Paid for the click.Why pay again for the sale?
Proxi discounts only when needed.
A store-wide discount treats every shopper the same. Proxi decides, session by session, who gets an offer, how deep it goes, and who gets none because they were going to buy anyway.
- Paid for the click
- $2.10
- List price
- $120.00
- Same 20% for everyone
- −$24.00
Proxi decision- No offer
- Margin kept
- $24.00
What Proxi saw: third visit this week, browsing new arrivals at full price.
Example orders. Illustrative data only.
The leak is after the click.
Ads do their job and the traffic arrives. Then every visitor meets the same promotion, because the store cannot tell who needs it. Follow one promotion to see what that costs.
One promotion. 240 shoppers.
Each figure is a shopper who bought a $120 product during a 20%-off weekend.
Everyone got the same 20%.
The ring is the discount. $24 off, 240 times. That is $5,760 of margin spent to make these sales.
The shoppers were not the same.
Sort the shoppers by what it took to close the sale. Some needed the full offer. Some needed less. Some were going to buy anyway.
$2,592 bought nothing.
72 shoppers would have bought at full price. Another 72 needed half the discount. Together that is 45% of what the promotion cost.
Proxi makes the call before the offer is shown.
No offer where none is needed. The minimum effective discount where less will do. The full offer where it earns its cost.
Lifetime value hides it.
A shopper who only buys on discount can look like the best customer in the file: many orders, high lifetime revenue, thin margin. Proxi ranks shoppers on two things instead. How likely they are to pay full price, and the margin the store actually keeps.
Shopper A
$1,400Shopper B
$1,000Example: a $200 product that costs $100 to make and ship.
Proxi decides during the visit.
Proxi reads what a shopper is looking at and how much demand that product already has. Then it decides.
- No offerProxi held the discount back.
- 5% offerProxi showed an offer and set its depth.
- HoldoutThe store’s normal offer. This is the baseline every result is measured against.
What Proxi decides for each shopper.
No offer.
The shopper is on course to buy. Proxi stays out of the way and the store's normal experience applies.
The minimum effective discount.
The lowest discount that still closes the sale. The rest stays as margin.
The full offer, at the right moment.
On the product page, in the cart or at checkout, for the shoppers and products that need it.
It starts with the product.
Proxi scores every product on demand. The score reads page views that were not driven by ads, seasonality, price tier and list price. Products in high demand keep their price. Offers go where demand is soft, and shoppers are grouped by what they browse.
List prices never move.
Proxi manages offers. The price on the shelf is the same for every shopper.
People set the limits.
The team approves the discount ceilings. Proxi decides inside them.
Every offer is a test.
Each one is measured against traffic Proxi did not touch.
It starts without touching a single shopper.
Proxi earns its place in shadow mode first. Nothing changes on the storefront until the team has seen what it would have done.
Deploy where the business already runs
Proxi deploys inside the environment the business already operates. The data never leaves that tenant.
Connect the store
Proxi connects to a Shopify or BigCommerce store and reads the catalog, traffic and orders.
Run in shadow mode
Proxi proposes an offer for every session and shows none. Its engine learns from what actually sold and adapts.
Go live, with a holdout
Offers go live inside the approved ceilings. A share of traffic is left untouched as the baseline.
Read results against the baseline
Every headline number is shown next to the traffic Proxi did not touch, so the difference is Proxi's and nothing else's.
Estimate the margin Proxi can recover.
Enter a store’s figures. The model shows what a year of discounts costs, and what comes back if Proxi holds some of them, under three conversion scenarios.
GuardrailsMaximum discount 30%, conversion within 15% of baseline
Baseline is the traffic Proxi does not touch. At 15%, conversion on the orders Proxi acts on stays within 15% of that.
Change the underlying assumptionsIntervention rate 20%, gross margin 55%
Margin recovered per year. A model, not a forecast and not a measured result. Each bar shows margin recovered against discounts given. A full bar is half of the discounts given.
Check the calculation
Assumptions. Every sale is discounted at the average rate. Where Proxi intervenes, the full discount is removed and a share of those orders is lost. That share is the conversion drop. Margin recovered per dollar of list value = discount rate − conversion drop × gross margin.
Guardrails. If a case’s conversion drop is above the conversion limit, Proxi intervenes on fewer orders so the drop equals the limit. If the average discount rate is above the maximum discount, the other orders are trimmed to the maximum, with a conversion drop in proportion to the cut.
Questions?
Do shoppers see different prices?
No. List prices never move. Proxi decides whether to show an offer and how deep it goes. The price on the product is the same for everyone.
What does “no offer” mean for the shopper?
Proxi does not step in. The shopper sees whatever the store shows by default, including any standing coupon.
Does Proxi need to know who the shopper is?
No. It works from what a shopper browses in the session, so it applies to first-time and anonymous visitors as well as known customers.
Where do offers appear?
On the storefront: on the product page, in the cart or at checkout.
Where does the data live?
In the customer’s own tenant. Proxi deploys where the business already operates, and the data never leaves it.
Who stays in control?
People do. A person approves the guardrails: the maximum discount and the limit on conversion. Proxi decides inside them. As the engine learns, the team can let it run with less review. The guardrails stay in place.
Which platforms does it run on?
Shopify and BigCommerce.
Find out the margin you’re giving away.
Book a discovery session. We walk through the product, look at how the store discounts today, and show where Proxi can add value.