Why Marketplace Personalisation Is Essential

Personalisation is becoming essential because digital marketplaces are no longer short of choice; they are short of relevance. When thousands of listings, services and opportunities are available, the central problem is not displaying more results. It is understanding which results fit a person’s situation, location, budget, timing and underlying intent.

This does not mean every user should be enclosed in an invisible algorithmic bubble. Useful personalisation is better understood as a reduction in effort: the platform remembers appropriate preferences, interprets natural-language requests and adapts the experience while leaving the user in control. As expectations shift towards simpler, more responsive services, this capacity is becoming part of the basic infrastructure of a next-generation marketplace.

Personalisation is moving beyond recommendations

For many years, personalisation mainly meant recommending products similar to those previously viewed. That model remains useful, especially in retail and resale, but it is too narrow for the range of everyday needs now moving online.

A person looking for a bicycle may need one that fits their height, lies within five kilometres and can be collected after work. Someone seeking help with a move may care about availability, vehicle size, physical assistance and a fixed deadline. These requests cannot be represented adequately by a generic list of popular results.

From profile-based prediction to contextual understanding

Modern personalisation increasingly combines several forms of context:

the user’s explicitly stated preferences;
the intent expressed in the current request;
location, timing and budget constraints;
the type of transaction, such as buying, renting or booking;
previous interactions that the user has agreed may be remembered;
trust, availability and practical compatibility.

The important shift is from predicting what people might click to helping them accomplish what they actually want to do. This is closely connected to the new standards of marketplace user experience, where relevance, clarity and continuity matter as much as the size of the catalogue.

Conversational search is particularly significant. Instead of selecting a category, applying several filters and trying different keywords, a user can explain a need in ordinary language. The assistant layer can then translate that request into structured criteria. This is why conversational search offers a new way to find what users need: it starts with intent rather than vocabulary.

Why the traditional search model is under pressure

Keyword search was designed for relatively clear requests. It performs well when the user knows the name of an item and the marketplace has a predictable catalogue. It becomes less effective when the request is ambiguous, multi-part or spread across several categories.

Consider the difference between searching for “suitcase” and asking: “I need a cabin suitcase by Friday, preferably second-hand, within five kilometres, and I can collect after 6 pm.” The first produces a catalogue. The second describes a real-life constraint system.

Search engines provide a useful, although imperfect, analogy. They evolved from presenting lists of links towards offering direct answers and summarised guidance. Marketplaces are undergoing a related transition from lists of offers to assistance with decisions. The limit of the analogy is important: a marketplace must also manage availability, payment, trust and exchanges between people. It cannot simply generate an answer.

The broader history is explored in how marketplaces evolve with digital habits. As with feature phones becoming smartphones or scheduled television giving way to streaming, the older format does not cease to be useful. Instead, a new interface emerges around changing expectations.

Different platforms personalise in different ways

Established marketplaces have built strong habits and educated millions of people in peer-to-peer exchange. Their approaches reflect their categories, communities and histories; personalisation is not equally valuable in every context.

| Approach | Representative platforms | Genuine strength | Personalisation opportunity | |---|---|---|---| | Category-led specialised platform | Vinted, Depop, Beebs, Opla | Focused journeys, recognisable communities and category-specific conventions | Better sizing, style, life-stage or condition matching | | Large generalist marketplace | eBay, Leboncoin | Broad supply, established search behaviour and access to many categories | Better local matching and cross-category understanding | | Socially connected marketplace | Facebook Marketplace | Convenient discovery within an existing social environment and strong local reach | Recommendations informed by location and immediate browsing context | | Conversational multi-universe platform | Emerging models such as WEVONE | One interface can connect goods, services, missions, mobility and housing | Intent can be interpreted across several types of need |

Vinted’s focused resale experience and large fashion-oriented community make listing and browsing familiar. Depop brings a distinctive culture around vintage, style and seller identity. Beebs concentrates on the practical needs of families and children’s goods. Opla represents a focused second-hand approach. eBay contributes extraordinary category breadth and mature transaction mechanisms, while Leboncoin is deeply associated with local trade and general classifieds. Facebook Marketplace makes local discovery accessible from a social product many people already use.

These strengths do not disappear because conversational systems arrive. Indeed, a specialised platform can personalise deeply within its domain. A fashion platform may understand brands, sizes and style signals better than a broad service with insufficient category knowledge.

The challenge for a generalist marketplace or multi-universe platform is different: breadth can create complexity. Personalisation becomes the mechanism that prevents a broad ecosystem from feeling like an untidy collection of sections. This tension between focus and breadth is examined in why multi-universe platforms are gaining importance.

Two everyday scenarios

Scenario one: preparing for a family weekend

Imagine that Sam is travelling with two children. Sam needs a roof box, a pet sitter and possibly a larger car for three days. In today’s fragmented environment, that may require separate searches across resale, rental, local services and mobility apps. Each service asks for dates, distance, budget and other preferences again.

A personalised assistant layer could interpret the complete request, separate it into tasks and search the relevant universes. It might prioritise a roof-box rental rather than a purchase because Sam says it is needed only once. It could show pet carers available on the exact dates and mobility options suitable for the family’s luggage.

The value does not come from predicting that Sam “likes travel”. It comes from retaining the practical context of one project. This illustrates why users increasingly want to solve several needs from a single app.

Scenario two: turning free time into local income

Aisha is free on Saturday afternoon, has basic photography skills and owns a drill. She does not necessarily think of herself as a seller or service provider. A conventional marketplace may wait for her to choose a category and create an advert.

A contextual platform could instead help her identify relevant opportunities: a nearby mission photographing items for a small business, a request for help assembling furniture, or local demand for short-term tool rental. Recommendations could respect her chosen radius, availability and preferred level of commitment.

Here, personalisation connects capacity with demand. It can support the collaborative economy by recognising that one person may be a buyer in the morning, a renter in the afternoon and a service provider at the weekend. The platform’s role expands from hosting listings to enabling appropriate local matching.

Personalisation matters more in multi-universe ecosystems

When goods, services, missions, mobility and housing coexist, the same request can have several valid solutions. A person who needs a carpet cleaner could buy one second-hand, rent one locally or book someone to do the work. A rigid category structure requires the user to decide the transaction type before seeing the alternatives.

A well-designed multi-universe platform can begin with the outcome: “I need to clean the carpet before Monday.” The assistant layer can then present the possible routes, including their likely cost, timing and effort. The user still decides; personalisation simply organises the options around intent.

This is one reason platforms are becoming intelligent assistants. Their emerging function is not only to retrieve records but also to coordinate choices across a complete ecosystem.

“One account, one reputation” could reinforce that continuity. With suitable safeguards, a reliable history in one universe may help users establish trust elsewhere. However, reputation must remain contextual. Being an excellent second-hand seller does not automatically demonstrate competence as a tradesperson or driver. Portable identity can reduce friction, but category-specific verification and evidence still matter.

Good personalisation requires boundaries

Personalisation creates genuine risks when it is opaque, excessive or designed solely to maximise attention. A marketplace may narrow options too aggressively, infer sensitive information or make it difficult for users to understand why an offer appears.

Control should remain with the user

Responsible design should include:

clear explanations of why recommendations are shown;
easy ways to edit or reset preferences;
the ability to search without extensive profiling;
proportionate collection and retention of data;
visible distinctions between organic, sponsored and personalised results;
opportunities to widen the search beyond previous behaviour.

European data protection rules, including the General Data Protection Regulation, establish important principles around lawful processing, transparency and data minimisation. The EU Digital Services Act also places transparency obligations on online platforms and introduces additional requirements for the largest services, including around recommender systems. Compliance is only the baseline: trust also depends on understandable product design.

Personalisation should therefore be based increasingly on what people choose to express now, not merely on what a system silently infers from their past. Conversational interfaces can support that model because users can state and correct their requirements directly. As explained in how an AI understands user needs, useful interpretation should involve clarification rather than unsupported assumptions.

WEVONE as one illustration of the shift

WEVONE is a young platform built around the ambition of placing Mia, its conversational AI, at the centre of the experience. Its proposed model brings goods, services, missions, mobility and housing into a multi-universe platform designed around everyday needs.

In that model, personalisation is intended to make breadth manageable. Mia can act as an assistant layer, interpreting a request, asking clarifying questions and directing the user towards relevant options across several universes. The aim is a smoother, more personalised and contextual experience, supported by one account, one reputation where appropriate.

This remains a design direction rather than proof of market leadership or established results. WEVONE is one concrete illustration of how a platform can be built for usages emerging with AI; other generalist marketplaces, specialised platforms and new entrants may pursue different combinations of conversation, recommendation and human choice. The wider development is best understood as part of marketplaces in the era of generative AI, not as the story of one company alone.

Frequently asked questions

Is personalisation the same as targeted advertising?

No. Targeted advertising selects promotional messages, whereas marketplace personalisation can organise search results, remember declared constraints and adapt workflows. Advertising may use personalisation, but assistance and relevance are broader functions.

Will personalisation replace filters and categories?

Not entirely. Filters remain precise, familiar and easy to verify. Conversational search can complement them by translating intent into criteria, while categories continue to provide structure. The most useful interfaces are likely to let users move between conversation, filters and direct browsing.

Can specialised platforms personalise better than generalist marketplaces?

They can have an advantage within a narrow domain because their data model and community conventions are more focused. A generalist marketplace has greater breadth, while a multi-universe platform can connect several needs. Quality depends on category knowledge, interface design and responsible use of context, not breadth alone.

How can users avoid a recommendation bubble?

Platforms should provide non-personalised or minimally personalised views, explain recommendations and let users broaden or reset results. Users should also be able to state new intent explicitly so that past behaviour does not determine every future suggestion.

Conclusion

Personalisation is becoming essential because marketplace complexity is growing faster than users’ willingness to manage it. The next step is not simply to recommend more attractive products. It is to understand intent, preserve context and present appropriate routes through goods, services, missions, mobility and housing.

The strongest model will not be the one that knows the most about a person in secret. It will be the one that asks useful questions, applies context proportionately and keeps choice visible. In that sense, personalisation is not an ornamental feature. It is becoming the organising layer that can make a next-generation marketplace feel coherent, useful and genuinely responsive to everyday needs.

Why Personalisation Is Becoming Essential