All answers

The Prevailing
Answers for
Business model:
Business model, pick up to two

Should we test discount codes and promotions for incrementality?

The PrevailingPublished Oct 29, 2026

Our answer

Yes. Test every discount you run often or that costs real margin, starting with the costliest. Withhold it from a random 10 to 20% of eligible customers and judge profit per customer, not redemptions.

Confidence
Medium
Strongest proof
4 official rules
2 datasets, 2 practitioners, 11 opinions or claims
Sources
13 sources
19 statements

New here? Take the 1-minute tour

Why this is the answer

3 reasons and our judgment

A code records who used it, not who bought because of it (BlueAlpha, a mix modeling and testing vendor, and Uniqodo, a promotion code platform). Codes are redeemed mostly by people already close to buying, so redemption reports overstate what an offer added.

A holdout shows what the offer added. Talon.One (promotion and loyalty software) advises withholding a promotion from 10 to 20% of the eligible audience and scaling it only when the holdout shows positive incremental contribution after discount cost.

Sale periods read differently from normal weeks. Common Thread Collective (CTC) measured incrementality 30 to 50% away from normal weeks in its sale-period lift tests, in either direction, so a sale's platform numbers can't be trusted on their own.

Our judgment

  • Test first the discounts you run repeatedly or that give away the most margin. That is where a wrong read costs most.
  • For offers sent to known customers, hold the offer out from a random 10 to 20%. Keep the split across every flow and channel that could deliver it, and judge on profit per eligible customer through a tail of at least seven days after the offer ends.
  • Read a public sitewide sale against matched regions or a baseline of past weeks, and treat the read as weaker than a holdout.
  • Decide the offer's future on the holdout, not the codes. Keep codes or offer links to see who redeemed, keep the offer's dates and promotion attached to the result, and retest when the offer, audience or season changes.

Our judgment, built from Uniqodo, Talon.One, Made With Intent, Quikly, BlueAlpha and Common Thread Collective: our reading of these sources, not a finding from any one of them.

What the sources say

19 statements from 13 sources
  • ShopifyUndatedSupports

    Shopify Experiments test theme, checkout, catalog or discount changes against a control, by default with a 50/50 traffic split and an end date 90 days after the start.

    “By default, a new Experiment uses a traffic split of 50% control and 50% treatment, and an end date 90 days after the Experiment starts.”

    Read it on help.shopify.com

  • ShopifyUndatedCaveat

    Shopify's built-in Experiments need the Grow plan or higher and cannot test headless checkouts or Liquid template changes. During a test, ads and feeds may show a different offer from the store.

    “Product ads, listings, feeds, and third-party apps might not display the same offer.”

    Read it on help.shopify.com

  • LebesgueNov 12, 2026Caveat

    Lebesgue, an analytics seller, says in BFCM data from nearly 1,000 stores, sales grew 1.95 times at about 15% off and no more at 40% or 50% off, while gross profit grew 2.05 times at 15% off against 1.75 at 40%.

    “stores at about 40% off and 50% off or more grew 2.0x, no more than stores at 15%.”

    Read it on lebesgue.io

  • Common Thread CollectiveOct 22, 2026Supports

    CTC says a ROAS target from a Black Friday lift test may not fit January; keep the test dates and promotion attached so a seasonal result does not become a permanent budget rule.

    “Keep the test dates and promotional context attached to the result so a seasonal finding doesn't become a permanent budget assumption.”

    Read it on x.com

  • Taylor Sicard ConsultingOct 21, 2026Supports

    Taylor Sicard, a DTC consultant, suggests checking a past sale by comparing customers first acquired in it, split by discount use, with a normal month's cohort on gross profit per customer over your repeat cycle.

    “Compare gross profit per customer, not revenue, for each cohort over the same number of months.”

    Read it on taylorsicard.com

  • Common Thread CollectiveOct 20, 2026Supports

    CTC's Luke Austin says a handful of its lift tests in promotional periods showed incrementality 30 to 50% different from normal periods, in either direction, which should change BFCM ROAS targets.

    “it's 30, 40, 50% variation in, in some direction against your normal”

    Read it on commonthreadco.com

  • Talon.OneOct 19, 2026Supports

    Talon.One says scale a promotion only when a randomized holdout shows positive incremental contribution after discount cost. In its hypothetical example, a discount adds 50 orders but cuts contribution by $3,750.

    “A promotion should not scale because redemption rises.”

    Read it on talon.one

  • Growth SuiteOct 18, 2026Supports

    Growth Suite, which sells targeted-offer software, says a few genuine events a year is the ceiling for storewide sales, because frequency trains customers to wait. It says to measure the weeks after each campaign.

    “The campaign's true cost shows up in the full-price dip that follows it.”

    Read it on growthsuite.net

  • ShopifyOct 2, 2026Supports

    Shopify added discounts to Rollouts in October 2026, so stores can test offers against existing ones and compare conversion before launching widely. Existing eligibility rules still apply.

    “Test discount changes against your existing offers and review conversion results to help decide which offer to keep.”

    Read it on changelog.shopify.com

  • IntelligemsSep 8, 2026Disagrees

    Over Labor Day 2026, 34% of 3,000+ Shopify brands ran an offer. Discounters' revenue lift fell to 22% from 31%, brands without offers gained 3%, and 20%+ discounts lifted revenue 89%. Revenue, not profit.

    “a token discount costs you margin and buys you nothing. Go meaningful or skip it.”

    Read it on intelligems.io

  • UniqodoSep 3, 2026Supports

    Uniqodo, which sells promo code software, says discount reports can't show whether an offer created new sales, and last click can give a code 100% of a sale. Start holdouts with the costliest promotions.

    “Even a simple holdout on one campaign can provide more useful evidence than a long series of redemption reports.”

    Read it on uniqodo.com

  • QuiklyAug 19, 2026Supports

    Quikly, which sells promotion software, says to split email holdouts by customer ID before the send and keep them out of flows with the same offer; Klaviyo's campaign revenue can't prove a send caused a sale.

    “Klaviyo's campaign revenue can describe what happened after a send, but it can't by itself prove that the send caused the purchase.”

    Read it on hello.quikly.com

  • Digital AppliedMay 29, 2026Supports

    Digital Applied, an agency, says only a holdout shows whether a promotion made new sales, and gives the break-even math: a brand at 48% gross margin running 15% off needs about 50% more units to hold gross profit.

    “Without holdout groups, you cannot answer whether a promotion generated truly new sales or simply cannibalized purchases that would have happened anyway.”

    Read it on digitalapplied.com

  • Digital AppliedMay 29, 2026Caveat

    Digital Applied, which sells promotion audits, shows 20% off needs 40% more units to break even at a 70% margin and 133% more at 35%, and relays 20 to 60% cannibalization for broad coupons.

    “Reach for free shipping or bundles instead of deep percentage-off.”

    Read it on digitalapplied.com

  • Made With IntentMay 22, 2026Supports

    Made With Intent suggests suppressing a random 10% of recipients before a promo email, then comparing revenue per recipient, including a seven-day tail, adjusted for the margin cost of the discount.

    “Before your next promotional email, suppress a random 10% of eligible recipients.”

    Read it on madewithintent.ai

  • BlueAlphaMay 12, 2026Supports

    BlueAlpha, which sells incrementality testing, says direct mail promo codes measure redemption, not cause: they take credit from other channels and miss buyers who skip the code. A holdout should decide budget.

    “Once a holdout has run, the ratio between the two becomes a calibration factor you can apply to subsequent drops without retesting every cycle.”

    Read it on bluealpha.ai

  • Talon.OneMay 11, 2026Supports

    Talon.One, which sells promotion software, says redemption rates do not show whether a coupon changed behavior. Hold the offer back from 10 to 20% of the audience and track sales 30 to 180 days after it ends.

    “A promotion generating 30% sales lift may represent zero net new revenue if customers bought earlier and then went quiet.”

    Read it on talon.one

  • IntelligemsJan 20, 2026Supports

    Intelligems, which sells offer testing, says brands running promotions monthly or more see full-price purchase rates fall within 6 to 12 months, while quarterly or tentpole-only brands fare better. It shows no data.

    “If 15% converts nearly as well as 25%, you've found margin you can protect”

    Read it on intelligems.io

  • KlaviyoSep 23, 2025Caveat

    Klaviyo's global holdout groups need at least 400,000 profiles. It recommends holding out about 5% of marketable profiles for 3 months, and the lift counts all conversions, attributed or not.

    “We additionally recommend that you run holdout groups for 3 months to gather enough data.”

    Read it on help.klaviyo.com

In the full answer

For readers who sign in
  • What to do this week3 steps
  • The numbers behind it3 figures from cited sources
  • The strongest disagreementThe best case against the answer, from a cited source
  • What the research saysWhat academic research adds, kept apart from the answer
  • How sure we areWhat its confidence rests on, and what would change the answer

Read next

More in Incrementality tests