Google's own support documentation describes the Gmail tab that automated confirmations land in as holding messages that "may not need immediate attention." That tab is called Updates, and it sits beside the Primary tab, which is the one people actually read.

One line of product documentation says more about receipt delivery than most comparison charts do. A receipt is not a formatting problem, and the choice between paper, email, and phone is not really a choice about file types. It is a question of whether the receipt survives the trip from your till to your customer's attention.

Three methods dominate retail today, and each one breaks at a different point in that trip. This article walks each of the three gates a receipt has to clear, scores the three methods honestly against all of them, and names the limits of the approach we build ourselves.

A receipt has to clear three gates

Every receipt method, no matter how modern, runs the same gauntlet. Miss any one of the three and the receipt has failed, even if your POS reports it as sent.

Gate one is capture. At the moment of sale, does your store end up holding a valid, correctly-spelled way to reach this person? This is the gate most comparisons skip entirely, and it is the one that fails most often.

Gate two is delivery. Does the receipt land somewhere the customer will plausibly notice it, in a place they check without being prompted?

Gate three is retrieval. Three months later, when the customer needs that receipt for a return, a warranty claim, or an expense report, can they actually find it?

A useful fourth question sits underneath all three: after the sale, do you know who bought from you? A receipt that clears all three gates and still tells you nothing about your customer is a delivered document, not a relationship.

Printed slips: perfect capture, guaranteed loss

Paper wins gate one by not competing in it. There is nothing to capture, no address to type, no spelling to confirm. You print, you hand it over, and delivery is physically certain in a way no digital channel can match.

Then it falls apart. Thermal receipts are printed with heat rather than ink, and the image fades with time, sunlight, and heat. The retrieval gate is not merely difficult, it is on a timer. Most customers solve this by not trying: the slip goes in a bag, a pocket, or a bin within the hour.

Paper also carries a problem most store owners have never been told about. Thermal receipt paper has historically been coated with bisphenol A, and the EU restricted BPA in thermal paper to below 0.02% by weight from 2 January 2020 under Regulation (EU) 2016/2235, amending Annex XVII of REACH. The dossier behind that restriction, submitted by France, identified cashiers as a group at risk from handling receipt paper all day. Many suppliers substituted BPS, a closely related compound, which is why the coating is also the reason thermal receipts contaminate the paper recycling stream. We covered the full environmental arithmetic in the real environmental impact of paper receipts.

The deeper cost is the one that does not show up on any invoice. A printed receipt is the point where the relationship ends. You served a customer, and thirty seconds later you cannot say who they were, what they bought last time, or whether they will come back.

Emailed receipts: two failure points, one of them invisible

Email looks like the obvious upgrade. It is searchable, it is free to send, and every POS supports it. It also breaks at both of the first two gates, and the second break is one you cannot see.

Capture fails at the counter

An email address has to be spoken aloud, heard correctly, and typed correctly, by a cashier, with a queue forming. Every part of that sentence is a failure mode.

Addresses are hostile to spoken transmission. Dots, hyphens, and underscores have to be dictated. Homophones multiply: two spellings of the same name, or a domain the cashier assumes rather than confirms. Nothing validates the result, because a syntactically valid address that belongs to nobody accepts no correction. It simply bounces into a log your cashier never reads.

The friction is not evenly distributed either. It costs the customer more effort to give an email address than it costs them to decline, so the customers most likely to hand one over are the ones already inclined to hear from you. That skews your customer list toward people you had already reached.

Delivery fails in the inbox

Assume the address is correct. Gmail sorts incoming mail into five categories, and Google describes Updates as "automated confirmations, notifications, statements, and reminders that may not need immediate attention." A receipt is, by that definition, exactly an Update. Google is not misfiling it. Google is filing it correctly, into a tab designed for things that can wait.

Deliverability adds a second layer. Google's email sender guidelines require senders of more than 5,000 messages a day to Gmail addresses to authenticate with SPF and DKIM, publish a DMARC record, and hold their spam rate below 0.30%, with 0.10% recommended as a working target. A busy chain crosses 5,000 receipts a day without noticing. Meeting those requirements is domain and infrastructure work, not something a store owner can fix from the POS settings screen.

Then there is the part that makes email genuinely hard to manage: you cannot reliably tell whether any of it worked. Apple's Mail Privacy Protection hides the recipient's IP address and, in Apple's own description, "prevents senders from seeing if you've opened the email message they sent you." Open rates stopped describing readership. If your emailed receipts are landing in a tab nobody opens, your reporting will not be the thing that tells you.

Email does win gate three. A receipt sitting in an inbox is searchable years later, and that is a real advantage over paper. It just has to arrive first.

Receipts on the phone: the number you already have

The third method sends the receipt to the customer's phone number, as a message in the thread where they already talk to people. This is the approach Waraqa takes, and its advantage is concentrated almost entirely at gate one.

A phone number is built for spoken transmission. It is digits only, with no domain to guess and no punctuation to dictate. It is usually already in the POS customer record, because loyalty schemes, delivery, and callbacks all needed it long before receipts did. And when it is wrong, the customer usually finds out immediately, because the message does not arrive while they are still standing at the counter.

Gate two changes character too. Meta's documentation defines utility templates as messages "typically sent in response to a user action or request, such as an order confirmation or update," and they are delivered into the customer's ordinary message thread rather than a promotional bucket. There is no Updates tab in a messaging app. The receipt arrives where the customer's conversations are.

Retrieval works on two levels. The message stays in the thread, searchable like any other conversation. Customers who want more can install the Waraqa app and get every receipt from every participating store in one searchable place, along with their points balance. That step is optional, which matters more than it sounds: the receipt has already been delivered by the time the app is a consideration.

Where receipts on the phone have real limits

Any comparison that ends with "and the third option is perfect" is selling something. Three constraints are worth knowing before you switch.

You need the phone number in the sale. If your staff close sales without assigning a customer, no delivery method reaches anyone. This is a counter-habit change, and it is the single thing that determines whether a rollout works. The good news is that it takes a couple of seconds and your POS almost certainly has the field already.

Message templates are governed, not free-form. Meta reviews and approves templates, and it will recategorize a template as marketing if it contains promotional content. A receipt has to read like a receipt. In practice this is a guardrail rather than an obstacle, but it does mean you cannot bolt a sales pitch onto the bottom of a delivery confirmation.

And messaging reach varies by country. WhatsApp penetration is close to universal in some markets and partial in others. Somewhere in your customer base is a person who will want paper, and the sensible transition keeps printing on request while the default shifts.

The three gates, scored

	Printed slip	Emailed receipt	Sent to the phone
Capture at the till	Nothing to capture	Dictated, mistyped, often declined	Digits only, usually already on file
Delivery to attention	Physically certain	Sorted into a low-priority tab; deliverability rules apply	Lands in the customer's message thread
Retrieval months later	Faded or discarded	Searchable, if it arrived	Searchable in-thread, plus optional app
You learn who bought	No	Only if the address was captured	Yes
Verifiable that it landed	At handover only	Open tracking is no longer reliable	Delivery status per message

Read the table by column and the pattern is clear. Paper is excellent at the gate that costs nothing and fails at the one that matters later. Email is excellent at retrieval and weak precisely where it needs to be strong. Sending to the phone is the only column without a hard failure, and its weak point is a workflow habit rather than a structural flaw.

What changes when the receipt actually arrives

Delivery is not the goal. It is the precondition for everything a store actually wants.

Once a receipt reliably reaches a known customer, the sale stops being anonymous. You can see which customers return and how often, what a typical basket looks like, and who your best thirty customers are. None of that is available from a printed slip, and it is only partly available from email, because your emailed list is skewed toward the customers who were easy to reach.

It also makes loyalty work without new hardware. Points can attach to the receipt itself, so the customer sees their balance at the moment they are thinking about the purchase, rather than hunting for a card they left at home. Waraqa runs tiered points with your own tier names and automatic birthday rewards, where 100 points equals 1 unit of your local currency. If your POS has its own loyalty module, turn it off before going live so the two do not double-count.

And the economics move in the right direction. You stop buying thermal rolls, stop servicing the printer, and stop paying for a document your customer discards on the way out. Waraqa's first 100 receipts each month are free, and after that you pay only for what you use, with the per-receipt rate varying by country. The pricing page shows the current range.

The decision, in one sentence

Choose the method that clears the gate where your current one fails. If you print today, your failure is retrieval and customer knowledge, and either digital method beats paper. If you email today, your failure is capture and inbox routing, and switching the identifier from an address to a number fixes both at once.

The concrete next step is smaller than a platform migration. Ask your staff to attach a phone number to every sale for one week and see what share of transactions come back with one. That number is your real ceiling on any receipt method, and it costs nothing to measure.

When you are ready, connect your POS to Waraqa and your first hundred receipts are on us.