There is a particular kind of expensive mistake that is almost impossible to see from inside a platform dashboard. It looks exactly like success. The ROAS is strong, conversion volume is solid, your campaigns hit their efficiency targets, and every month you confidently pour more budget into the channels that are "working." The problem is they are not working. They are claiming credit for work done somewhere else. I call these Attribution Lies, and once you learn to spot them, you cannot unsee them.
An Attribution Lie is a channel that looks productive in your reporting because it captures conversions, but did not actually create the customer intent behind them. The clearest examples are brand search and bottom-of-funnel retargeting. Someone sees your YouTube ad, thinks about your product for a few days, then Googles your brand name and clicks the paid result. In Google's reporting, Google gets the conversion. In reality, YouTube created the customer and Google just collected the toll.
At scale, this quietly wrecks your budget. If you allocate based on platform-reported ROAS, you systematically starve the channels doing the real demand creation and over-invest in the ones doing the credit-claiming. Every month the imbalance compounds. Eventually growth stalls, but the dashboards still look fine, so you never see it coming.
A DTC pet brand that the performance agency Tier 11 and Wicked Reports worked with had been spending $52,500 a month on Amazon advertising. Platform metrics said it was productive, so they kept spending. The Tier 11 team had a hypothesis: what if Amazon was not creating customers, just intercepting them at the point of purchase?
So they tested it properly. They cut Amazon spend by 50% and watched the data. New customer acquisition held steady at around 500 a month. So they cut again. And again. And again. After four consecutive cuts, Amazon ad spend had dropped 91%, from $52,500 to $4,700 a month, and new customer numbers were essentially unchanged. Amazon revenue actually rose 33%. They had not lost Amazon customers. They had just stopped paying Amazon to take credit for customers their other channels already created.
The same pattern held for Google Brand search. A 95% cut in brand-search spend led to only a 17% drop in clicks. The brand had been paying a steep premium to appear for searches that would have found them organically anyway. Between those two channels, the team freed up tens of thousands of dollars a month and redeployed it into actual demand creation.
You do not always need a full incrementality test to sense whether you have a problem, though you should eventually run one. A few signals are worth looking for.
Your bottom-of-funnel channels post suspiciously high ROAS. Brand search and retargeting showing 8x, 10x, or 15x ROAS should raise eyebrows, not champagne. People who already know and want your brand are cheap to convert, so you are measuring the cost of the final step, not the cost of the journey that created the demand.
Cutting bottom-of-funnel spend feels terrifying, even in small amounts. If reducing your retargeting budget by 20% is anxiety-inducing, ask yourself exactly what you are afraid will stop happening. If you cannot answer that clearly, it is worth investigating.
New customer growth is flat while blended ROAS looks healthy. This is the clearest signal of all. A healthy business grows its new customer base. Retargeting-heavy accounts often show great blended metrics while quietly shrinking the pool of new customers they draw from.
Your top-of-funnel campaigns keep getting killed for poor ROAS. If you judge awareness campaigns by the same last-click ROAS standard as retargeting, you are comparing two things doing completely different jobs, and you will keep cutting the ones that actually create demand.
Standard platform ROAS will never reveal Attribution Lies, because platforms only measure what happens inside their own walls. A customer who saw a native ad, searched Google two weeks later, clicked an email, then bought is four channels all claiming credit, with none of them showing the full picture. Three metrics actually expose the problem.
New customer acquisition cost. This isolates the cost of acquiring someone who has genuinely never bought from you, based on first-party order and customer data rather than modeled estimates. If nCAC is rising while your ROAS looks stable, you have an attribution problem.
New visitor rate by channel. What share of a channel's traffic is people who have never been to your site before? Native ads and broad social should be high, often 80% or more. Retargeting and brand search should be low. If your "top-performing" channel has a 20% new visitor rate, it is recycling existing customers, not finding new ones.
Incrementality. The gold standard: holdout groups and controlled experiments that measure what would have happened without a given channel. No platform offers this natively, because it requires a measurement layer that sits outside the platforms themselves.
That is what Wicked Reports is built to provide: first-party click paths, real order and customer IDs, and the 5 Forces AI framework that assigns the right attribution window and the right KPI to each campaign based on its actual job, instead of forcing everything through one last-click model.
For the Tier 11 client, fixing the attribution layer and reallocating budget accordingly produced results across four straight months, including the two historically slowest months of the year for the brand. New customer acquisition cost fell from $193 to $128, a 34% reduction and an all-time low. New customers grew 25%, with 10,870 acquired. Revenue grew 21.9%, all on a 2.7% increase in total spend. And the channels that received no additional investment, organic, email, and Amazon, all grew too. That is what real demand creation does. When you bring genuinely new people into your ecosystem, the whole business benefits, because you are growing the pool rather than just moving conversions around inside it.
THE BIGGER PICTURE
Attribution Lies are everywhere in digital marketing, and they are not going away. Platforms will always be incentivized to claim credit, last-click models will always favor the final touch, and as long as marketers are graded on platform-reported ROAS, they will keep over-investing in channels that look efficient but are not effective. The way out is not to stop using these platforms. It is to stop letting them grade their own homework. You need a measurement layer that sits outside the platforms, one that starts with your actual customers, real order IDs, real people, real behavior, and works backward to understand which channels actually created them. That is the foundation of every good marketing decision, and it is what Wicked Reports was built to provide. See how it works on the platform overview, or book a demo.
An Attribution Lie is a channel that looks productive in your reporting because it captures conversions, but did not create the customer intent behind them. Brand search and bottom-of-funnel retargeting are the classic examples: they collect the final click from customers that other channels, like YouTube or native ads, actually brought in. Judging budget on platform-reported ROAS rewards these credit-claimers and starves the channels doing the real demand creation.
Look at new visitor rate and new customer acquisition cost by channel, not just ROAS. A genuine demand-creation channel drives a high share of first-time visitors, often 80% or more, while a credit-claiming channel like retargeting or brand search shows a low new-visitor rate. The definitive test is incrementality: cut or hold out the channel and see whether new customer acquisition actually drops. If it holds steady, the channel was claiming credit, not creating customers.
Because platforms only measure activity inside their own walls, so each one claims the conversions it touched with no view of the rest of the journey. A single customer's path can have four channels all claiming credit. Only a measurement layer outside the platforms, built on first-party order and customer data, can reconcile the full journey and show which channel actually created each new customer.