SSubida Skills

Esc to close · results from all notes, skills and people

Marketing›Measurement

Should we add a new ad channel? Scale, return, effort — and small platform syndrome

How to decide whether a new or smaller ad channel (CTV, AppLovin for e-commerce, chatbot ads, retail media) is worth testing — judge scale, achievable ROAS and team effort, require it to beat (not match) incumbents, and budget for a real test plus data integration.

Updated 2026-10-24·4 sources·measurementchannel-mixctvapplovinnew-channelssmall-platform-syndromecapi

Should we add a new ad channel?

TL;DR

  • Score every channel on three things: scale (can it absorb meaningful spend?), return (can it hit the ROAS/CPA target?), and effort (team hours per dollar of profit).
  • A small channel must beat a big one, not just match its ROAS. Each extra channel costs attention and integration work, and equal ROAS on a small budget yields little absolute profit.
  • Small platform syndrome. A new platform performs poorly until you send it full conversion data via CAPI, but you won’t integrate until it performs. Decide up front whether you’ll give it the data it needs to be fairly tested.
  • Plan the test properly. CTV-type channels need 4–6 weeks and enough budget. Excitement isn’t spend: macro shocks send budgets back to Meta and Google.
  • AppLovin for e-commerce (Haus meta-analysis, 15 months). Efficiency was 11% above the average of other channels tested. It beat concurrent tests 60–85% of the time in 2025, falling to a coin flip in 2026 as spend scaled.
  • Tiny budgets. Find a messaging angle with organic TikTok posts first, then put paid money on Meta.

Details

Scale, return, effort (Nebojsa Radovic, Zynga). He says the dedicated “new channel testing” teams of the 2010s are gone. He once spent hours optimising an incentivised network driving 10 purchases a day instead of Facebook driving 1,000: “think about the ROI of effort”.1 High-ARPU users are hard to find outside the walled gardens, which know who plays and pays. Testing outside them is a “shotgun approach” suited to teams with time to experiment.2

Small platform syndrome (Eric Seufert). Platforms scale their targeting data from advertiser spend, and advertisers won’t add spend if ROAS dips: a circular dependency.3 CAPI makes it worse. The advertiser tests with a limited data connection, sees weak results, and declines to send the feedback that could have improved them.4 And a smaller channel “may have to demonstrate better performance to win advertiser budget”, because equal ROAS doesn’t mean equal commercial value once you count opportunity cost and absolute profit.5 Practical takeaway: if you test a channel, either commit to its CAPI/pixel integration for the test period or accept the result as a lower bound.

CTV and new video channels (Olivia Kory). Advertisers arrive keen to crack CTV, YouTube and chatbot ads to reach new audiences. Then tariffs, a war or a missed goal hit, and they “end up back in Meta and Google”; she isn’t seeing CTV take 20% of app advertisers’ budgets.6 Sellers who position CTV as a performance channel, with low minimums, Meta-like operation and funded third-party tests, are winning trials.7 See Designing incrementality tests (duration, budget, CTV and brand) for test length.

AppLovin outside gaming. Haus’s 15-month meta-analysis of tests (mostly e-commerce/DTC brands) put AppLovin as a strong number 2 or 3 channel, with efficiency 11% better than the average of other channels the same brands tested. It works fast and mostly drives direct-to-consumer sales, which simplifies measurement.8 As brands scaled, the share of concurrent tests it beat fell from 60–85% (2025) to about 50% (2026).9 Its share of trackable digital spend rose from ~5% (Jan) to over 8% (May 2026).10

From $0 to $5K (Thomas Petit). The worst stage, with no data. Try non-paid first: post many videos on a fresh TikTok account, since organic reach still exists there, to find the angle that catches attention. Then spend on Meta, whose algorithm finds interested users fastest. Google and AppLovin need more volume.11 Founders who arrive on Facebook with two untested videos burn the first $5–10K.12

Disagreements / caveats

Footnotes

  1. S07, E33: An oral history of mobile user acquisition (with Nebojsa Radovic) @ 39:43 — Nebojsa Radovic ↩

  2. S07, E33: An oral history of mobile user acquisition (with Nebojsa Radovic) @ 40:30 — Nebojsa Radovic ↩

  3. S07, E32: The CAPI revolution @ 16:03 — Eric Seufert ↩

  4. S07, E32: The CAPI revolution @ 17:44 — Eric Seufert ↩

  5. S07, E32: The CAPI revolution @ 20:15 — Eric Seufert ↩

  6. S07, E21: AppsFlyer and the new dynamics of mobile attribution (with Olivia Kory) @ 27:09 — Olivia Kory ↩

  7. S07, E21: AppsFlyer and the new dynamics of mobile attribution (with Olivia Kory) @ 37:04 — Olivia Kory ↩

  8. S07, E21: AppsFlyer and the new dynamics of mobile attribution (with Olivia Kory) @ 45:14 — Olivia Kory ↩

  9. S07, E21: AppsFlyer and the new dynamics of mobile attribution (with Olivia Kory) @ 46:03 — Olivia Kory ↩

  10. S07, E21: AppsFlyer and the new dynamics of mobile attribution (with Olivia Kory) @ 46:54 — Olivia Kory ↩

  11. S07, E30: MDM Mailbag #8 (with Thomas Petit) @ 25:00 — Thomas Petit ↩

  12. S07, E30: MDM Mailbag #8 (with Thomas Petit) @ 27:22 — Thomas Petit ↩