This Humanizing Growth session, hosted by the Institute for Real Growth, is a working map of commerce rather than a forecast. The guest sets out four modes: physical retail, e-commerce, discovery commerce, and conversational commerce.
The framing rule matters more than any single mode. Commerce is evolving, but one behavior does not replace another. All four will coexist, which is what makes the terrain hard to plan for.
Key findings
- Physical retail still carries the majority of purchases, put at around 70% outside the pandemic, and it is not going away. It is described as over deployed today rather than obsolete.
- The role of the store has changed even though its share has not. Conversion in physical retail is twice what it was ten years ago, because shoppers research online first and arrive intending to buy.
- Brand substitution on e-commerce platforms is estimated at around 20%, and the speaker expects it to reach a third or more. It is one of the fastest growing trends he names.
- Scale is what makes discovery commerce work. He cites 1.8 billion people engaging with Facebook’s properties per week, which is what generates the signals the recommendation works from.
The store, now the last step of an online journey
Because the shopper arrives having already decided, two things become critical. What is advertised must be in stock at the advertised price, and the in-store experience has to be consistent with what was seen online. Marketing and sales have to be synchronized for the channel to pay off.
E-commerce, where brand value leaks
The economics differ from a store trip. A shopper who has driven to a mall amortizes the effort by buying more, so baskets and margins are larger. Online purchase is focused on the item needed now, so both basket size and margin are lower. The incidental items around the main purchase are where the margin sits.
Strategy on the platforms is bottom of funnel rather than full funnel. That is where the substitution happens. A shopper who searched for an advertised product is shown similar items in the cart, backed by thousands of peer ratings. They then buy from a brand they have never heard of, on the strength of those reviews.
The conclusion he draws is a trade-off rather than a verdict. These routes to market are efficient at selling and detrimental to brand value, so the strategic question is how many units against how much long term brand value.
Discovery commerce, where the funnel collapses
In discovery commerce the purchase happens inside a social feed, and the consequence Perraudin emphasizes is that intent capture disappears. There is no search, so there is no signal that someone is in market. The system infers readiness from behavior instead.
The funnel collapses into a single moment. Awareness, consideration, preference, and purchase happen in the same millisecond as the ad is served. He calls this the revenge of the marketers, because discovery and conversion now occur together rather than in separate channels.
The practical implication is that assets have to be built to sell. Marketing has to be shoppable while still carrying the brand message, the preference message, and the purchase message in the same moment.
Conversational commerce, where service level is the brand
The fourth mode runs through messaging. Broadcast is one to millions and digital is one to many with personalization. A conversation is one to one, and that is where the difficulty sits. The biggest challenge he names is the service level: who answers when someone messages the brand, and whether that exchange is consistent with everything the brand has built.
What a marketing leader should actually do
The first move is organizational. In many companies sales, marketing, and e-commerce are separate teams with separate technology, agencies, and measures. His advice is to look hard at roles and responsibilities and not to be territorial about them.
The second is to optimize for the platforms rather than compete with them. Feed them assets, product information, and pricing data so their algorithms can represent the brand well.
The hiring advice follows from that and cuts against the fashion. What most companies need is data architects who can get content and pricing into a database and out through an interface. The data science can be left to the platforms, which will always have more of it.
The limit he puts on all of it
Asked about technology racing ahead of regulation, and citing 626 million facial recognition cameras operating in China, Perraudin puts the line plainly. Technology has to remain in service of people rather than the reverse, and he says that is the regulators’ job. Good intentions and self-regulation are not the answer, and he says regulation is welcome.
Scope and limitations
The substitution figure is an estimate rather than a published number, and the speaker says so. The platforms operate in silos and do not share that data.
The speaker works for one of the platforms he is describing, and he separates his personal view from his employer’s when he gives buying advice.
Source
This page summarises HGS #22: Lucas Perraudin (Head of AR/VR, Facebook) on the future of commerce and customer journeys, by Marc de Swaan Arons.