Skip to content
Bayesin

Marketing Mix

Sample dataset

The next dollar earns most in Affiliate at 3.06× marginal return, and least in Display at 0.83×. 2 of 5 channels are past 75% saturation.

Modeled

You are not signed in, so this is a sample dataset rather than your own. Sign in and upload weekly data to fit a model against real spend.

Media ContributionModeled
$88.8Mincremental revenue
Modelled Spend
$23.8Mobserved
Baseline ShareModeled
54.0%not paid-driven
Saturated
2/5past 75%

Marginal return on next dollar

Return on the next dollar of spend, after diminishing returns. This ordering, not average ROAS, is what should drive reallocation.

View as table
Marginal return by channel
ChannelMarginal ROASAverage ROASSaturation
Affiliate3.06×3.02×45%
Paid Social2.64×3.23×56%
Paid Search2.33×5.77×78%
Video1.91×1.69×39%
Display0.83×2.89×84%

Channel decomposition

Credible intervals are the 80% posterior range on average ROAS. A wide interval means the model has seen too little variation in that channel's spend to be confident.

ChannelSpendContributionROAS80% intervalSaturation
Affiliate$2.3M$7M3.02×2.56×3.48×
45%
Paid Social$6.2M$20.1M3.23×2.71×3.76×
56%
Paid Search$7.8M$45.1M5.77×4.90×6.64×
78%
Video$4.2M$7M1.69×1.43×1.95×
39%
Display$3.3M$9.6M2.89×2.53×3.25×
84%