The promotion you did not have to write
Every owner knows which afternoons are dead and means to do something about it. The something involves picking an item, deciding a discount, writing the copy, making an image, sending it somewhere, and then working out whether anybody who turned up came because of it. Six jobs, and the promotion is worth less than the evening it takes.
So it does not happen, or it happens once with a code nobody could find and gets filed under things that do not work for a business this size.
This is the version where the promotion writes itself, the discount is attached to the item rather than to a code, and the customer who taps it lands in a conversation with the thing already in front of them.
Book a Free DemoThe offer is attached to the item, not printed on a code
Everyone has run the coupon code version of this and watched it fail the same way. The customer cannot find the code, or types it wrong, or gets to the counter and your staff have to key it in and guess whether it applies to that item.
An offer here is structured data attached to a specific item. There is no code for the customer to remember and nothing for staff to key in. The discount is a property of the offer and the item, so it either applies or it does not, and neither the customer nor your team has to adjudicate anything.
The arithmetic is bounded on purpose. The discount applies to the base price only, modifier surcharges are never discounted, and a line can never go negative. Money is handled in integer cents throughout.
- An offer is structured data on a specific item, not a coupon code
- Nothing for the customer to remember and nothing for staff to key in
- Discount applies to the base price only, never to modifier surcharges
- A line can never go negative, and money is handled in integer cents
They tap the post and land in the chat with the item already loaded
A customer sees a promoted item in a post or a message and taps it. They land in the chat with that exact item already in front of them, ready to order. Not on a menu page they have to search, and not in a generic conversation where they have to explain which thing they meant.
Then the part that decides whether any of this works: when the order confirms, the discount is applied automatically. No code, no reminder to mention it, and no moment where the customer wonders whether they got the deal.
Compare that with the usual arrangement, where the promotion and the ordering system are different products that have never met, so the customer arrives holding a code the ordering flow does not recognise.
- Tap a promoted item and arrive in the chat with it already loaded
- The discount applied automatically when the order confirms
- No code to find, no code to type, no code to explain at the counter
- Ordering and the offer are the same system, so they cannot disagree
It notices the quiet Tuesday before you do
Most software reports a problem. This proposes the fix and waits for you to say yes.
The platform reads your real booking data and compares your quiet periods against your busiest, so a recurring quiet day or hour is detected from what actually happened rather than from a feeling. When it finds one, it proposes a promotion to fill it. It does the same for a group of customers who have gone quiet, and for recent completed visits with no review request.
The detection is deterministic and reads real data. It is not a model guessing at what your business might need this week, which is the difference between a suggestion you act on and one you learn to ignore.
You approve, and approving routes straight into the real action: the promo is generated, the win-back messages are released, the review requests are created. Nothing fires without that approval, and there is at most one open suggestion of each type, with a cooldown, so it never becomes noise.
- A recurring quiet day or hour detected from observed bookings
- A promotion proposed to fill it, a win-back proposed for a lapsed group
- Recent visits with no review request proposed for sending
- Deterministic detection on real data, not a model guessing
- You approve and it acts. One open suggestion per type, with a cooldown.
Who it goes to, and how you know it worked
A promotion is only as good as the list it reaches. Distribution goes to real behavioural segments built from what customers actually did, over email, SMS, and WhatsApp. Opted-out customers are excluded from every segment, enforced in code.
Measurement starts at the click on a tracked link the platform owns and runs through to the order. Because that order sits on a real customer record, the promotion is followed through to revenue rather than stopping at a send count. See how attribution ties a click to a paying customer.
How it fits the rest of the platform
A promotion is the sharp end of everything else. The item comes from the menu that Clover or Square holds, the order it produces runs through the ordering flow, the customers it reaches are segments of your own list, and the quiet slot it fills was found in your real booking data. Broadcasts, segments, and the measurement behind all of it live on marketing and attribution, and the win-back side of the suggestions engine is covered on retention.
Questions owners ask
Do I have to write the promotion?
No. Promotions and daily specials are generated from your real menu and real business data. You approve what goes out, and nothing sends without that approval.
Is there a coupon code my customers have to remember?
No. The offer is structured data attached to a specific item rather than a code. The discount is applied automatically when the order confirms.
What exactly gets discounted?
The base price only. Modifier surcharges are never discounted and a line can never go negative. Money is handled in integer cents throughout, so totals stay exact.
What happens when a customer taps a promoted item in a post?
They land in the chat with that item already loaded, ready to order, and the discount applies automatically when the order confirms.
How does it know which afternoons are quiet?
It compares your quiet periods against your busiest using observed booking data, deterministically, rather than guessing.
Will suggestions start sending things on their own?
No. A suggestion is a proposal that waits for you, and approving it routes into the real action. There is at most one open suggestion of each type, with a cooldown.
See it working on a business like yours.
Book a free fifteen minute demo. We will show you where customers are slipping away today and what closing that gap looks like for your shop.
Book a Free Demo