Skip to content

We move Life, We bring Life

Where trips end,commerce begins

When a customer sets a destination in Kakao Map, Navi, T or Bike, the shops and brands around the arrival point calculate what they can offer. The customer reviews only the offers that fit, and pays without leaving Kakao.

A fee is charged only after pickup and purchase are confirmed.

Arrival in 12min Yeongildae Beach
5-min walk
Yeongilman Coffee Iced americano · 2 min ₩1,500 off
Guryongpo Bake Two salt breads · 4 min ₩1,000 off
  1. 01 · Destination The customer enters the destination Setting a destination in Kakao Map, Navi, T or Bike fixes both the arrival radius and the expected time. No inference required.
  2. 02 · Pricing Each merchant sets its own offer Time remaining, walking distance, stock and margin determine the optimal discount rate, with a purchase rate estimated from the route.
  3. 03 · Choice The customer sees a shortlist Only the strongest offers within a five-minute walk are shown. Trip data is used for that trip and nothing else.
  4. 04 · Settlement The fee follows the purchase Impressions and clicks are not billed. Fees apply only to orders where pickup and purchase are both confirmed, and the merchant is settled through Kakao.

Defining the problem

Independent shops and brands face different constraints. Dangdo analyses the two separately and addresses each with its own mechanism.

Independent shops

Acquisition spend carries all the risk

A shop can sit on a busy corner or a major taxi drop-off and still see little footfall. When the spend comes first, nothing about the outcome is guaranteed once it is committed. Dangdo charges only on orders where pickup and purchase are confirmed.

Brands

Where the offer actually lands

Brands already commit substantial budget to events, discounts and giveaways. Without raising the discount at all, placing that offer on the customer's route improves what the spend returns.

Outer districts

What arrival data is worth

Customers take taxis out to areas well beyond the centre. For shops there, where the local advertising market is small, knowing that someone is on their way is a workable route to footfall.

How it works

The loop opens fifteen minutes before arrival and closes shortly after. What moves at each step is set out below.

  1. 01

    Receive destination and arrival time

    A destination set in Kakao Map, Navi, T or Bike is received together with its expected arrival time. Discovery is active only within fifteen minutes and a five-minute walk.

    One trip, one window, one offer

  2. 02

    Price an offer per merchant

    For each merchant our model calculates the optimal discount rate and estimates a purchase rate from the route and the time remaining. It selects the amount that maximises expected profit, capped by the campaign budget still available.

    Expected profit = conversion × contribution

  3. 03

    Surface the shortlist

    Two or three offers that fit the conditions, rather than a ranked list of ads. If the customer selects nothing, nothing is charged.

    No impression or click billing

  4. 04

    Confirm payment and pickup

    Payment is completed through Kakao Pay. Only orders with pickup and purchase both confirmed count toward performance, and the merchant is settled through Kakao.

    Cancelled orders are excluded

Signals that set the amount

Time to arrival Conversion falls when the offer lands too early or too late.
Walking distance Each additional minute on foot lowers conversion.
Stock and open seats How much is left to sell governs how much is worth spending.
Cost and margin The margin rate sets the ceiling on any offer.
Conversion by hour The same offer performs differently by hour.
Remaining budget Offers stop once the budget is spent.

Everything stays inside Kakao

No separate app to install. Setting the destination, reviewing the offer and paying all happen in Kakao services the customer already uses.

  1. Kakao Map · Navi The destination is set
  2. Kakao T · Bike The trip and its arrival time are committed
  3. Dangdo Shops and brands in range submit offers
  4. Kakao Pay Payment clears and settles

Because it integrates directly with systems Kakao already operates, customers never switch apps and merchants never adopt a new settlement process.

Pricing engine

The calculation a merchant runs fifteen minutes before arrival. Adjust the inputs and the model recomputes on the server.

4,500 KRW

List price. Below ₩3,000 the minimum-spend rule is not met.

62%

Contribution margin after cost of goods. It sets the ceiling on any offer.

12 min

Conversion peaks around twelve minutes and falls either side of it.

2 min

Conversion falls as the walk gets longer.

How busy the shop is
The engine's call

Make an offer

700 KRW
Discount
500 KRW
Points
200 KRW
Rider pays
4,000 KRW
Share of price
16%

Estimated conversion

Same axis
No offer shown 6.3%
Offer shown 13.7%
73 KRW
Added expected profit per trip
1,292 KRW
Cost per incremental order
1,810 KRW
Contribution per order

Merchant-funded model: the shop carries the discount, the points, a 5% Dangdo order fee and 2% payment processing. The conversion model and its constants are assumptions, to be replaced with measured values during the PoC. Amounts are rounded to the won.

Who pays what

The settlement flow for a single ₩4,500 iced americano order, on the brand-funded model. Amounts are proposal assumptions.

Rider pays via Kakao Pay Rider
+3,000
Brand-funded discount Brand
+1,500
Payment processing · 2% assumed Merchant
-60
Merchant settles at
4,440 KRW

Billed separately to the brand

Points granted
150
Dangdo performance fee · after pickup
1,000

In the merchant-funded model the shop funds the discount and points and pays Dangdo 5% of the order. In the brand-funded model the brand funds discount, points and the performance fee, and the shop is not charged twice. Either way the shop's real revenue settles through Kakao.

What counts as performance

  • An offer was selected, then at least ₩3,000 paid
  • Kakao Pay authorisation confirmed
  • Pickup completed in store after arrival
  • Cancellations excluded; one order per trip

These figures show only that a transaction took place. They cannot establish whether the customer would have come without the offer, which is what the controlled experiment is designed to settle.

See the eight-week design

Team

Developed in Pohang and tested in Pohang. The roles are set; we are finalising the names.

  • Product & lead Kim Jion Korea Advanced Institute of Science and Technology
  • Engineering Yoon Seokyoung Pohang University of Science and Technology