By December 31, 2027, flat monthly fees will be the default way businesses spending under $50k a month on Meta buy management. Percent of spend will survive only where an agency can explain what extra work the extra spend creates.

Most agencies will not tell you this plainly: the percent fee was built to price labor that Meta now does. For years, 10 to 20% of spend, often with a $1,000 to $5,000 minimum, was standard for Meta management. It is still quoted that way in ecommerce pricing guides. The model made sense when someone built audiences by hand, split placements and adjusted bids account by account. Advantage+ now handles audience selection, placement, budget allocation and bidding. I used to bill that way myself. I was wrong to keep doing it as long as I did.

Say you spend $20k a month. At 15%, you pay $3,000 whether the agency ships two new creatives or twenty, whether CPA falls or drifts. Raise the budget to $40k because Christmas worked, and you just gave your agency a raise for your own good news. Here are six dated predictions about what replaces that deal, why I expect each one, and what would prove me wrong.

What Meta took over, and what still takes work

Keep that $20k account in mind. In 2019, it paid for hours of manual assembly: stacking interest audiences, excluding purchasers one list at a time, splitting Facebook from Instagram from Audience Network, and nudging bids when delivery stalled. Advantage+ now handles audience selection, placement optimization, budget allocation and bidding. Advantage+ Audience treats age and interests as soft suggestions, with location and minimum age as the hard constraints. One campaign can test up to 150 creative combinations, with an existing-customer cap around 25 to 30%.

This is not a pilot program. Advantage+ campaigns represent 62% of ecommerce ad spend and deliver 22% higher ROAS than manual campaigns on average. Some 82% of advertisers use Advantage+ automation in some form, and Conversions API adoption hit 89% by late 2025. Manual detailed targeting is being systematically retired.

Meta’s Andromeda engine handles retrieval: it decides which ads reach auction and lifted retrieval recall 6% and ad quality 8%. It reads the message, visuals and format to decide who is likely to respond. Creative is the targeting now. What still takes work is creative strategy and production, offer design, Pixel and CAPI upkeep with Event Match Quality above 6.0, and budget judgment. That is the work a fee needs to buy.

Prediction 1: Flat fees become the SMB default by December 2027

By December 31, 2027, most serious proposals for Meta management under $50k a month in spend will lead with a flat fee. Percent of spend will appear only where an agency can tie it to additional work, such as more creative volume.

The mechanism is arithmetic. Meta workload now tracks creative volume, testing cadence and tracking maintenance more closely than budget size. Divide a flat fee by the percent rate to find the breakeven: $5,000 versus 15% breaks even at $33,333 a month in spend. Buyers in Dubai have already published the same comparison using AED 5,000 flat versus 15%. Above that line, the percent fee costs more for the same scope. At $100k spend, 15% is $15,000 versus about $5,000 flat. The gap repeats every month.

The objection is already showing up in pricing arguments: under a percent fee, the manager gets paid the same whether an account produces 50 leads or 500, and a recommended budget increase is also a pay rise, even though creative, testing and tracking drive much of the workload. groas already offers Meta Ads management for $1,499 a month, with no cut of spend and no setup fee. Watch what fee structure appears first on the proposal, not what an agency says it can negotiate later.

I will be wrong if, in Q4 2027, you collect ten SMB proposals and more than half still lead with percent of spend as the default.

Prediction 2: Creative concepts become the priced unit by June 2027

By June 30, 2027, most competent Meta proposals will quote new concepts tested per month, not a percentage of budget. Andromeda reads an ad’s message, visuals and format to decide who sees it. Feed that system one or two polished videos a month and you give it little to work with. Brands testing 20 or more new ads monthly post 65% higher ROAS than those testing fewer than 10; top-third Andromeda-era accounts run about 395 live ads versus 296 for the bottom third. I stopped counting audiences in audits last year. I count live creatives and kill rate.

That changes what you are buying. An agency that writes hooks and briefs new ads weekly costs more than one that only manages campaigns. Current creative prices run around $150 to $500 per static, $500 to $2,000 per video, and about $99 per UGC video before edits. Expect tiers that distinguish 6 ideas turned into 30 assets from 12 ideas turned into 60, with high- and low-production treatments spelled out.

If a proposal cannot tell you how many new concepts your fee includes, it is selling 2019 labor at 2026 prices. I will be wrong if percent-of-spend proposals in late 2027 still routinely bundle unlimited creative without a volume cap.

Prediction 3: Pure pay per acquisition stays niche through 2027

By December 31, 2027, true pay per lead or pay per sale will still cover less than 10% of managed Meta spend outside affiliate and lead arbitrage. I get asked who charges on CPA instead of spend almost weekly. I wrote out which fee types actually price on cost per acquisition because the short answer disappoints people.

The structures exist: cost per lead at $50 to $200 plus, CPA priced to margin, or revenue share at 3 to 10% of tracked sales. What does not exist is an easy version of that deal for an unproven offer with messy tracking. An agency paid only when a lead fires has every reason to pick proven offers with clean funnels and to price its risk into the rate. It also has an incentive to chase volume or last-click ROAS at the expense of lead quality and top-of-funnel work. That is not villainy. It is the contract doing what the contract says.

The version I expect to survive is a hybrid: a base that covers creative and measurement, plus a bonus tied to qualified pipeline or closed revenue rather than raw lead count. If an agency offers you pure CPA with no base, ask what offers it rejects. I will be wrong if unknown brands with new pixels can get pure CPA terms from reputable managers without paying a premium over flat fees.

Prediction 4: Media-only shops lose modest-budget clients first

By March 31, 2027, media-only Meta shops will be visibly losing accounts spending $5k to $20k a month to shops that write, shoot and cut creative. The mechanism is turnaround time. When an account needs 20 or more new tests a month to hold ROAS, every round trip to a separate freelancer or back to the client for UGC adds time while tired ads keep spending. A shop that briefs hooks on Monday, shoots lo-fi on Tuesday and launches on Friday can run a 3 motivations × 2 treatments × 5 outputs system: 30 assets from 6 ideas. A shop that manages bids and waits for you to supply video cannot.

The price gap explains why modest spenders feel this first. Static creative at $150 to $500 each, video at $500 to $2,000, and about $99 per raw UGC video before edits make two vendors hard to carry on a modest budget: one charging 15% for media and another charging by the asset. That split still appears in ecommerce cost breakdowns. Those clients will look for one flat fee that covers both. Larger spenders can carry separate contracts longer.

If your agency asks you for creative more often than it sends you creative, you are the creative department. I will be wrong if standalone media management retains most sub-$20k accounts through 2027.

Prediction 5: Audience expertise disappears from pitches by October 2027

By October 31, 2027, no credible Meta proposal will lead with interest stacks, lookalike percentages or placement splits. The mechanism is removal, not opinion. Meta merged manual and Advantage+ Shopping into one flow, removed dozens of interest categories in January, and treats detailed targeting you add as a suggestion while Advantage+ Audience remains the default for Sales and Leads.

I used to sell audience builds as a skill. That skill was tied to the interface. When the interface goes away, the line item has to go with it. The $20k account still needs someone to decide what to say and make the ads that say it. It does not need a pitch deck full of interest stacks.

Ask for the last three audiences an agency built from scratch, and watch how fast the call turns to creative. I will be wrong if detailed targeting returns as a hard filter or winning pitches in late 2027 still list audience architecture as a separate deliverable.

Prediction 6: Weekly action logs beat monthly reports by September 2027

By September 30, 2027, retention will belong to shops that show what changed each week, not shops that send a polished PDF each month. Creative can fatigue in days, yet many agencies still send client-facing Facebook reports monthly or quarterly. Impressions and reach arrive without a dollar outcome, creative gets little analysis, and a monthly cadence catches a tired ad weeks after it started wasting money.

An action log tells a different story: killed 4 creatives under 0.8% CTR, scaled 2 over 2.1% CTR, added 6 hooks, fixed CAPI match rate from 5.4 to 7.1. It tells the business owner what the fee bought. I made a similar call in Google Ads agent predictions you can check by 2027, for the same reason: when execution moves quickly, a slow report hides the work.

If your report does not name what was killed, what was scaled and what ships next, it is a receipt, not management. I will be wrong if monthly static reports still retain most SMB accounts through 2027.

How I would know I am wrong

I keep these predictions where I can check them. The point is to be measured, not to sound right. Flat fees lose their case if Meta puts real labor back into media buying. Through 2027, I will watch whether Q4 SMB proposals still lead with percent of spend, whether pure pay per acquisition spreads from proven offers to new advertisers, and whether monthly PDFs without weekly logs still retain accounts.

A related breakdown of why retainer models fail growth-stage advertisers names the same structural problem on Google: fees rise with spend while work stays flat. If agencies rebuild audience work as a paid line item or Meta reverses its Advantage+ defaults, I will update this call in public. The proposals and contracts, not the agency slogans, are the evidence to collect.

What to do now: put three clauses in your next Meta contract

If I am right, you do not need to wait until 2027 to buy differently. Use these three clauses the next time you hire:

  1. Flat fee with a creative quota. Name dollars per month and new concepts tested per month: for example, 8 concepts and 24 assets. Specify whether filming and editing are included or priced by asset.
  2. Weekly log, not monthly summary. Require kills, scales, spend moved, tracking fixes and next tests every seven days, tied to CPA or ROAS and Event Match Quality.
  3. No pay rise for raising budget. State that budget increases inside an agreed range do not change the fee. Tie any bonus to qualified pipeline or closed revenue, not spend or raw lead count.

That flat-fee structure is already how groas runs Meta Ads for $1,499 a month: no cut of spend, with audiences, budgets and ads worked around the clock and new tests scaled as the log shows. Ask any percent-fee bidder to re-quote with those three clauses. The ones who can do the work will say yes.

Keep this page and check it in December 2027. If flat fees with a creative quota, weekly logs and no budget-triggered raise are standard, you bought early and paid less for it. If percent of spend is still the default, send me the proposals. I will write the correction myself, with the numbers attached.