Jon Loomer's 2027 Meta Forecast: Stop Tuning Targeting, Start Feeding Andromeda
Jon Loomer published a 2027 forecast on May 12 that effectively retires the manual targeting role inside Meta accounts. Andromeda, the retrieval engine Meta detailed in its December 2024 engineering post, is doing the audience work now: +6 percent recall and +8 percent ads quality on selected segments, per Meta's own published numbers. The remaining lever for the advertiser is feeding the system 10 to 15 conceptually distinct creatives per Advantage+ campaign.
What Loomer is actually saying
The premise of Loomer's post is uncomfortable if you are still pricing your services around audience research. By 2027, in his read, Meta will keep collapsing the inputs the advertiser sets: detailed targeting expansion goes from optional to default, lookalikes get folded into Advantage+ Audience, and Advantage+ Shopping spreads to objectives where it used to be a niche choice.
I think most paid social managers will read that and tell themselves it does not apply to their book. It probably does. The Meta engineering post on Andromeda is doing the same job a CMO does when they reorganize a team: the work has not disappeared, it has just been reassigned. From you to the model.
The honest read is that Loomer is calling the same shift Meta's engineering team already documented on its own blog. Andromeda is built to handle the explosion of creative options, and the engineers' framing makes clear the assumption is that you supply the variety. Less targeting, more variants. That is the trade.
The numbers Meta itself published, and what they obscure
Meta's December 2024 Andromeda post is the one to read if you have not. It quantifies the system as +6 percent retrieval recall and +8 percent ads quality on selected segments, plus a separate claim that Advantage+ creative AI lifted ROAS roughly 22 percent for advertisers who had not used it before.
Those numbers seem to depend a lot on what you compare them to. A 22 percent ROAS increase against a manual baseline that was already healthy is one thing. The same 22 percent against a struggling account that was missing creative diversification to begin with is a different conversation. The averages also hide which verticals lost, and you can be sure some did, because no algorithm change moves every account in the same direction.
Anyway, the structural read is clearer than the per-account read. Meta is investing compute on Grace Hopper hardware to do retrieval at a scale no human media buyer can replicate. Once that retrieval lift exists, the budget the advertiser used to spend on audience research has no place to go inside the platform. It either leaves the channel or gets reallocated to creative.
Where to actually redirect the targeting budget
Logical Position's 2026 paid social playbook puts a sharper number on this than most. Brands testing 20-plus new ads per month showed 65 percent higher ROAS than those testing fewer than 10. That is roughly the line, from what I have seen, between accounts that are quietly being repriced by Andromeda and accounts that are still benefiting from it.
Here is the move I would make this week if I ran a paid social P&L. Take whatever you currently spend on audience research, manual interest building, and lookalike refresh. Cut it by 60 to 70 percent. Move that line item to either an in-house creative production sprint or an external production partner that can ship 3 to 5 new concepts per week. Not iterations of the same concept. Conceptually distinct: different hook, different format, different proof angle.
If your team is small, the cheapest version of this is one weekly creative review where you kill anything that has not produced a cost per result inside your account threshold over the last 14 days. We covered this earlier in the breakdown of Andromeda's 14 day creative refresh window, and the math has not gotten any kinder since.
The reason most accounts I have looked at still have room here is unspectacular. Most teams set their creative cadence in 2022 when 4 to 6 ads per month was sufficient, and nobody updated the production schedule when the retrieval system changed underneath them.
The brief Andromeda actually wants
The other piece of Loomer's read that nobody is talking about enough is the brief itself. The creative brief that worked in 2022 was built around a single hero asset and 2 to 3 size variants. The brief that Andromeda actually wants is closer to what a TV agency would call a campaign system: one core promise, 4 to 5 hook variants, 3 format treatments, and at least 2 proof angles. That math gives you the 10 to 15 conceptually distinct outputs per campaign without burning the creative team out reinventing strategy every two weeks. Loomer's diversification breakdown shows seven concrete examples of what this looks like in production.
I have started running my own briefs this way and the difference, honestly, is more about velocity than quality. The creative director stops being the bottleneck because they are not approving 15 separate concepts. They are approving one system and 15 expressions of it. Probably the biggest shift for small teams is just naming this distinction out loud, because most creative leads still treat every variant approval as a separate decision.
A second thing Loomer's 2027 framing implies, although he does not quite say it: agency retainer math gets recut. If audience setup goes from a 20 hour project to a 2 hour briefing, and creative volume needs to triple, the retainer that priced the old ratio is upside down. From what I have heard at a couple of agency dinners this spring, the conversations are already happening. Most of them have not surfaced publicly yet.
Why this prediction is less of a leap than it looks
Loomer has been on this thesis publicly for at least a year, going back to his post on how his own Meta approach changed. The Andromeda piece on his site walks through how Advantage+ Creative variations multiply one ad into hundreds.
On paper, that sounds like a clean playbook. And sometimes it is. The complication is that buyers I have talked to are quietly running into the same problem: their creative pipeline cannot keep up with the cadence Meta wants. The bottleneck moved from media buying to production, and most agency retainer contracts are not priced for the new ratio.
Social Media Examiner covered the same shift from a different angle, and their framing was useful. The advertiser's job is now closer to a creative director's, not a media buyer's. That is the part Loomer's 2027 prediction makes concrete. He is not predicting a new feature. He is predicting that the people who keep pricing their work around manual targeting will get repriced out of the channel.
I do not think the winners here are necessarily the ones with the biggest budgets. From what I have seen, it is probably just the ones who fix their creative production cadence before Q3 planning, when next year's budget allocation locks in.
For context on what winning looks like in raw account performance, look at the public split between Advantage+ Shopping and manual. We covered the 42/58 results across 640 tests earlier this quarter. It is closer than it sounds, but the trajectory is one way.
The unglamorous next 90 days
Loomer's piece will probably be quoted out of context by half the agency LinkedIn world over the next two weeks. That is normal. The part worth quoting is not the 2027 prediction itself. It is the implication for what you ship before September.
If your team produced fewer than 30 conceptually distinct ad variations this quarter, you are already inside the bracket Andromeda is silently penalizing. The fix is not strategic. It is operational, and it is mostly about whether your creative pipeline can handle a cadence change that has already happened.
Notice Me Senpai Editorial
