scope: the ELC announcement is dated June 17, 2026; the Nightflare case page gives no project publication date. Both were read October 17, 2026. The example is hypothetical. No live-service test, ONCE client result, scent-matching validation, health benefit or sales guarantee is claimed.
