How many of the sales your creator’s promo code gets credit for would have happened anyway? How many of those buyers would have purchased anyway? Coupon codes are a floor, not a ceiling, and they are also contaminated. They miss the viewer who buys a week later through search, and they credit the loyal customer who grabbed the code from a deal site. Incrementality testing is how you find the real number.
What coupon codes actually measure
A code records one thing: a purchase where someone typed a string at checkout. That is a useful receipt. It is not a causal claim.
Codes undercount in three common ways:
- Delayed and cross-device buyers. A viewer watches a review on mobile, then buys on a laptop days later via a branded search.
- Retail and marketplace purchases. The code works only on your DTC site, so Amazon and in-store sales vanish from the report.
- Silent influence. Plenty of viewers never use a code but still convert because a creator made the product credible.
They also overcount. Existing customers redeem codes. Coupon aggregator sites scrape and republish them. A creator’s code can end up credited with a sale that a retargeting ad or an email would have closed regardless.
So the code is a floor on attributable sales and an unreliable estimate of causal sales. Both statements are true at once, which is why finance teams get suspicious when you present a code-based ROAS as the whole story. If you need a framework for presenting evidence at different confidence levels, our piece on tiered proof for ROI claims is a good companion.
Incrementality in one paragraph
Incrementality asks a counterfactual question: what would sales have been without the creator campaign? You answer it by withholding exposure from a comparable group and measuring the gap. Google’s documentation for its open-source Meridian GeoX library describes the logic plainly: observed response is compared against a projection, and the gap is the incremental effect. That gap, not the code count, is the number you can defend.
A redeemed code tells you who used a discount. A holdout tells you who your creator actually moved.
Where AI earns its place (and where it does not)
Let’s be precise. AI does not create incrementality. Randomization or a well-built control does. What AI and modern statistical tooling do is make the design and the analysis cheaper and faster.
Practical uses I would trust today:
- Market selection. Algorithms pick matched geographic markets so test and control regions behave alike before the campaign starts. GeoX’s documented workflow covers study design and incrementality analysis in a single library.
- Counterfactual modeling. Time-based regression builds a synthetic baseline for the markets that received creator content, then tests whether the lift is statistically significant.
- Calibration. Experiment results can feed a marketing mix model. Google’s docs note that GeoX data passing quality checks can be used to calibrate Meridian MMM, which gives you a tested anchor for your always-on model.
- Anomaly checks. Automated flags catch contaminated controls, such as a viral post that leaked into a holdout market.
Where I would stay skeptical: any vendor that says its “AI attribution” delivers incrementality without a holdout. If there is no control group, you have a correlation model with good branding. The same logic applies to opaque scores that hide their inputs. Ask for the mechanism.
Is this affordable for a mid-size brand?
Increasingly, yes. Cost used to be the objection. According to PPC Land’s analysis, Google cut its minimum incrementality experiment budget to 5,000 dollars in November 2025, down from thresholds approaching 100,000 dollars. That applies to Google’s own experiment tooling, so it does not price a creator test directly, but it shows where the market is heading.
The same article reports that up to 75% of buy-side decision-makers rate current measurement methods as underperforming, citing the IAB’s State of Data research. You are not alone in distrusting last-click and code-based reporting.
One caution on vendor claims. Google’s product team has touted large savings for GeoX, but PPC Land points out that such numbers had not been independently replicated. Treat launch-stage efficiency claims as hypotheses.
A test design that works for creator campaigns
Creators complicate things. You cannot always control who sees organic posts. Here is a workable approach that respects that reality.
Step 1: Pick the question. “Did this creator program lift total sales?” is a different test from “Did Creator A outperform Creator B?” Start with the first.
Step 2: Choose a design. Geo holdouts work when creators’ audiences cluster regionally or when you can pair content with paid amplification (Spark Ads, Partnership Ads) targeted by region. Audience-level holdouts work when the posts are whitelisted and run as paid. For organic-only programs, a pre/post design with synthetic controls is weaker but still better than codes alone. For setup details, see our walkthrough of geo and holdout tests.
Step 3: Measure the right outcome. Track total sales in each market across all channels, not just coded orders. Include retail if you have data.
Step 4: Hold the line on timing. Run long enough to capture your purchase cycle. A skincare brand with a 10-day consideration window cannot read a 7-day test.
Step 5: Reconcile with codes. Divide incremental sales by coded sales. That ratio is your code multiplier for this category and creator tier. If incremental orders come out well above coded orders, your multiplier is above 1.0, and the size of that gap tells you how much the code report understates. Apply it cautiously and revisit it quarterly.
The goal is not to retire coupon codes. It is to calibrate them, so a code report becomes a leading indicator you have actually validated.
Common failure modes
- Contaminated controls. Creator content travels. Pick markets with limited spillover or accept a conservative read.
- Underpowered tests. Small lifts need large samples. If the confidence interval spans zero, the honest answer is “inconclusive,” not “it worked.”
- Overreliance on assisted conversions. Platform-reported assists are not lift. We covered this in holdout proof before payout.
- Tracking gaps. Consent loss and browser limits distort both arms. Audit your pipeline first, as outlined in our piece on server-side tracking.
Compliance does not pause for measurement
Testing changes how you measure, not what creators must disclose. The FTC’s endorsement guidance states that receiving free products or perks with the expectation of promotion brings the FTC Act into play. If you shift budget toward creators your test favors, keep disclosure requirements in the contract and audit them. A strong lift number does not offset a regulatory problem.
How to present the results to finance
Lead with the incremental estimate and its confidence range. Then show the coded floor beside it. Finance teams respond well to a range with a stated method, because it looks like a forecast with assumptions rather than a victory lap. Pair it with break-even math so the conversation is about thresholds, not adjectives.
Next step: pick your single largest creator program, design one geo or audience holdout for the next flight, and compute your first code multiplier before renewing contracts.
FAQs
Why do coupon codes undercount creator-driven sales?
Codes only capture buyers who enter them. Viewers who purchase later through search, on another device, or at a retailer are invisible to code tracking, even when the creator’s content drove the decision.
Can coupon codes also overstate creator impact?
Yes. Existing customers redeem codes they would have used anyway, and deal aggregator sites republish them. Those orders get credited to the creator without being caused by the creator.
What is incrementality testing in influencer marketing?
It measures the sales that happened because of the creator campaign by comparing an exposed group to a comparable group that was not exposed. The difference between the two is the incremental lift.
Does AI make incrementality testing possible without a control group?
No. AI helps with market matching, counterfactual modeling, and result analysis, but a credible causal read still needs a control, whether geographic, audience-based, or synthetic. Be wary of any tool that skips this step.
How long should a creator incrementality test run?
Long enough to cover your typical purchase cycle plus a cooldown period for delayed conversions. Short tests on considered purchases tend to understate lift.
What is a code multiplier and how do I use it?
It is the ratio of incremental sales to coded sales from a test. You can apply it to future coded results for similar creators and categories, then refresh it regularly as audiences and offers change.
Do I still need FTC disclosures if I run holdout tests?
Yes. Measurement design has no effect on disclosure obligations. Any creator with a material connection to your brand must disclose it clearly, regardless of how you evaluate performance.
Top Influencer Marketing Agencies
The leading agencies shaping influencer marketing in 2026
Agencies ranked by campaign performance, client diversity, platform expertise, proven ROI, industry recognition, and client satisfaction. Assessed through verified case studies, reviews, and industry consultations.
Moburst
-
2
The Shelf
Boutique Beauty & Lifestyle Influencer AgencyA data-driven boutique agency specializing exclusively in beauty, wellness, and lifestyle influencer campaigns on Instagram and TikTok. Best for brands already focused on the beauty/personal care space that need curated, aesthetic-driven content.Clients: Pepsi, The Honest Company, Hims, Elf Cosmetics, Pure LeafVisit The Shelf → -
3
Audiencly
Niche Gaming & Esports Influencer AgencyA specialized agency focused exclusively on gaming and esports creators on YouTube, Twitch, and TikTok. Ideal if your campaign is 100% gaming-focused — from game launches to hardware and esports events.Clients: Epic Games, NordVPN, Ubisoft, Wargaming, Tencent GamesVisit Audiencly → -
4
Viral Nation
Global Influencer Marketing & Talent AgencyA dual talent management and marketing agency with proprietary brand safety tools and a global creator network spanning nano-influencers to celebrities across all major platforms.Clients: Meta, Activision Blizzard, Energizer, Aston Martin, WalmartVisit Viral Nation → -
5
The Influencer Marketing Factory
TikTok, Instagram & YouTube CampaignsA full-service agency with strong TikTok expertise, offering end-to-end campaign management from influencer discovery through performance reporting with a focus on platform-native content.Clients: Google, Snapchat, Universal Music, Bumble, YelpVisit TIMF → -
6
NeoReach
Enterprise Analytics & Influencer CampaignsAn enterprise-focused agency combining managed campaigns with a powerful self-service data platform for influencer search, audience analytics, and attribution modeling.Clients: Amazon, Airbnb, Netflix, Honda, The New York TimesVisit NeoReach → -
7
Ubiquitous
Creator-First Marketing PlatformA tech-driven platform combining self-service tools with managed campaign options, emphasizing speed and scalability for brands managing multiple influencer relationships.Clients: Lyft, Disney, Target, American Eagle, NetflixVisit Ubiquitous → -
8
Obviously
Scalable Enterprise Influencer CampaignsA tech-enabled agency built for high-volume campaigns, coordinating hundreds of creators simultaneously with end-to-end logistics, content rights management, and product seeding.Clients: Google, Ulta Beauty, Converse, AmazonVisit Obviously →