E-commerce & Omnicanal 2026 : ce qu'on retient de la journée de l'Institut du Commerce

We were proud sponsors of the Institut du Commerce's annual conference on e-commerce and omnichannel. A packed day, featuring contributions from Criteo, NielsenIQ, iProspect/dentsu, Unilever, TikTok, Sensefuel, Petrel, and many others. We also had the opportunity to take the stage ourselves, alongside Léa Nature and Mars Petcare, on the subject of product recommendation as a driver of e-commerce performance.

Here are the key takeaways we came away with, and what they tell us about the months ahead for retailers and brands.

A record year, but growth that is shifting

The mass-market e-commerce sector generated 21.4 billion euros in FMCG revenue, representing 14% of total category sales, and even more in certain verticals such as fragrances and skincare. A strong dynamic, but one that masks a more nuanced reality: in France, e-commerce growth remains more modest than elsewhere in the world.

That said, the weight of click & collect in France is unique: click & collect and home delivery operated by mass-market players have more than doubled in under ten years. A French specificity that is redrawing the competitive landscape between channels. 29% of the revenue lost by hypermarkets and supermarkets flows directly to click & collect and home delivery, a cannibalisation effect that NielsenIQ put squarely on the table during the day.

Another strong signal: recruiting new customers is no longer the primary growth driver. It is now purchase frequency that makes the difference. A shift in paradigm that places the shopper experience, and the ability to bring customers back, at the heart of e-commerce strategy. This is precisely the subject we came to defend, alongside Léa Nature and Mars Petcare: how balanced, shopper-centric product recommendations can turn a single visit into lasting loyalty.

Search is reinventing itself, and it has gone cross-platform

Traditional search still plays a central role: 80% of French consumers use Google every week. But that figure is declining compared to 2024, a sign that a recomposition is underway. Usages are diversifying: search is becoming social, driven by content and communities rather than keywords, with notable growth on Instagram, particularly among younger generations.

TikTok Shop illustrates this shift in spectacular fashion: within six months, the platform surpassed established players such as Zara and eBay in market share. 10% of French consumers say they have already made a purchase directly via TikTok Shop, and that figure rises to 26% among 15-24 year olds. The purchase journey, from inspiration through to conversion, is increasingly happening without friction, all within a single platform.

This recomposition of search does not only concern brands. It also directly challenges mass-market retailers, whose internal search engines must evolve to meet shopper expectations that are increasingly close to those found on these new platforms: more intuitive, more contextual, less dependent on exact keywords. This is a space where Matcha is already making progress, with new technologies in the pipeline to make product search more relevant. More on that very soon.

Being visible in AI-generated responses: the new frontier

Search is also becoming conversational: 79% of ChatGPT users use it at least once a week, and 95% say they are satisfied with the responses they receive. A rapid adoption that is fundamentally changing the nature of product visibility. On average, a conversational search surfaces only 2.4 brands in its response. A scarcity that changes everything: being identified, read, and recommended by AI agents is becoming a visibility challenge in its own right, much like SEO was for traditional search.

Unilever has made this a concrete priority. The company has built a dedicated team to rethink its entire product content with a view to improving visibility with AI agents, particularly Rufus on Amazon. It is also working with specialist tools to track and optimise this visibility within AI recommendations, an approach it considers one of its most promising growth levers for the years ahead.

Agentic commerce, mentioned repeatedly throughout the day, is no longer a distant prospect. It is a challenge to tackle now.

Retail media: shopper data before format

The workshop led by iProspect/dentsu delivered a series of clear convictions on retail media best practices. The central message: always start from business objectives before choosing a format, and leverage retailer shopper data to refine targeting, category by category, retailer by retailer.

Another key point: diversifying formats. Relying on a single lever, or replicating the same targeting from one campaign to the next, significantly limits performance. The recommendation is to test continuously and to ring-fence part of the budget for experimenting with new technologies. All within a rigorous measurement framework, an essential condition for genuinely iterating from one campaign to the next.

Towards a more circular e-commerce?

Slightly away from the core performance agenda, the question of reusable packaging deposit schemes was also raised, notably by Petrel, a mission-driven company seeking to transform existing players rather than compete with them. Initiatives already exist on certain circuits (Le Fourgon, Le Drive tout nu), but they remain underdeveloped in mass-market click & collect and home delivery. The barriers are multiple: changing consumer habits, reverse logistics, integration into product listings.

Beyond the logistics, the smooth functioning of deposit schemes also depends on the quality of product data and shopper education: how do you explain clearly, at the right moment in the journey, that a product is part of a deposit scheme and how the return process works? A question of pedagogy and clarity that is not unlike the challenges we address at Matcha every day, as Catherine Blandin (Mars Petcare) aptly highlighted during our roundtable, in the context of petcare and oral hygiene in particular, where recommendation and retail media address that same need for guidance. And who knows, perhaps one day the "Consigne pour réemploi" logo will appear on our product tiles too.

What this means for retailers and brands

Beyond the diversity of subjects covered, a common thread runs through all of these contributions: the intensification of competition, whether from pure-players, international marketplaces, or new entrants like TikTok Shop, is forcing every player to be impeccable on the fundamentals. That includes product availability and content richness, but above all the quality of the online shopping experience.

Because whether the goal is to build loyalty on click & collect, make search more relevant, or simply guide a shopper through an increasingly dense online aisle, the conclusion is the same: a high-performing purchase journey is first and foremost one that understands and supports the shopper, at every step.

Frequently asked questions

What are the key e-commerce trends to watch in 2026?

One particularly significant signal for retailers: the shift from recruitment to loyalty. Purchase frequency is now what makes the difference, and with it, the quality of the shopper experience at every visit. Beyond that, search is fragmenting: Google remains central but is losing ground, while social platforms and conversational search engines are gaining traction. In this context, very few brands surface in AI-generated responses, making visibility within these new channels a new strategic priority that cannot be overlooked.

What are the most effective levers for building shopper loyalty in e-commerce?

As recruitment slows, purchase frequency is becoming the primary growth driver. To activate it, retailers need to focus on the quality of the journey: a range that meets demand, an intuitive search experience, and personalised product recommendations that help shoppers find what they are looking for quickly, and keep coming back.

How can brands optimise their e-commerce visibility through data and retail media?

Two dimensions need to be worked on in parallel. Product data first: enriched, well-structured product listings are a prerequisite for appearing in search results, whether traditional or conversational. Shopper data second: drawing on real purchase data from the retailer to refine retail media targeting, diversifying formats rather than relying solely on sponsored products, and measuring rigorously to iterate. Retail media performance is built over time, not in a single campaign.