Skip to content

Discount incrementality for ShopifyDiscount incrementality for Shopify

Which discounts actually create a sale?

Invisible Offer measures which discounts change purchasing behaviour — then helps you stop paying for the ones that don’t.Measure which discounts change purchasing behaviour — and stop paying for the ones that don’t.

Measure my discount waste

Don’t change a single discount until you have evidence.

  • Starts in Invisible ModeInvisible Mode
  • No shopper changesNo changes
  • Control-testedControl-tested
LIVE EXPERIMENTJust now
Running shoesSize 9 / black
$180
Your normal offer10% OFF−$18

Did this discount actually create the sale?

CONTROL10% OFF↓Purchase
Randomized comparison
INVISIBLENO DISCOUNT↓Purchase
Same purchase.$18 more margin. This discount didn’t change the outcome.
Illustrative demo data — not merchant performance.

Example figures from a demonstration store.

$42,180Discount spend measured
$16,420Non-incremental spend
+38.90%Discount waste rate
$11,840Estimated incremental margin

Example experiment · illustrative

Prediction tells you what might happen.

Incrementality tells you what actually changed.

The counterfactual

Would they have bought anyway?

You can see who purchased with a discount. You can’t see what would have happened without it.

WHAT YOU SEECustomer↓10% OFF↓Purchase ✓
?Would they have purchased
without the discount?

That’s the $18 question.

Randomized test

CONTROL10% OFF↓Purchase?
INVISIBLENO OFFER↓Purchase?

Invisible Offer creates the counterfactual your Shopify dashboard can’t show you.

Your economic baseline

How much are you spending on discounts you haven’t measured?

Enter a few numbers to estimate the discount spend whose incrementality you don’t yet know.

$
%
%
%

An assumption, not a measurement. Invisible Mode replaces it with your own control group.

You spend

$17,640/ month on discounts

If 40% are non-incremental

$7,056/ month may not be changing a single purchase.

Estimate based on your inputs — not a guarantee.

Replace this estimate with evidence That’s an assumption. Invisible Mode measures it.

The method

Measure first. Automate second.

A controlled path from behaviour to evidence to protected margin.

01

Watch

Understand the shopper. Behaviour becomes structured shopper state.

  1. Product
  2. Reviews
  3. Size guide
  4. Cart
  5. Shipping
  6. Checkout
Structuring behaviour…0 / 6
02

Test

Test the counterfactual with the same type of shopper and different treatment.

ELIGIBLE SHOPPERRandomized
CONTROL10% OFF
INVISIBLENO OFFER
Same type of shopper. Different treatment.
03

Learn

Learn what actually changed.

CONTROLPurchase−$18NON-INCREMENTAL
INVISIBLEPurchase$0Same outcome
CONTROLPurchasevs.INCREMENTAL: keep it
04

Protect

Protect margin where the evidence is strong.

SHOPPER↓Behaviour↓Historical evidence + current intent↓NO OFFER+$0 retained
We don’t ask you to trust the AI.We measure what actually happened.

The learning loop

Every experiment makes the next decision better.

01Behaviour
02Decision
03Experiment
04Outcome
05Learning

Invisible Offer doesn’t just predict. It learns from what your customers actually did.

Last 30 days · demo storeDEMO STORE
$84,260Discount spend measured
$51,420
Incremental discount spend
$32,840
Non-incremental spend
39.0%
Waste rate
+$21,600
Incremental margin protected

Discount spend

Incremental61%
Non-incremental39%
Conversion guardrailΔ +0.02pp
CONTROL 3.41%INVISIBLE 3.43%
Illustrative demo data — not merchant performance.

Economics evidence

Less discount spend. Conversion held.

Incremental and non-incremental spend stay visible together, with conversion treated as a guardrail.

Run Invisible Mode No customer-facing changes during analysis.

Restraint is the product

The best offer is often no offer.

Most discount tools optimise what to offer. Invisible Offer first measures whether the offer changes anything at all.

Most discount tools optimise:“What should we offer?”
Invisible Offer measures first:“Did the offer change the purchase?”

If it doesn’t change the purchase, don’t pay for it.

Start with evidence

Don’t change a single discount.

Run Invisible Offer silently first. Measure where your discount spend is actually incremental before you automate anything.

  1. 01Connect ShopifyRead-only install.
  2. 02Run Invisible ModeNo customer-facing changes.
  3. 03Establish baselineUnderstand current discount behaviour.
  4. 04Run controlled testsMeasure incrementality.
  5. 05Protect marginAutomate only where evidence supports it.
Measure my discount waste

Nothing changes for shoppers until you decide.

FAQ

Clear answers before you connect your store.

How is this different from a personalized discount app?

Personalized discount tools decide which offer to show. Invisible Offer starts one step earlier: it measures whether showing an offer actually changes purchasing behaviour.

How do you know a discount was unnecessary?

We don’t infer that from a purchase alone. Eligible traffic is compared using controlled cohorts, allowing the system to estimate the incremental effect of the intervention.

Does AI decide what happens to every shopper?

Semantic shopper signals inform the decision, but merchant-defined policies and experiment rules control what actions are allowed.

Will this immediately change my store?

No. Invisible Mode runs silently first. Nothing changes for shoppers until you have evidence and decide to enable live protection.

Will Invisible Offer remove all discounts?

No. Discounts shown to be incremental are protected. The product targets spend that does not change the purchase.

The next economic question

You know what you spend on discounts. Now find out what they actually buy you.

Measure the incremental effect of your discount spend before you automate anything.

Measure my discount waste
  • Shopify
  • Invisible Mode
  • No customer-facing changes