You can run a thousand delivery orders a month and still not be able to name a single customer. That is not a glitch — it is the business model. The most expensive line item on a marketplace invoice is not the commission. It is the relationship you never get to keep.
So before you argue about percentages, answer the more important question: when an order comes in, whose customer is it?
The 40-second answer: who owns the data in each model
On a delivery marketplace (DoorDash, Uber Eats, Grubhub), the platform owns the customer relationship. You receive the order details needed to fulfill it and some aggregated, mostly anonymized insights — but not a portable customer record you can freely market to. With direct ordering through your own website or app tied to your POS, you are the data controller: the contact, the full order history, and the loyalty all belong to you, and you can bring that customer back yourself.
Both models have a place. The mistake is not knowing which one you are operating in — and quietly building your whole future on the one where you own nothing.
Does DoorDash give restaurants customer data?
Partially, and the kind of data is the whole story.
For Marketplace orders (someone finds you in the DoorDash app), the merchant gets what is operationally necessary to prepare and hand off the order — the customer’s name, delivery address, and phone number — plus a layer of aggregated analytics in the Merchant Portal. DoorDash’s own customer-analytics tooling shows you what percentage of your customers are new, occasional, or frequent, and visual heatmaps of which zip codes drive the most orders (DoorDash Merchant Learning Center). That is genuinely useful for understanding demand. What it is not is an exportable list of customers with emails that you can drop into your own marketing tool and own forever.
This aggregated-not-identifiable posture is by design, and it is industry-wide. When the National Restaurant Association and the major delivery companies agreed on public-policy principles for third-party delivery, the data commitment was that platforms should offer restaurants access to anonymized information about their orders — when orders were placed, where they originated, and whether they came from a new or existing customer (National Restaurant Association). Anonymized is the operative word.
There is an important wrinkle that trips up a lot of owners. DoorDash also sells Storefront / Online Ordering — a commission-free ordering page powered by DoorDash that lives on your website. It feels like “your” channel. But DoorDash’s Storefront merchant terms are explicit: the merchant receives “sufficient information to prepare the order but will not own such Customer Data,” and may not use that data for remarketing or to build user profiles (DoorDash Storefront Merchant Terms of Service). “Direct” and “owned” are not the same thing. Always read the data clause.
Can I email customers who ordered through Uber Eats?
Not the way you can with your own list.
Uber Eats has leaned into giving merchants more visibility and marketing capability — it has publicly committed to “best-in-class insights” on how customers interact with a storefront, and in early rollouts said merchants graded 72% of those insights as useful (Restaurant Business). Inside Uber Eats Manager, you can build offers and campaigns and target groups like “all customers in your delivery zone” or “new customers that haven’t ordered from your restaurant” (Uber Help).
That is real value — but notice where it happens. You reach those customers through Uber’s app, on Uber’s terms, within Uber’s platform. You do not get their email addresses to load into Klaviyo or Mailchimp and own independently. The relationship — and the channel back to the customer — stays with the platform. Uber has also moved to limit data exposure generally (its “View as Delivery Person” transparency feature lets customers see exactly what couriers can access at each stage), which is good for privacy but underscores the direction of travel: the platform, not the restaurant, is the steward of the customer (Restaurant Dive).
Let’s be fair: what marketplaces actually do give you
This is not a hit piece on delivery apps. For an independent café or restaurant, marketplaces provide things that are genuinely hard to replicate:
- Reach and discovery. Millions of hungry people are already in those apps. New customers find you there who never would have otherwise.
- Aggregated analytics. New-vs-returning splits, peak times, and geographic heatmaps help you read demand.
- In-platform campaign tools. Offers and promos that can win back lapsed orderers or acquire first-timers — run inside the app.
- Logistics. Drivers, tracking, and support you don’t have to staff.
The honest framing is not “marketplaces bad, direct good.” It is: marketplaces are a customer-acquisition channel, not a customer-ownership channel. The smartest independents treat them exactly that way — using them to find new diners, then giving those diners a reason to come back through a channel the operator owns.
The comparison, row by row
Here is what each model actually gives you. Entries are kept honest — where a marketplace provides something, it says so.
| What’s at stake | Delivery marketplace (DoorDash / Uber Eats app) | Direct ordering (your site/app + your POS) |
|---|---|---|
| Customer contact info | Name + delivery address + phone for fulfillment; no exportable email list for your own use | Full contact (name, email, phone) saved to your customer directory |
| Order history | Aggregated insights (new vs. frequent, zip heatmaps); not a per-customer record you own | Complete per-customer order history, yours to keep |
| Re-market via push/email | Only inside the platform’s offer/campaign tools; can’t message off-platform | Yes — your push, your email, your SMS, on your schedule |
| Loyalty ownership | Platform-run programs; the loyalty relationship sits with the app | Your loyalty program, running off your POS and customer record |
| Who controls pricing & menu | Platform sets the rules; commission shapes your margins | You set pricing and menu; no per-order commission |
| Reviews | View and respond inside the platform; ratings live there | You own first-party feedback and where it’s displayed |
| Who the customer thinks they bought from | Often “DoorDash” / “Uber Eats” — the platform owns the brand moment | Your brand — they bought from you |
The single most quietly damaging row is the last one. On a marketplace, the customer’s relationship is frequently with the app, not your restaurant. That is the relationship you are renting.
A framework: the Customer Ownership Ladder
Not every order sits in the same place. Use this ladder to grade where each one lands — and to see where you want them to climb.
Rung 1 — Anonymous marketplace order. You filled an order, but you cannot identify, reach, or bring this person back on your own. They are the platform’s customer. Most third-party delivery orders live here.
Rung 2 — Known but rented. You can see some info and maybe reach them through the platform’s tools (an Uber Eats offer, a DoorDash promo), but you do not hold the contact independently and cannot move it. Better than anonymous — still not yours.
Rung 3 — Captured. You now have the customer’s consent and contact details in your system (a direct order, a loyalty sign-up, an email opt-in). You can reach them for free, forever.
Rung 4 — Owned regular. They have a profile, an order history, and a loyalty balance with you, and you actively bring them back with push and rewards. This is the asset that compounds — and the one that survives if a platform changes its fees or policies tomorrow.
The job is not to abandon Rung 1. It is to build a path that moves people up the ladder — from a rented order to an owned regular.
Quick scorecard: “Do I actually own this customer?”
Score one point for each yes:
- Can you contact this customer without going through a third-party platform?
- Do you hold their order history in a system you control?
- Can you enroll them in your loyalty program?
- Did they experience the order under your brand, not the platform’s?
- If that platform vanished tomorrow, could you still reach this person?
4–5: Owned. 2–3: Rented with a foothold — push them higher. 0–1: You’re renting; start capturing.
The number that should change your mind: rented vs. owned lifetime value
Commission math gets all the attention. The lifetime-value gap is bigger — and it is the part the marketplace model quietly costs you. Let’s model it with round CAD figures.
The rented customer. Someone orders a $30 lunch through a delivery app. After a roughly 25% commission, you net about $22.50 on that order. Fine. But because you cannot reach them, your expected future value from this specific person is close to one order. To get a second order, you typically pay commission again — or hope the algorithm resurfaces you. Call the durable value of a rented customer ≈ $22.50, because you can’t reliably bring them back yourself.
The owned regular. Now imagine that same person orders through your own channel, lands in your customer directory, and joins your loyalty program. Repeat customers spend meaningfully more than one-timers — by one widely cited estimate, 67% more (ChowNow). Say they come back 4 times a month at a $25 average ticket. That’s $100/month, or $1,200 a year in top-line — and you can prompt each return with a free push or email instead of paying to re-acquire them. Even modeled conservatively over a single year, an owned regular is worth tens of times a rented order.
The lever behind that gap is retention, not acquisition. The classic finding — increasing customer retention by just 5% can lift profits by 25% to 95% (Bain & Company / Harvard Business Review) — only works if you can reach the customer to retain them. On Rung 1, you can’t. That is the real cost of renting: not the commission on the order you got, but the dozen owned orders you’ll never get.
How to convert rented marketplace customers into owned regulars
You don’t have to quit the apps. You have to stop letting them be the only place your customers live. A practical checklist:
- Stand up a direct ordering channel tied to your POS. Order-ahead pickup and your own web ordering, running off the system you already use, so every order creates a customer record you own. (Here’s how to set up a commission-free ordering channel.)
- Give people a concrete reason to switch. A loyalty program, a pickup-only perk, a slightly better price than the app, or a free item on first direct order. Make “order direct” the obviously smarter choice.
- Capture consent at the moment of intent. Collect email/phone with permission to message at sign-up or checkout — so you can legally re-market (more on the rules below).
- Use the apps for what they’re good at. Let marketplaces do discovery and acquisition. Then route repeat business to your owned channel. Cap third-party at a share of volume, not the whole thing.
- Insert your brand into the marketplace order. An in-bag card or sticker — “Order direct next time and earn rewards” with your QR code — quietly moves Rung 1 customers toward Rung 3. (Marketplace terms restrict using the platform’s name/logo in your outreach, so brand it as yours.)
- Bring them back on a schedule you control. Once they’re owned, a push notification or email costs you nothing and turns a one-time order into a habit.
Step one is the unlock. Until each order produces a customer record you control, every other tactic is downstream of a leak.
Where Square-based direct ordering changes the ownership question
If you run Square, the data-ownership picture flips in your favor — because of who the controller is. Details a customer shares when they order through your own Square-powered channel are saved in your Square Customer Directory, which you can use with Square Marketing or export for use with another tool, and your customer data is portable via Square’s public APIs (Square Support). In plain terms: Square acts as your processor, and you are the controller of the relationship. (In Canada, PIPEDA governs how you collect, use, and disclose that personal information; in the US, equivalent norms and state laws apply. None of this is legal advice — but the ownership direction is clear: on your own channel, the customer is yours.)
This is the worked example of owning the channel rather than renting it. A tool like Tany gives an independent café or restaurant on Square its own branded iOS and Android ordering app — order-ahead pickup, self-running loyalty, push, eGift cards, and web ordering — where the customer record and their order history are yours because they’re tied to your Square account. Loyalty and push run off that owned record, so a first order can actually become a regular. That is the entire point of climbing the ladder: not to fight the marketplaces, but to build the asset they will never hand you.
Marketplaces are worth using. Just know what they are: a place to meet customers, not a place to keep them. The keeping — the contact, the history, the loyalty, the next ten orders — is what you build on a channel you own. For the flip side of this trade-off in pure dollars, see the real cost of delivery apps.
The order you ran today is already gone. The only question that matters is whether you can get the next one — for free, on your terms, because the customer is finally yours.