Proof, not promises.
Incrementality and media mix studies with the actual numbers: geo holdouts, Meta and Google lift tests, and what brands did with the results.
Want the how-to behind these? Read the blog.

PopSockets turned off Google in a set of matched markets and measured what disappeared.
Pulling Google Ads for 28 days cut PopSockets.com revenue about $34K and 631 orders, at 99% significance.

Tushy held linear TV out of its markets for 28 days. Revenue dropped where it went dark.
Pulling linear TV for 28 days cut Amazon revenue an estimated 9.34%, about $47K, at 90% significance.

Do YouTube ads actually work? What 92 incrementality tests show
Attribution can love YouTube or barely see it. 92 holdouts put the median incremental return at 2.01x. What that number does, and doesn't, prove.

How Rylee + Cru showed brand shopping ads drove revenue
Pausing Brand Shopping in 15 markets cost about $139K in 20 days. Incremental ROAS 6.57.

2025 DTC Digital Advertising Incrementality Benchmarks
For every dollar spent, the typical campaign generates $2.31 in truly incremental revenue (IQR: 1.36x - 3.24x)

How incremental are podcast ads? A 3-month DTC case study
Promo codes said podcasts returned 1.0 to 1.5. Our models estimated 2.33 across three months, with one month close to break-even and plenty of caveats.

Incrementality Case Study: Longer Video, Larger Impact: Rigorous Geo-Testing Reveals :30s Outperform :15s in Top-of-Funnel Awareness Channels
Despite a 28% higher CPM, with $32 for :30s versus $25 for :15s, the longer ads delivered a 15% higher return

Incrementality Study: How Incremental is Meta Really?
Most iROAS values fell between 1.84 and 3.18, with outliers ranging from 0.91 to 5.63.

Stella Case Study: Scaling Google Ads Profitably by Reframing ROAS Targets
MER grew from 3.00 to 4.45 in 5 Months