DoorDash added Macy's, Anthropologie, The North Face, Vans and Timberland to its retail marketplace this month, launched a Dasher Returns service that coordinates a pickup for items shoppers want to send back, and expanded Ask DoorDash so customers can place orders through simple text communication without opening the app. The company describes the ambition as building a mall in your pocket, and the texting layer is the part worth studying closely.
What Changed in a Single Month
Mike Goldblatt, vice president of enterprise partnerships at DoorDash, told Modern Retail that the retail and technology expansions are part of a continued effort to move beyond restaurant delivery. He said this year's focus is expanding partnerships to serve retailers and customers better as user behavior shifts from one-off convenience orders toward a more routine shopping cadence. Mike Goldblatt said the expansions extend DoorDash beyond restaurant delivery
The month also included a milestone partnership with Costco, which had been at the top of the company's wish list of retailers to add, and an expansion of its Snap and EBT program with seven new grocers including regional chains such as Albertsons Market, Amigos and Boyer's Food Markets. With the warehouse club addition, DoorDash now carries eight of the top 10 U.S. grocers. DoorDash now has eight of the top 10 U.S. grocers on the platform
Why Texting Beats Browsing for Routine Orders
A text thread removes the two costs that make app shopping slow: navigation and decision fatigue. A shopper who already knows what they want does not need a catalogue, they need a confirmation. Converting that intent into an order through a message is faster than any interface improvement, because it skips the interface entirely.
The trade-off is that a text thread gives the platform far less control over what gets discovered. In a browse experience, placement is a merchandising decision the platform can sell. In a conversation, the recommendation emerges from the exchange, and the platform's ability to steer it depends entirely on how the assistant is built and what data it draws on.
Three Effects on Retail Media
Sponsored placement loses its anchor
Retail media is priced against visible slots: search results, category pages, home screen tiles. A conversational order has no slots in the conventional sense. Platforms will need a new unit of inventory, likely defined by intent category rather than position, and advertisers will need a new way to value it.
Attribution moves to the conversation
When an order originates in a message, the attribution question becomes which part of the exchange drove the decision. That is closer to a qualitative transcript than a click path, and it makes the measurement of incrementality harder rather than easier, which is precisely why platforms are keen to own the conversation layer.
Returns become part of the funnel
Dasher Returns addresses the inconvenience of sending items back by coordinating a pickup. Returns have historically been treated as a post-purchase cost center, but in a conversational commerce model they are a trust input: a shopper who knows returns are easy is more willing to order a size they are unsure about.
The Routine Cadence Is the Real Prize
Goldblatt's phrasing about shifting from one-off convenience orders toward a routine shopping cadence is the strategic center of gravity here. Convenience orders are opportunistic and price-insensitive but infrequent. Routine orders are frequent, predictable and far more valuable over a year, but they require the platform to be reliable in a way that a Friday night dinner order never demanded. DoorDash is building what it calls a mall in your pocket
Reliability in grocery means inventory accuracy at the item level. A missing item in a restaurant order is an annoyance; a missing item in a weekly grocery order breaks the routine and sends the shopper back to a store. This is why the grocer partnerships matter more than the fashion brands, even though the fashion brands generate the headlines.
What Retailers Should Ask Before Joining
The first question is who owns the customer relationship inside the conversation. If the assistant answers product questions using the platform's data rather than the retailer's, the retailer becomes a fulfillment node. The second is what data comes back: item-level demand signals are worth far more than order totals, and they are the difference between participating in a channel and learning from it.
The third is how the conversation is documented. Transcripts are the only durable record of why a shopper chose one item over another, and they are also the raw material any future assistant will be trained or grounded on. Retailers that negotiate access to that record will be able to improve their own product pages, while those that do not will keep guessing at intent. PepsiCo is betting on refrigerated fresh foods to drive growth
