Step two · before any modelling
Can this data answer the question at all?
A fit takes minutes and hands back confident-looking numbers whether or not the design could ever have identified them. So we ask whether it could first, and it costs a second. Nothing here has been modelled — this is arithmetic on your file.
The verdict
No-go
The model will still run — it never refuses — but this data cannot tell your channels apart. Read the total, not the split, and fix the columns named below before you quote any single channel.
The problem is with the shape of the data itself, not with one channel.Two channels are near-duplicatesIll-conditioned design matrixNothing you do to a single column fixes that.
How this verdict is decided
There is no score to improve and no percentage to argue with. A caution about one channel never demotes the whole dataset, because it has somewhere better to live: that channel’s own row, and its trust grade after the fit.
Takes about two minutes, and it is the step that catches impossible assumptions before they become results.
Five questions we ask of every dataset
2.74 · line at 4.00
across the whole dataset
13 channels flagged
Two channels are near-duplicates, Two channels are near-duplicates, Two channels are near-duplicates, Two channels are near-duplicates, Two channels are near-duplicates, Two channels are near-duplicates, Two channels are near-duplicates, Two channels are near-duplicates, Two channels are near-duplicates, Two channels are near-duplicates, Two channels are near-duplicates, TV — Cable and Ill-conditioned design matrix
3 channels flagged
TV — Broadcast, TV — Cable and Out of Home
Present
a demand control is mapped
No extreme weeks
nothing the outlier check flagged
Where channels moved together
- TV — Broadcast moved with TV — Cable (r=1.00)
- Social — Meta moved with Social — TikTok (r=1.00)
- Social — Meta moved with Search — Brand (r=0.96)
- Social — Meta moved with Search — Non-brand (r=0.96)
- Social — Meta moved with Programmatic Display (r=0.96)
- Social — TikTok moved with Search — Brand (r=0.96)
- Social — TikTok moved with Search — Non-brand (r=0.96)
- Social — TikTok moved with Programmatic Display (r=0.95)
- Search — Brand moved with Search — Non-brand (r=1.00)
- Search — Brand moved with Programmatic Display (r=0.93)
- Search — Non-brand moved with Programmatic Display (r=0.93)
17 things to know, each with what to do about it
TV — Broadcast and TV — Cable correlate at r=1.00 after transformation. No model can attribute between them; combine them or run an experiment.
What to do: These two were bought as one thing, so model them as one thing: combine the columns. If you need them split, the only way to get there is an experiment that moves one without the other.
Two channels are near-duplicates0.998 measured · 0.90 thresholdOpen
Social — Meta and Social — TikTok correlate at r=1.00 after transformation. No model can attribute between them; combine them or run an experiment.
What to do: These two were bought as one thing, so model them as one thing: combine the columns. If you need them split, the only way to get there is an experiment that moves one without the other.
Two channels are near-duplicates0.959 measured · 0.90 thresholdOpen
Social — Meta and Search — Brand correlate at r=0.96 after transformation. No model can attribute between them; combine them or run an experiment.
What to do: These two were bought as one thing, so model them as one thing: combine the columns. If you need them split, the only way to get there is an experiment that moves one without the other.
Two channels are near-duplicates0.960 measured · 0.90 thresholdOpen
Social — Meta and Search — Non-brand correlate at r=0.96 after transformation. No model can attribute between them; combine them or run an experiment.
What to do: These two were bought as one thing, so model them as one thing: combine the columns. If you need them split, the only way to get there is an experiment that moves one without the other.
Two channels are near-duplicates0.958 measured · 0.90 thresholdOpen
Social — Meta and Programmatic Display correlate at r=0.96 after transformation. No model can attribute between them; combine them or run an experiment.
What to do: These two were bought as one thing, so model them as one thing: combine the columns. If you need them split, the only way to get there is an experiment that moves one without the other.
Two channels are near-duplicates0.959 measured · 0.90 thresholdOpen
Social — TikTok and Search — Brand correlate at r=0.96 after transformation. No model can attribute between them; combine them or run an experiment.
What to do: These two were bought as one thing, so model them as one thing: combine the columns. If you need them split, the only way to get there is an experiment that moves one without the other.
Two channels are near-duplicates0.960 measured · 0.90 thresholdOpen
Social — TikTok and Search — Non-brand correlate at r=0.96 after transformation. No model can attribute between them; combine them or run an experiment.
What to do: These two were bought as one thing, so model them as one thing: combine the columns. If you need them split, the only way to get there is an experiment that moves one without the other.
Two channels are near-duplicates0.955 measured · 0.90 thresholdOpen
Social — TikTok and Programmatic Display correlate at r=0.95 after transformation. No model can attribute between them; combine them or run an experiment.
What to do: These two were bought as one thing, so model them as one thing: combine the columns. If you need them split, the only way to get there is an experiment that moves one without the other.
Two channels are near-duplicates0.998 measured · 0.90 thresholdOpen
Search — Brand and Search — Non-brand correlate at r=1.00 after transformation. No model can attribute between them; combine them or run an experiment.
What to do: These two were bought as one thing, so model them as one thing: combine the columns. If you need them split, the only way to get there is an experiment that moves one without the other.
Two channels are near-duplicates0.929 measured · 0.90 thresholdOpen
Search — Brand and Programmatic Display correlate at r=0.93 after transformation. No model can attribute between them; combine them or run an experiment.
What to do: These two were bought as one thing, so model them as one thing: combine the columns. If you need them split, the only way to get there is an experiment that moves one without the other.
Two channels are near-duplicates0.929 measured · 0.90 thresholdOpen
Search — Non-brand and Programmatic Display correlate at r=0.93 after transformation. No model can attribute between them; combine them or run an experiment.
What to do: These two were bought as one thing, so model them as one thing: combine the columns. If you need them split, the only way to get there is an experiment that moves one without the other.
TV — Cable has VIF 5170.9 after transformation; it is nearly a linear combination of the other regressors.
What to do: Usually the same cause as near-duplicate channels: this one is reconstructable from the others. Combine it with whichever channel it shadows, or drop a control that is standing in for it. A channel that launched partway through the window hits this too, because the baseline absorbs the period it was dark.
Condition number 216 means the design is close to singular.
What to do: The design as a whole is near-singular, so no single column is to blame. Simplify: fewer channels, fewer controls, or a longer window.
104 observations against 38 parameters is 2.7x, under the 4x rule of thumb. Expect wide intervals.
What to do: Merging two closely related channels would buy back headroom. Otherwise expect wide intervals and treat them as the honest answer rather than a defect.
It runs in 39% of weeks, in bursts rather than at the start or end of the window. The effect is identified from the handful of on/off transitions rather than from the whole period, so its interval comes back wide even when the fit itself is sound.
What to do: Nothing to fix in the data — this is a media plan, not a defect. Expect a wide range on this channel and weight it accordingly; a lift test on a single burst would pin it faster than more history would.
TV — Cable is off air most weeks0.567 measured · 0.50 thresholdOpen
It runs in 39% of weeks, in bursts rather than at the start or end of the window. The effect is identified from the handful of on/off transitions rather than from the whole period, so its interval comes back wide even when the fit itself is sound.
What to do: Nothing to fix in the data — this is a media plan, not a defect. Expect a wide range on this channel and weight it accordingly; a lift test on a single burst would pin it faster than more history would.
Out of Home is off air most weeks0.596 measured · 0.50 thresholdOpen
It runs in 40% of weeks, in bursts rather than at the start or end of the window. The effect is identified from the handful of on/off transitions rather than from the whole period, so its interval comes back wide even when the fit itself is sound.
What to do: Nothing to fix in the data — this is a media plan, not a defect. Expect a wide range on this channel and weight it accordingly; a lift test on a single burst would pin it faster than more history would.
What we will do about it
Findings can change the model, never your numbers — we never impute, winsorise or drop a row. A change a finding calls for is only a candidate until a full fit has tested it on your own spend, and every entry says what it would cost you.
A full fit will test these on your own data before choosing
None of these is a change yet. Before fitting, a full fit hides returns we chose in copies of your own spend and tries setups built from these changes on them, side by side with your model exactly as you specified it. It uses a change only if its setup does better, and holds back from the budget plan any channel whose return it cannot find. A quick preview tests nothing and uses none of them.
One combined return for two channels bought together
To be testedraised by two channels bought togetherfor each ofTV — Broadcast + TV — CableSearch — Brand + Search — Non-brandSocial — Meta + Social — TikTokThese two channels were bought so closely together that the data cannot tell their effects apart, so any split between them would be a guess. We estimate one return for the pair from their combined spend, starting from an assumption that blends the two channels' own, weighted by how much each spent.
What it would cost you, if a full fit uses it
You get one number for the pair and no split within it. If you need each channel's own return, only an experiment that moves one without the other can give it to you.
One carryover shared by every channel
To be testedraised by a short history for everything the model estimatesYour history is short for everything the model has to estimate. Instead of learning separately how long each channel's ads keep working after they run, we estimate one carryover for all of them, which leaves more of your data for the returns themselves.
What it would cost you, if a full fit uses it
Channels that genuinely differ in how long their ads keep working are treated alike, so each channel's timing is assumed rather than measured.
A firmer starting assumption for a channel the data barely sees
To be testedraised by spend that runs in short burstsOut of HomeThis channel's spend barely moves, runs in short bursts or covers only part of the window, so your data holds little evidence about what it returned. We narrow its starting assumption, so a few weeks of noise cannot hand it credit that belongs to the baseline or to another channel.
What it would cost you, if a full fit uses it
This channel's number leans on the starting assumption more than on your data. Read it as an informed default, not a measurement.
A shared starting assumption for two channels bought together
To be testedraised by two channels bought togetherfor each ofTV — Broadcast + TV — CableSearch — Brand + Search — Non-brandSocial — Meta + Social — TikTokThese two channels were bought so closely together that the data struggles to tell their effects apart. Each keeps its own return, but both start from the assumption that they return about the same, so the split between them leans on that rather than on noise.
What it would cost you, if a full fit uses it
The two returns are pulled toward each other. If one channel genuinely returns far more than the other, this setting narrows the gap between them.
A smoother baseline with fewer bends
To be testedraised by a short history for everything the model estimatesYour history is short for everything the model has to estimate. We let the baseline — the sales you would have had without media — change direction at fewer points in time, which leaves more of your data for the channel returns.
What it would cost you, if a full fit uses it
A slow shift in underlying demand that the smoother baseline cannot follow may be credited to media instead.
An assumed diminishing-returns curve for a channel the data barely sees
To be testedraised by spend that runs in short burstsOut of HomeThis channel's spend barely moves, runs in short bursts or covers only part of the window, so your data cannot trace how its returns fall off as spend rises. We fix that curve at the starting assumption — half its full effect at about its typical weekly spend — and estimate only how much the channel returned.
What it would cost you, if a full fit uses it
How fast this channel runs into diminishing returns is assumed, not measured, so its number, and any budget advice built on its curve, leans on the starting assumption.
How much each channel's spend moves about
Ups and downs are what let the model see what a channel does. Spend the same amount every week and there is nothing to learn from — the return comes back as whatever we assumed going in, however large the budget.
| Channel | Spend share | Variation | Overlap cost | Weeks with nothing spent |
|---|---|---|---|---|
| TV — Broadcast | 22.6% | 1.34 | 5169.3 | 61% of weeks had no spend |
| TV — Cable | 12.3% | 1.34 | 5170.9 | 61% of weeks had no spend |
| Social — Meta | 15.5% | 0.25 | 311.4 | spent every week |
| Social — TikTok | 8.2% | 0.25 | 310.4 | spent every week |
| Search — Brand | 7.4% | 0.45 | 505.1 | spent every week |
| Search — Non-brand | 11.1% | 0.42 | 486.9 | spent every week |
| Programmatic Display | 7.9% | 0.19 | 15.6 | spent every week |
| Video — CTV | 9.4% | 0.60 | 9.1 | spent every week |
| Out of Home | 5.6% | 1.23 | 1.2 | 60% of weeks had no spend |
How closely your channels move together
Read a cell as: when this channel went up, did that one go up too? 1.00 means always, 0.00 means never. Anything above 0.80 and your data cannot tell the two apart. Measured after allowing for advertising that keeps working once it stops running and for the tenth million doing less than the first — checking raw spend instead is the usual reason a tool declares your data fine and then fails to separate anything.
| TV — Broadcast | 1.00 | 1.00 | 0.18 | 0.19 | 0.21 | 0.20 | 0.14 | -0.02 | 0.08 |
|---|---|---|---|---|---|---|---|---|---|
| TV — Cable | 1.00 | 1.00 | 0.18 | 0.18 | 0.21 | 0.20 | 0.14 | -0.02 | 0.08 |
| Social — Meta | 0.18 | 0.18 | 1.00 | 1.00 | 0.96 | 0.96 | 0.96 | 0.24 | 0.05 |
| Social — TikTok | 0.19 | 0.18 | 1.00 | 1.00 | 0.96 | 0.96 | 0.95 | 0.24 | 0.07 |
| Search — Brand | 0.21 | 0.21 | 0.96 | 0.96 | 1.00 | 1.00 | 0.93 | 0.23 | 0.07 |
| Search — Non-brand | 0.20 | 0.20 | 0.96 | 0.96 | 1.00 | 1.00 | 0.93 | 0.23 | 0.08 |
| Programmatic Display | 0.14 | 0.14 | 0.96 | 0.95 | 0.93 | 0.93 | 1.00 | 0.19 | 0.02 |
| Video — CTV | -0.02 | -0.02 | 0.24 | 0.24 | 0.23 | 0.23 | 0.19 | 1.00 | -0.04 |
| Out of Home | 0.08 | 0.08 | 0.05 | 0.07 | 0.07 | 0.08 | 0.02 | -0.04 | 1.00 |
| TV — Broadcast | TV — Cable | Social — Meta | Social — TikTok | Search — Brand | Search — Non-brand | Programmatic Display | Video — CTV | Out of Home |