When we started working with our early-access accounts in late 2025, we were not trying to quantify misallocation. We were testing whether pre-spend ROAS forecasts would actually change how teams moved their budgets. What we found instead was something more fundamental: most DTC performance teams have very little information about where their current week's budget should go until the week is already half over.
The 25 to 35 Percent Problem
Across the six accounts in our early-access program, budget allocation decisions were being made on the basis of the previous month's data, updated with whatever signals leaked through from the current week's performance reports. On average, we estimated that 25 to 35 percent of weekly ad spend was going to channels, campaigns, or ad sets that were trending toward poor ROAS before any money was actually committed. Not dramatically poor. Just consistently below what the team could have achieved by redistributing the same total budget based on forward signals.
That number is not a dramatic finding. It is actually a reasonably optimistic assessment of how accurately teams allocate without predictive tooling. The more interesting question is why it happens, and why it is so hard to catch through the standard monthly review process.
Why Misallocation Is Hard to Catch in the Moment
Performance marketing teams are not making careless decisions. The people running DTC ad accounts are typically experienced practitioners who review their dashboards daily and have strong intuitions about how their channels behave. The problem is structural, not attentional.
The budget cycle locks spend for periods of time that are much longer than the meaningful signal window. A DTC brand running $50,000 per week across Meta, Google, and TikTok might review performance every 30 days and adjust allocations at that cadence. But the ROAS signals that would justify a reallocation start showing up 5 to 10 days into the cycle. By the time the review meeting happens, the team has already spent three more weeks on the original allocation.
There is also a dashboard fragmentation problem. Platform-reported ROAS, which is what most teams look at daily, is measured using different attribution windows on different platforms. Meta defaults to 7-day click and 1-day view. Google Ads uses a 30-day window by default for search campaigns. TikTok uses a 7-day click window but processes conversions with significant reporting lag. When a performance marketer looks at all three platforms on a Tuesday morning, they are not looking at the same thing. They are looking at three different ROAS figures that use incompatible methodologies and time periods. Making allocation decisions based on cross-platform ROAS comparisons at that point is, at best, an educated guess.
A Concrete Look at How This Plays Out
One of our early-access accounts, a DTC outdoor apparel brand managing around $100K per month in ad spend, came in with a typical allocation pattern: roughly 55 percent to Meta, 35 percent to Google, and 10 percent to TikTok. The allocation had been set based on the prior quarter's performance and was reviewed every 30 days.
Over the first four weeks of working with their data, Flyweel's models flagged that Meta's prospecting campaigns were entering a period of audience saturation that typically preceded a ROAS decline of 30 to 40 percent in that account's historical patterns. The signal was visible in the model about 10 days before it showed up clearly in campaign performance. The Google account, by contrast, was showing early positive signals from a newly expanded keyword set.
The team had not shifted budget. They were operating on last month's allocation and had no system for seeing those early signals. By the time the monthly review came around, Meta prospecting had already underperformed for three weeks. The 35 percent Google allocation had underserved a higher-converting opportunity for the same period.
The estimated misallocation cost over that 30-day period, based on projected versus realized ROAS at the channel level, was roughly 18 percent of total spend. Not catastrophic. But for a $100K-per-month account, that is around $18,000 in inefficient allocation, compounded across 12 months.
The Cost Compounds Beyond Direct Spend Waste
The direct cost of misallocation, measured as the gap between realized and achievable ROAS, is only part of the story. The downstream effects are harder to quantify but matter at least as much.
When performance teams consistently underdeliver on ROAS during certain cycles, they lose the credibility to request budget increases. A marketing director who sees two consecutive months of flat ROAS from a channel that is "supposed to be working" will typically cut that channel's allocation rather than redistribute within it. This can cause teams to abandon channels that were actually performing well at certain budget levels but were being overloaded relative to their saturation point.
There is also a compounding effect on creative testing. When a team misallocates toward a channel that is entering audience saturation, fresh creative typically gets tested against a degraded audience pool and performs worse than it would have at the right time. The team draws the wrong conclusion about what creative works, which affects the next creative cycle.
What Misallocated Means and Does Not Mean
We want to be precise here. Saying that 25 to 35 percent of spend is misallocated does not mean that money was wasted on fraudulent placements or that the campaigns were mismanaged at the execution level. Most DTC teams run well-managed accounts with sound creative testing frameworks and reasonable bid strategies.
Misallocation, as we measure it, is the gap between how budget was distributed and how it would have been distributed if the team had reliable forward signals at the channel level before the spend was committed. It is a structural measurement problem, not an operational one. The best performance marketer in the world cannot redistribute budget they do not yet know needs to be redistributed.
This distinction matters because the solution is not better account management. The solution is better information, earlier in the decision cycle.
What Changes With Pre-Spend Signals
When teams have access to channel-level ROAS forecasts before the budget goes out, the allocation decision changes from a retrospective calibration to a prospective one. Instead of asking "what did Meta do last month," the question becomes "what does Meta look likely to do this week, and how does that compare to Google and TikTok?"
That question, answered 3 to 5 days before the weekly budget cycle starts, gives teams enough runway to act. Not to make enormous strategic shifts, but to make the incremental redistributions that add up over a full quarter.
In our early-access data, teams that were using forecast signals to inform weekly allocation decisions reduced their estimated misallocation from 25 to 35 percent of spend to roughly 10 to 15 percent. That is still imperfect. Forecasts have error bands and some signals are genuinely hard to read. But cutting misallocation by more than half, at the same total spend level, has a measurable effect on realized ROAS across a quarter.
The structural problem with the monthly review cycle is not that performance marketers are reviewing their data wrong. It is that the cycle was designed for a world where budget decisions took a week to execute. In a world where reallocations can happen in an afternoon, the 30-day cadence leaves most of the efficiency gains on the table.