When the customer stops visiting
Shopping agents could use a retailer’s expertise and take the purchase elsewhere. What happens to the economics of retail when advice and the sale become easier to separate?
BY ROB OTERO VIZ / FORO
7 minutes
At an event last week, I heard a founder being asked what shopping agents might mean for his business.
My reading of his answer was that he sees agents as an inevitability. The business is preparing by doubling down on being a trusted source of information and advice, giving customers and their future agents a reason to shop there first.
I left wondering whether being a trusted source of advice and being the preferred place to buy would remain the same thing. Retailers fund expertise and discovery partly through the purchases they generate. If agents use that expertise but take the transaction elsewhere, who pays for it?
What happens when the customer stops visiting but keeps buying?
Meet customers where they are
I’ve spent much of my career watching retailers work to earn preference. At Burberry, that meant greater control of the brand experience and building community through social. At Selfridges, it meant extending a remarkable physical experience digitally, with screens of pixels replacing physical marble. At Charlotte Tilbury, it meant creating new ways to discover and buy beauty digitally. The instruction was consistent: meet customers where they are. There was still a person experiencing what we had built, and the question was always, “Why choose us?”
At Selfridges, availability and exclusivity were part of the answer. So were taste, curation and the sense that somebody had brought together things worth discovering. Online, as in stores, we needed reasons for people to visit, buy something and come back.
Across retail, we have become increasingly proficient at the expensive business of winning or buying digital attention. Performance media brings people in; the shopping experience and loyalty proposition help turn those visits into valuable purchases and repeat business.
With agents researching and buying on a customer’s behalf, do we still need the same level of investment? Does the landing page, with its curated storytelling and expensive shoots, still warrant its place? The concern is that we create beautiful experiences customers no longer see.
The concern is that we create beautiful experiences customers no longer see.
Two requests and a different role for the retailer
“Find me this Burberry coat.”
The customer has chosen the product. The seller may matter less than price, availability, delivery and loyalty benefits. An agent could compare those conditions from a single instruction. The brand has won the product decision; the retailer still has to win the sale.
Now consider a different request.
“Find me a black coat for a wedding. Something that suits me.”
Judgement suddenly enters the chat. An agent needs to understand the occasion, taste and budget. Whose expertise will it use? Its own knowledge of the customer, a retailer’s understanding of fit and style, or a specialist the customer already trusts?
One possible arrangement is a customer’s agent working with a retailer’s agent, which puts together an offer the business can deliver. But that second agent has to earn its place. For a straightforward purchase, the customer’s agent might only need access to product, stock and ordering systems. A specialist styling agent could supply the advice.
In the coat example, advice about fit could be used to buy elsewhere. A coat with alterations arranged before the wedding gives the customer a reason to buy from that retailer. The expertise is attached to something the business can deliver, with a completion date the customer cares about.
Existing customer information could help too. With permission, previous purchases and loyalty benefits could inform the offer. But those advantages need to be visible while the agent is choosing. A benefit revealed only at checkout may arrive too late.
Agents and competing interests
Are these agents helping each other or competing? I guess it’s the same question as whether you think a sales associate is there to help or hard sell.
The buyer’s agent is meant to serve the customer’s request. The retailer’s agent represents a business that wants to convert and grow the basket. Why should my agent believe yours, and what can you offer that makes listening worthwhile?
An agent could research with one retailer and buy from another, or divide a basket across several sellers. Delivery charges and timing would still matter, but the work of comparing offers could become much easier.
I remember the concern about “showrooming”: customers walking the store, mobile in hand, then buying elsewhere. Should we put a sign-up requirement on the Wi-Fi so that at least we could capture customer information? We were considering adding friction to stop sales leaking elsewhere, which we realised was counterintuitive. Our focus became how we could intervene with human service and give customers a reason to take the product away there and then. When the exchange happens between agents, that opportunity may never arise.
There is also the purchase the customer had not planned to make. Someone arriving for a coat may discover shoes or a scarf. An agent with a narrow brief could suggest them, but what incentive does it have to do so? We cannot assume it will reproduce the browsing behaviour on which some of our basket economics depend.
Amazon makes the tension visible
The challenge is particularly acute for generalist retailers selling products customers can find in several places. On the face of it, Amazon should be well placed. Its range, delivery network and existing customer relationships give an agent good reasons to buy there.
But Amazon also earns money from advertising while customers decide what to buy. An outside agent arriving with the product and seller already chosen could leave Amazon fulfilling the order with less opportunity to sell visibility along the way.
Amazon’s dispute with Perplexity over its Comet shopping experience shows that access by outside agents is contested. Amazon cited transparency and customer experience concerns. Those concerns may be valid; the dispute does not establish an advertising motive. Its introduction of sponsored prompts also shows how advertising can follow customers into conversations with its own shopping assistant.
For me, the commercial question is who gets to influence the decision and when. Being a preferred place to fulfil an order may be less valuable than being the place where the customer decides what to buy.
Being a preferred place to fulfil an order may be less valuable than being the place where the customer decides what to buy.
Who pays for the value being created
Multi-brand retailers carry the costs of inventory, people, technology and fulfilment. Those with stores carry substantial property costs too. Adding another route to market does not remove the cost of serving the existing ones.
If decisions increasingly happen elsewhere, retailers could lose direct traffic, advertising income and additional purchases. Some of the digital experience they have invested in may have less influence on the order.
There is an upside. An agent could bring a ready-to-buy customer whom the retailer would otherwise have paid heavily to acquire, or someone it would never have targeted. Lower acquisition costs and better conversion could outweigh a smaller basket.
How much are we making per order? We may have saved on acquisition costs, but how does that change the economics of a fulfilled basket? Is the first purchase profitable, or do we need the customer to shop again to make the relationship pay? Those assumptions deserve another look if an agent can compare the market afresh each time.
That is where a new shopping channel becomes a business model question, with consequences for where the next pound of digital investment goes.
What needs to be true to get ahead
I still see trust as central, but retailers need to be precise about what they are trusted to do. A customer may name a retailer because they trust its taste or service. An agent comparing offers needs evidence that the stock is available and the promised service can be delivered.
Exclusive products, own label and specialist services can still give customers reasons to choose a retailer. The coat with alterations is one example. Even exclusivity has limits, though: it may win one item in a basket that the agent splits across several sellers.
Before investing, I would ask where we are today. What are we known for, and is it strong enough to protect our position? If that were the only reason someone shopped with us, would the basket still be profitable? I would also want to know what we learn after the sale. If the customer tells their assistant why the coat disappointed them, does that feedback reach us?
This will not unfold uniformly. Initially, agentic sales may be a share of an existing digital business. Shopping for pleasure is a pastime that ends in a purchase, and that is very different from replenishing household essentials. The same customer may want to browse on Saturday and delegate on Monday.
Retailers will have to serve both behaviours and understand the economics of each. The investment question is where we can still influence the choice, earn the sale and learn enough to be chosen again.