Push Delivery to This Ad: Meta's First Real Manual Override in Years, Explained

Meta's algorithm has, for years, quietly resisted anything resembling manual override, on the theory that it usually knows better than the advertiser does. That's starting to change. "Push Delivery to This Ad," which reached nearly all ad accounts by July 10, is the first real crack in that stance. It lets advertisers manually direct additional delivery toward a specific ad the algorithm may be under-delivering relative to how well it's actually performing. ## When it's genuinely useful You have a creative with strong early signal, good CTR, good CVR in initial testing, that isn't getting a proportional share of spend because the algorithm hasn't caught up yet to how well it's performing. That lag is real, and it can cost you a week or two of a winning creative's best window. ## When it's the wrong tool Don't use this to prop up a genuinely underperforming ad out of stubbornness or sunk cost. It's a delivery override, not a performance fix, and it will not make a weak ad convert no matter how much delivery you force toward it.  ## How to actually test it: - Pick a creative with early positive signal but a noticeably low share of spend. - Apply the override. - Compare a full week's performance against a matched ad running without it, not just a gut feeling at the end of the week.  It's a scalpel, not a hammer. Useful for a small handful of clearly-winning creatives at any given time, not a general fix for underdelivery across an entire account.