Incrementality Versus Last Click for Growth
A branded search campaign can show a 12x ROAS while contributing almost nothing to net-new revenue. That is the central problem in incrementality versus last click: one model credits the final touchpoint before a purchase, while the other asks whether the sale would have happened without the advertising.
For DTC supplement brands, the distinction is not academic. It determines whether you scale a profitable acquisition engine or keep increasing spend on campaigns that harvest demand created by your creative, affiliates, retail presence, email, and existing customer base. Last-click reporting can make a media account look efficient right up until the business runs out of genuinely new buyers.
What last-click attribution actually measures
Last-click attribution gives 100% of conversion credit to the final tracked ad interaction before checkout. A customer sees a Meta video ad for your creatine, reads reviews two days later, receives an email, then searches your brand name on Google and purchases after clicking a paid brand ad. Last click gives Google brand search the sale.
That data is not useless. It can tell you which channel closed a measurable transaction, identify high-intent queries, and help operators understand the path a customer took immediately before buying. The problem begins when a brand treats it as a complete explanation of what caused the purchase.
Supplement purchases rarely happen in one touch. A consumer may need to see your product benefits, ingredient transparency, flavor proof, third-party testing, and real customer results before they are ready to buy. For products involving recurring use – protein, hydration, gut health, sleep, or performance supplements – trust often develops over days or weeks. The last touch tends to receive credit for work performed much earlier in the journey.
This creates predictable reporting distortions. Branded search, retargeting, and email-assisted conversions often appear exceptionally efficient. Prospecting creative, creator partnerships, and upper-funnel video can look weak because their influence is difficult to capture inside a short attribution window.
Incrementality versus last click: the business difference
Incrementality measures the additional outcomes caused by marketing activity. Put simply, it asks a tougher question: if we did not run this campaign, how many purchases would we have lost?
That is a materially better question for a founder deciding where the next $50,000 of spend should go. A campaign with a 3x last-click ROAS may be highly incremental if it brings in buyers who would not otherwise discover the brand. A campaign with a 10x last-click ROAS may be mostly cannibalizing organic demand or converting customers who were already on their way to checkout.
The goal is not to declare last click “wrong.” The goal is to stop confusing conversion capture with demand creation.
Consider two common scenarios. In the first, a customer searches for your brand because they saw a high-performing TikTok creator ad. Paid brand search captures the final click. In the second, a returning subscriber types your brand name directly into Google and clicks an ad before reordering. Both sales may be credited to the same search campaign, but neither necessarily represents new demand created by that campaign.
If you optimize solely to last-click ROAS, budget naturally flows toward channels closest to conversion. Your dashboard improves, but prospecting gets underfunded, creative testing slows, and the pipeline of future customers weakens. That is how brands become dependent on branded search and retargeting while claiming they have a scalable acquisition strategy.
Why supplements make measurement harder
The supplement category adds complications that generalist ecommerce playbooks often miss. Purchase behavior is shaped by ingredient skepticism, claims compliance, repeat cadence, subscriptions, discount sensitivity, and product education. A first-time buyer may convert on a bundle after seeing five pieces of compliant creative, while a returning buyer may purchase after a simple replenishment reminder.
Those customers should not be measured the same way. A 20% off retargeting ad could receive last-click credit for an order from a customer who was already convinced. It may still be commercially useful, especially if it recovers abandoned carts, but it is not evidence that the campaign can profitably acquire cold traffic at scale.
Platform reporting compounds the issue. Meta, TikTok, Google, Shopify, and your attribution platform each use different windows, identity resolution, and modeling approaches. They will all claim revenue. If each channel is managed against its own reported ROAS without a business-level source of truth, total attributed revenue can exceed actual store revenue by a wide margin.
That does not mean platform data should be ignored. It means it needs context: blended performance, customer mix, geo trends, spend levels, and controlled tests.
How to measure incrementality without slowing growth
The strongest incrementality program is practical, not performative. You do not need a data science department to make better decisions. You need clean baseline metrics, disciplined testing, and enough patience to avoid declaring victory from one week of noisy results.
Start with blended economics
Use store-level metrics as the operating scoreboard. Track total revenue, new customer revenue, total ad spend, blended CAC, contribution margin after advertising, repeat purchase rate, and new-to-returning customer mix. These numbers cannot tell you exactly which ad drove each sale, but they reveal whether increased media spend is producing profitable business growth.
For example, if Meta spend rises 30% while new customer revenue rises 25%, blended CAC remains within target, and contribution margin holds, that is more meaningful than a reported 5x platform ROAS alone. If spend rises while new customer volume stays flat and branded search revenue grows, you may be paying more to capture demand you already had.
Use holdout tests where the decision matters
A holdout test compares an exposed group with a similar group that does not receive the campaign. Geographic testing is often the most accessible approach for DTC brands. Select matched regions, reduce or pause a specific channel in test markets, maintain normal activity in control markets, then compare the difference in outcomes.
The test must be designed around the channel and your scale. For branded search, you may test selected geographies or time periods carefully, while protecting against stock issues, promotions, and seasonality. For Meta prospecting, geo holdouts can indicate whether increased reach produces incremental new customer revenue beyond what would have happened anyway.
Do not judge a test on platform-reported conversions. Judge it on store-level revenue, new customer orders, and profitability across the test and control groups. The result may not be perfectly precise. It will still be more decision-useful than giving every sale to the final click.
Watch for diminishing returns
Incrementality changes as spend changes. The first dollars into a strong creator-style Meta campaign may reach new, high-intent audiences. The next dollars may repeatedly reach people already exposed to the brand. Reported ROAS might stay stable because retargeting and modeled conversions absorb credit, while marginal CAC quietly rises.
This is why budget decisions should be based on marginal performance, not only account averages. Ask what happened after the most recent increase in spend. Did new customer revenue grow at a profitable rate? Did blended CAC stay controlled? Did frequency rise faster than reach? Did discounting increase to preserve conversion rate?
A campaign can be incremental at $10,000 per month and inefficient at $100,000. Scale is not a switch. It is a series of changing marginal economics.
A better reporting framework for media teams
Last click should sit in your reporting stack, but it should not sit at the top of it. Use it as a diagnostic lens for bottom-funnel behavior, not as the sole allocation rule.
At the executive level, focus on profitable revenue growth: blended CAC, contribution margin, new customer acquisition, and payback period. At the channel level, evaluate platform metrics alongside creative fatigue, audience saturation, landing page conversion rate, and incrementality test results. At the campaign level, use last-click data to identify intent signals and operational opportunities, such as a high-converting keyword, an underfunded retargeting segment, or a broken checkout path.
This layered approach also changes how teams talk about creative. A prospecting ad should not be killed just because it does not win the last click. If it drives branded search lift, improves direct traffic, expands qualified remarketing pools, and increases new customer orders at an acceptable blended CAC, it is doing valuable work.
Conversely, a retargeting campaign should not receive unlimited budget because it reports a huge ROAS. Its job may be to efficiently convert existing demand. That has value, but there is a ceiling, and spending beyond it often increases frequency without producing proportional revenue.
The question to ask before increasing budget
Before scaling any channel, ask: are we paying this campaign to create demand, capture demand, or both? Then match the measurement method to the answer.
Demand-capture campaigns deserve careful efficiency controls, but their reported ROAS will often overstate their incremental value. Demand-creation campaigns require more patience and stronger creative, but they are usually where the next level of customer growth comes from. The winning media strategy does not choose one over the other. It funds both according to their real contribution to profitable growth.
For supplement brands, that means pairing policy-safe, convincing creative with measurement that respects how customers actually buy. The next budget increase should be earned not by the prettiest attribution dashboard, but by evidence that it brings more profitable customers into the business.