Commerce Media Networks Compared 2026: Uber, DoorDash, Instacart for DTC Brands

Commerce media networks are ad platforms owned by delivery and mobility apps that sell audiences built on completed orders and trips. They matter now because Uber, DoorDash, and Instacart together move roughly USD 200 billion of US gross bookings a year and report sales back against real orders. Start with Instacart for grocery CPG, DoorDash for food and beverage, Uber when you need reach beyond the basket.

Key takeaway: Commerce media is the cheapest place a DTC brand can buy an ad that sits two taps away from a checkout it can measure.

Quick answer: how do Uber, DoorDash, and Instacart ads compare in 2026?

All three sell in-app sponsored placements plus off-platform display against first-party order data. Instacart is the SKU specialist for grocery CPG at USD 10 to USD 18 CPM with reported ROAS of 3 to 6. DoorDash reaches about 42 million monthly actives at USD 12 to USD 22 CPM. Uber is widest, about 170 million users, at USD 16 to USD 30 CPM.

1. What is a commerce media network in 2026?

A commerce media network is an ad platform run by a marketplace or delivery app that monetizes its own transaction data and in-app real estate. The defining feature is proximity: the ad appears in the same session where the order gets placed, so attribution is closed-loop by default.

The category is distinct from its neighbours. Retail media (Walmart Connect, Target Roundel) sells inventory inside a retailer that owns the shelf. Bank-issued media (Chase, Capital One) sells reach against settled card spend with no shelf at all. Commerce media sits between them: the app owns fulfilment and the basket but usually not the product. eMarketer puts total US retail and commerce media spend above USD 65 billion for 2026, with delivery and mobility apps the fastest-growing slice. See IAB's retail media research hub for category definitions and measurement standards.

The three anchors are Uber Advertising, DoorDash Ads, and Instacart Ads. Everything else (Lyft Media, Gopuff, Grubhub) is a rounding error on a DTC media plan in 2026.

Key takeaway: If the platform delivers the order, its ad network can prove the sale. That is the whole thesis of commerce media.

2. Uber Ads: reach, formats, and CPM benchmarks

Uber Advertising is the reach play: roughly 170 million monthly platform users across Rides and Eats at CPMs of USD 16 to USD 30. Managed commitments start near USD 10,000 per campaign; self-serve Uber Eats sponsored listings start far lower.

In-app formats

Sponsored Items and Sponsored Brands run inside Uber Eats search and browse on a CPC basis, typically USD 0.45 to USD 1.20 per click. Uber Eats homepage takeovers sell as a fixed daily rate, often USD 15,000 to USD 40,000 per day in the US.

Journey Ads

Journey Ads have no clean equivalent elsewhere: full-screen video and static served during the wait and ride phases of a trip. Attention is high because the rider is holding the phone with nothing to do. We observed view-through rates 30 to 50 percent above standard in-app display when creative was built for the moment rather than resized from Meta. Uber also extends Journey audiences out of app to display and CTV through DSP partners. Full detail in our Uber Ads network playbook and on Uber Advertising.

Key takeaway: Buy Uber for attention and reach, not for basket-level attribution. Its strength is the captive ride window.

3. DoorDash Ads: sponsored listings and off-platform display

DoorDash Ads is the food and beverage specialist: roughly 42 million monthly actives, 22 million-plus DashPass subscribers, USD 12 to USD 22 display CPMs, and listing CPCs of USD 0.35 to USD 1.10. Self-serve starts near USD 100 per day, the lowest floor of the three.

Sponsored Products place your SKU above organic results inside convenience and grocery storefronts. Sponsored Brands buy category share of voice with a logo, tagline, and a three-product carousel. DashPass subscribers order 3 to 4 times more often than non-subscribers, so retargeting them on lapsed purchase is the highest-ROAS cell we run there.

The off-platform Display Network extends DoorDash first-party order audiences to web and app inventory, sold CPM with closed-loop measurement back to DoorDash orders. In our client accounts it returns 40 to 60 percent of in-app ROAS but adds new-to-brand buyers at a lower CPM. See our DoorDash Ads playbook and the DoorDash Ads Manager documentation for current placement inventory.

Key takeaway: DoorDash has the lowest entry cost of the three, which makes it the correct first test for brands under USD 10,000 monthly commerce media budgets.

4. Instacart Ads: sponsored products, brand pages, and shoppable display

Instacart Ads is the closed-loop leader for grocery CPG: about 14 million monthly actives, over 1,800 retail banners, reported ROAS of 3 to 6 on Featured Products, CPCs of USD 0.50 to USD 2.00, and a USD 100 self-serve minimum.

Featured Products is the workhorse, bidding a SKU into search and aisle browse. Display buys category and homepage banners at USD 10 to USD 18 CPM. Shoppable Video and Shoppable Display add a cart button on the creative and drive higher new-to-brand rates than search because they interrupt rather than answer. Carrot Ads extends the same stack into partner retailer sites as a white-label network, so a campaign can serve on a grocer's own domain.

Our Instacart Ads playbook covers bidding structure in detail, and Instacart's advertising platform docs list current formats and eligibility.

Key takeaway: If your product has a UPC and sits on a grocery shelf, Instacart is not optional. It is the only one of the three that reports sales at SKU level natively.

5. Head to head: CPMs, ROAS, and typical minimums

Instacart wins on measurement, DoorDash on entry cost, Uber on reach. The table below is what we brief clients with before a first test.

Currency note: UK and Australian floors track similarly at roughly GBP 8 to 20 and AUD 18 to 38 CPM, though Instacart has no UK presence and Uber Eats carries most available inventory in both markets.

Key takeaway: Reported ROAS falls as audience size rises. That is an attribution artefact, not a media truth, and it is why we compare on incremental CAC.

6. Which commerce media network wins by DTC vertical?

Vertical decides the winner more than budget does. Match the network to where your basket actually gets built.

Key takeaway: Pick by shelf, not by CPM. The cheapest impression against the wrong basket returns nothing.

Want our team to run all three commerce media networks for you?

We build the test matrix, fund the right offers, and report incremental CAC instead of three incompatible dashboards. Book a free 30 minute audit and we will map the plan.

7. Creative specs and copy patterns that convert

Vertical 9:16 video under 15 seconds and 1:1 static at 1080x1080 cover most placements across all three networks, and pricing badges lift click-through 15 to 25 percent in our tests. Listing thumbnails need the product filling 70 percent-plus of the frame.

Copy patterns that work here:

Ship 6 to 9 assets per cell and refresh every 4 to 6 weeks. Our performance marketing services cover production for these formats, and the discipline matches our Meta Ads agency work: more variants, faster kills, offer-led messaging.

Key takeaway: Commerce media creative is packaging, not brand film. Show the product, the pack size, and the price.

8. Measurement and closed-loop attribution across the three

All three report closed-loop sales, but only Instacart and DoorDash resolve to SKU. Instacart reports attributed sales and new-to-brand rate per UPC, DoorDash reports order-level conversions with SKU detail in grocery and convenience storefronts, Uber reports order and trip-level conversions plus brand lift through partners.

Build the stack in three layers: platform reporting for in-flight optimisation only, incremental ROAS through geo holdouts of 15 percent of DMAs for 60 to 90 days, and offsite lift studies. All three support Nielsen and Circana style measured partnerships at higher spend tiers, and Instacart offers Nielsen MPA integration for CPG brands. Our clients see 6 to 14 percent incremental lift, well below what platform-reported ROAS implies.

Key takeaway: Treat platform ROAS as a bidding signal and a geo holdout as the truth. The gap between them is usually 40 to 60 percent.

9. How to allocate budget across Uber, DoorDash, and Instacart

Start with USD 12,000 over 90 days split across two networks, not three. Adding the third before you have a clean read multiplies dashboards without multiplying learning.

1. Month 1: commit USD 3,000 to the primary network for your vertical (Instacart for CPG, DoorDash for food and beverage) across search or Featured Products only. Target platform ROAS above 2.5 before adding formats.
2. Month 1: hold USD 1,000 for the second network as a reach test on display, single audience, single creative set.
3. Month 2: scale the winning network to USD 5,000 and add display or Sponsored Brands. Freeze the losing network rather than optimising it.
4. Month 2: hold out 15 percent of US DMAs as control and stop all other channel changes in those geos.
5. Month 3: add Uber Journey Ads at USD 2,000 to USD 3,000 only if attention-led creative exists. Do not resize Meta assets into this slot.
6. Month 3: compare exposed versus holdout revenue, keep networks clearing 8 percent incremental lift, and reallocate the rest to your best existing channel.
7. Quarter 2: negotiate. Past USD 25,000 spent, all three will discuss rate cards, co-op funding, and measurement credits.

Key takeaway: Two networks, one holdout, one written kill rule, and a scale decision at day 60. Commerce media pilots fail on governance, not on media.

Commerce media networks FAQ

Which commerce media network has the lowest minimum spend in 2026?

DoorDash and Instacart tie at the bottom, both opening self-serve campaigns at roughly USD 100, DoorDash on a per day floor and Instacart per campaign. Uber is the expensive entry because Journey Ads and homepage inventory sell managed, usually from USD 10,000 per campaign. For a first test under USD 5,000, start with DoorDash or Instacart.

How does commerce media attribution work compared to Meta or Google?

Commerce media attributes against orders the platform itself fulfilled, so the conversion is a settled transaction, not a pixel event stitched across domains. Instacart resolves to UPC, DoorDash to order and SKU, Uber to order and trip. That removes the signal loss Meta and Google fight, but it flatters the network by counting orders that would have happened anyway. Geo holdouts close that gap.

Should a small DTC brand test all three networks at once?

No. Under USD 15,000 quarterly, three networks means three underfunded tests and no statistical read. Run the network matching your shelf for 60 days at USD 3,000 per month, prove platform ROAS above 2.5, then add a second. We have never seen a sub USD 10,000 monthly budget produce a clean three-network comparison.

Ready to make your commerce media spend pay for itself?

Three networks, three attribution models, one budget. We will name the two worth testing in your category, build the holdout, and report incremental CAC next to your Meta and Google numbers.

Book a free 30 minute audit and we will map the 90 day plan with you.

Written by Nand Patel, Founder, New Public Media. More about us.