Every café owner knows the cheapest customer to win is the one who already came in. The hard part is reaching them again without paying for it. A delivery marketplace will happily re-target your past customer — and charge you a commission to do it. Email lands in a promotions tab. SMS costs per message and gets reserved for emergencies. Push notifications are the one channel where reaching a past customer costs essentially nothing per send — but they come with a real requirement most articles gloss over.
This guide is the honest version: how push actually works, how it really compares to email and SMS, what to send, and what it takes to run. It’s written for an independent café or restaurant owner, usually on Square, deciding whether push is worth it.
What a push notification actually is (and requires)
A push notification is a short message delivered straight to a customer’s phone — a banner on the lock screen, a badge on an app icon. No inbox, no phone number dialed, no carrier fee per message.
The catch, stated plainly: true mobile push requires a branded app the customer has installed, and an explicit opt-in to receive notifications. You can’t push to someone who hasn’t downloaded your app and tapped “Allow.” This is the single most important thing to understand before you get excited about push — it’s not a channel you can switch on for your whole customer list tomorrow. It’s a channel you earn as people adopt your app.
That requirement is also why push is so valuable once you have it. The audience is pre-qualified: every person on it chose to put your app on their phone and chose to let you message them. That’s a fundamentally warmer list than an email blast to addresses collected at a Wi-Fi gate.
Push vs. email vs. SMS: the honest comparison
Each channel has a different cost, reach, and tone. None is strictly best; they do different jobs.
| Channel | Cost to send | Requires | Typical open/visibility | Best for |
|---|---|---|---|---|
| App push | Effectively free per message | A branded app + opt-in | ~7.8% avg open, wide range | Timely, relevant nudges to engaged customers |
| Low (often free at small scale) | An email address | Promo opens often in the teens to ~20% | Longer messages, receipts, newsletters | |
| SMS | Per-message fee | A phone number + opt-in | Very high open, but intrusive | Rare, high-value or transactional alerts |
Benchmarks are cross-industry aggregates, 2025. Your café’s numbers depend on your list quality, timing, and relevance — treat these as directional.
A few honest points the table compresses:
- Push open rates are modest on average (~7.8%) but enormously variable — from under 1% to over 50% — driven almost entirely by relevance and timing. A “your order is ready” push gets opened; a generic “we miss you” blast doesn’t.
- Opt-in rates differ sharply by platform. Android users opt into push at much higher rates than iOS users (broad benchmarks put Android opt-in around 90%+ and iOS closer to the mid-40s%). So your reachable push audience skews Android, and you should keep email and in-app messaging as complements, not assume push reaches everyone.
- SMS opens are high but the channel is expensive and easily resented. Save it for the rare message that truly warrants interrupting someone.
The realistic posture: push is your low-cost, high-frequency-tolerant nudge channel for engaged app users; email is your catch-all for everyone with an address; SMS is your break-glass option.
What to actually send (and what to never send)
The fastest way to kill a push channel is to blast it. The opt-out is one tap, and once it’s gone it’s gone. Earn the attention.
Send these:
- Transactional, expected messages. “Your order’s ready for pickup,” “your card was charged,” “your loyalty reward unlocked.” People want these, and they keep your app’s notifications switched on.
- Timely, relevant offers. A 2pm “slow afternoon: any pastry half off till 4” to people near your shop. A new seasonal drink the week it launches. Relevance and timing do all the work.
- Win-backs with a reason. “Haven’t seen you in a while — here’s a free drip on your next visit.” Tied to actual lapsed-customer data, not sent to everyone.
- Loyalty milestones. “One more visit till your free coffee.” This pairs directly with a loyalty program that turns first-timers into regulars — push is how the loyalty program speaks.
Never send these:
- Daily generic “good morning!” blasts with no offer or reason.
- The same message to your whole list when only a segment is relevant.
- Anything that reads as desperate or spammy. Each bad push raises the odds of an opt-out, and opt-outs are nearly permanent.
Push is a privilege, not a megaphone. Every message either earns the next open or spends it.
The economics: why push beats renting your audience
Here’s where push earns its keep. Once you have an app and an opted-in audience, reaching every one of them costs essentially nothing per send. Compare that to the alternatives for re-reaching a past customer:
- A delivery marketplace will re-surface you to that customer — and take 15–30% commission on the order it drives.
- SMS costs a few cents per message, which adds up across a list and a year.
- Paid social charges per impression to reach people who already know you.
Push, by contrast, is a fixed-cost channel: you pay for the app platform once, and the marginal cost of telling 2,000 opted-in regulars about today’s special is zero. That’s the same “own the channel instead of renting it” logic behind marketplace vs. direct ordering and who owns your customer data. Push is the clearest expression of owning the relationship: a direct line to your customer that no platform meters.
The honest counterweight: that fixed cost is only worth it if your app gets adopted and your pushes stay relevant. A branded app with twelve installs and spammy notifications is worse than a good email list. Push rewards cafés that earn an audience and respect it.
What it takes to run push at a café
To send push, you need three things in place:
- A branded app customers install — on both iOS and Android, since your reachable audience splits across them.
- An opt-in flow that asks for notification permission at the right moment (after a first good experience, not on first launch).
- A way to segment and schedule — so you can send the 2pm offer only to nearby lapsed customers, not your whole list.
Building and maintaining all of that in-house is a real project. The pragmatic path for a Square café is a platform that ships the app, the opt-in, and the messaging tools together on top of your existing POS — so push, ordering, loyalty, and gift cards all share one customer record.
That’s what Tany does: a branded iOS and Android ordering app with push notifications, self-running loyalty, eGift cards, and web ordering, built on your existing Square POS and live in about a day for $99 CAD/month per location. The push channel isn’t a bolt-on; it’s the app talking to customers who chose to be there.
Push notifications won’t save a café on their own, and they’re not a switch you flip on an existing customer list. But for the customers who put your app on their phone, push is the cheapest, most direct way to bring them back — a channel you own outright instead of renting one nudge at a time.