One angle not really covered here: a lot of people frame this as "native is dead vs native still works," but the real variable is which network you're testing on for your specific vertical. MGID, Taboola, Outbrain, and Revcontent have meaningfully different approval bars, minimum CPCs, and demand density by niche — testing the same advertorial across two networks can give wildly different results even with identical creative and funnel.
If budget is the concern (since everyone's right that testing isn't cheap anymore), starting with the lower-CPC networks like MGID or Revcontent to validate the funnel/angle before throwing money at Taboola/Outbrain tends to be the more capital-efficient approach. Burning budget on premium native to test an unproven angle is usually the mistake people make when they say native "got too expensive."
If budget is the concern (since everyone's right that testing isn't cheap anymore), starting with the lower-CPC networks like MGID or Revcontent to validate the funnel/angle before throwing money at Taboola/Outbrain tends to be the more capital-efficient approach. Burning budget on premium native to test an unproven angle is usually the mistake people make when they say native "got too expensive."