A media buyer is looking at a channel that's working and is thinking about increasing the budget. That decision depends on the shape of a curve, the curve that describes how each additional dollar performs inside the channel. Most teams scale based on last week's ROAS trend, which is a single point on a curve nobody has actually measured at adequate resolution.
Portland Leather Goods took its TikTok budget from 0 to roughly 14% of total spend over seven months. During the same period, overall portfolio efficiency improved by eight points. The detail that matters when reading the story is why it worked: the team had visibility into the shape of the curve while they were scaling.
Welcome to The Halo. Each week: a clear take on measurement, the best from our blog, a real client result, and three links worth your time. We write for operators making weekly budget decisions under pressure.
IN THIS ISSUE
- The Take: Why saturation happens, and why the actual curve has a richer shape than what most textbooks teach.
- From the Blog: The difference between diminishing returns and saturation matters because most models conflate them, which distorts the budget decision.
- Discovery of the Week: Portland Leather grew TikTok from 0 to fourteen percent of total budget and improved portfolio efficiency by eight points along the way.
- Prescient Voices: Why doubling spend almost never doubles return.
- Three Things: Links worth your time this week.
01 · THE TAKE · WHY SATURATION HAPPENS
A saturated audience is the response to disproportionate scaling. The real curve has a richer shape than the version usually taught.
When budget in a channel grows faster than the audience available to respond inside that channel, the additional dollars start to reach people with lower purchase intent. Conversion per dollar falls because the marginal dollar is buying less receptive audience. That decline is saturation in practical terms, before any model and before any dashboard. The saturation curve is simply the drawing of how that decline looks as spend climbs.
THE CARICATURE AND THE REAL CURVE
The most common version of the saturation curve is a smooth S-shape or a simple concave line: the first segment rises quickly, then flattens, then stays flat. That shape comes from mathematical functions assumed before looking at the data. The problem is that the model assumes it and the team assumes it, and it ends up being the map used to allocate budget when in reality it rarely describes the actual curve of a digital channel.
In real data, the curve is rarely that clean. Some channels show multiple efficiency peaks: a channel can flatten, receive a new creative or an audience expansion, and recover slope over several more weeks. Other channels show nearly linear returns over extended stretches before any bend. The shape also changes with auction competition, platform algorithm adjustments, seasonality, and viral events that temporarily shift the composition of the responding audience.
Forcing every curve into the same theoretical shape produces a predictable bias: the team underinvests in channels that still have slope available because the model assumes they are already flat. There is a more technical reading of this point on the Prescient blog (section below).
WHY AVERAGE ROAS LAGS THE CHANGE
The ROAS shown in the dashboard is a weighted average of all the spend in the period. It mixes the first dollars, which are typically buying very receptive audience, with the marginal dollars at the end, which are already buying audience with lower intent. While most of the period's spend keeps landing on the receptive segment, the average stays steady and the team reads the channel as healthy.
When the marginal dollar's decline finally weighs enough to move the average, the curve's bend already has several weeks of history underneath. The metric that captures the change in real time is marginal ROAS, which is the return of the next dollar at the channel's current spend level. For a weekly decision about how much to invest, that is the useful information.
CREATIVE FATIGUE AND AUDIENCE SATURATION ARE DIFFERENT THINGS
Two phenomena produce efficiency declines and look similar in the dashboard, so it helps to distinguish them. Creative fatigue happens when a specific piece stops working within the same audience, which shows up on the dashboard as a progressive rise in frequency alongside a drop in CTR and a climb in CPM driven by the loss of useful competition in the auction. A new creative against the same audience tends to restore performance. Audience saturation happens when the responding audience is exhausting itself regardless of the creative. A new piece in this case postpones the decline by a few weeks without curing it, and the operational fix runs through expanding the audience, redirecting budget to the next channel with available slope, or accepting that the channel has entered a plateau and planning from there.
When a team treats a saturation problem by rotating creative, ROAS drops, climbs again for a couple of weeks with the new piece, and falls once more. The real curve predicts that pattern in advance, while creative rotation only postpones the bend without moving it.
WHY THIS MATTERS AT PRESCIENT
Prescient builds a saturation curve per channel with methods that let the real shape emerge from the data, instead of imposing an assumed shape on the channel. The curve updates daily with the current portfolio mix and recent market signals. For a budget decision next week, the team operates with the channel's current slope in hand, at a temporal resolution that gets close to the rate at which the portfolio actually changes.
02 · FROM THE BLOG
The difference between diminishing returns and saturation, and why mixing them up distorts the budget decision.
Diminishing returns is the economic process by which each additional dollar produces a smaller return than the previous one, a gradual and continuous property that is running all the time in any channel with finite audience. Saturation is the state a channel reaches when the audience responds so weakly that the marginal dollar is essentially lost. Both concepts are sometimes used interchangeably, and what matters when allocating budget is knowing which of the two states each channel in the portfolio is in right now.
The post goes deeper on why the textbook three-phase curve (high returns, diminishing returns, saturation) is a useful simplification for teaching the concept but a limited one when it comes time to decide. Real campaigns run inside multi-campaign systems with halo effects, competitive interactions, and efficiency peaks that recover after apparent plateaus.
7 min read
03 · DISCOVERY OF THE WEEK
Portland Leather grew TikTok from 0 to fourteen percent of total budget. Portfolio efficiency improved eight points during the increase.
Portland Leather Goods spent 2025 scaling TikTok while the platform's own dashboard suggested the channel was performing below its real potential. The team's cross-channel measurement showed something else: seventy percent of TikTok-generated revenue was landing outside TikTok itself, on DTC, Amazon, and retail, all invisible to native reporting.
On that basis they scaled. TikTok's share of total budget went from 0 to about fourteen percent over seven months, an increase that normally breaks the efficiency of the channel doing the scaling and, frequently, the efficiency of the whole portfolio.
In this case the opposite happened. Overall portfolio efficiency improved by eight points during the increase, with the index moving from 96 to 112. Cross-channel revenue share for TikTok Ads landed at 62.6 percent; for GMV Max, 66.0 percent. The most visible part of the increase was measurable: one million dollars in sales in twenty days, inside a TikTok Shop effort that drew on five hundred creators and around 3,800 videos.
The reason efficiency held during an increase of that size comes from visibility into the actual channel while it was growing. When a team operates with a curve that updates daily and can capture multiple peaks, it can keep scaling while the slope is still available and pull back before the dashboard's weighted average registers the change. By the time the reported ROAS finally adjusts, several weeks of spend at the wrong pace have already gone by.
The takeaway: every well-executed expansion rests on visibility into the channel's actual shape while it is happening. Portfolio efficiency holds during growth of that size when the team can read the curve at enough resolution to anticipate the bend before it arrives.
04 · PRESCIENT VOICES
Every brand that scaled too hard learned the same lesson the wrong way: the average dollar tells you almost nothing about where the next one lands. Grading a channel that already worked is one job. Forecasting whether it keeps working at higher spend is a completely different one, and you really need both running before you commit budget.
05 · THREE THINGS WORTH READING
1. The branded search problem is bigger than your SERP
Branded search is one of the easiest channels to misread. It captures demand the upstream work already created, so its ROAS looks high by structure more than by generating new revenue. When branded search ROAS climbs while upstream metrics stay flat, the channel is becoming more dependent on whatever is feeding it. The post breaks down how to isolate the real contribution.
2. How to measure Pinterest effectively
Pinterest is one of those cases where the saturation question shows up earlier than in other channels. It is a discovery channel, it compounds slowly, and most of its impact lands somewhere else in the portfolio. Native reporting captures a fraction of its total contribution. The post explains what to look for and where the platform's own numbers stop being useful for deciding whether to scale.
3. How to build a stack of marketing performance measurement tools
The measurement stacks most teams see are built to grade the past: how much was spent, what returned, where conversions landed. To decide next quarter the useful tool looks different, organized around forecasting the portfolio's current slope. The post details the layers most teams skip when they end up optimizing dashboards instead of budget decisions.
The Halo
Exclusive insights, every week.
Subscribe to The Halo for sharper marketing thinking.