By December 31, 2027, no small agency that knows its numbers will pay a percentage of ad spend for white-label Meta fulfillment. It will buy new ads tested each month for a flat fee, or it will switch providers.

I used to bill for the manual work behind Meta campaigns. That work has shrunk. Since February 2026, Advantage+ has been the default for new Sales, Leads and App Promotion campaigns, with targeting, budget, placements and creative automation on unless you opt out. Post-Andromeda delivery puts more weight on creative: Meta’s system pushes 8 to 15 distinct assets per ad set, and practitioners attribute 70 to 80% of performance to creative quality. The work left to sell is the work many white-label contracts still treat as an extra.

Here are five predictions an agency can check at the end of 2027. Each has a way to prove me wrong.

Prediction 1: By the end of 2027, flat fees beat percentage-of-spend pricing

White-label fulfillment today usually wholesales at $199 to $499 per account per month, which agencies resell at $800 to $1,500. That math worked when more spend meant more manual campaign splits, bid changes and budget babysitting. Retail Meta management is still quoted at $1,500 to $10,000 a month or 10 to 20% of spend, with $1,000 to $5,000 minimums. At $100,000 in spend, 15% is $15,000 against roughly $5,000 flat. The fee triples. The work does not.

The mechanism is simple: percentage pricing rises with spend even when the hours stay flat. Advantage+ handles budget shifts and placement choices that once helped justify the percentage. Creative still costs labor. Standalone work already runs $150 to $500 per static and $500 to $2,000 per edited video. By late 2027, agencies will ask why fulfillment takes a cut of media when the deliverable is a count of ads.

The alternative exists now: groas runs Meta Ads for $1,499 a month, at the same fee regardless of ad spend. No percentage cut. Pay for the work, not the size of the budget.

What would prove me wrong: Meta rolls back automation and hands enough manual control back to media buyers that hours scale with spend again. I do not see that happening.

Prediction 2: By the end of 2027, the creative count becomes the deliverable

I used to tell clients account structure was the product. I was wrong. Value has moved from hidden targeting tricks to creative throughput, offer clarity, landing pages and signal quality. An agency can still buy careful media management. It cannot buy its way around running the same few ads until they wear out.

The practical test is 10 to 20 distinct concepts a month instead of recycling 3 to 5 assets. Those concepts need different hooks and angles, not fresh filenames on near-identical images. When the delivery system treats look-alikes as one creative, making another version of the same ad is not much of a test. Count new ads tested per month, not campaigns launched.

By the end of 2027, I expect the white-label scope of work to name how many new statics, cutdowns and hooks the partner produces, and who kills the losers. A provider that bills $90 an hour for creative while the client supplies the assets is selling media-only work in a market where much of the media setup is automatic. Call it that, and price the missing work separately.

What would prove me wrong: Creative fatigue stops mattering because Meta starts rewarding long-running evergreen ads again. Watch frequency and CPA on month-old winners. If they keep holding up without fresh tests, my contract prediction weakens.

Prediction 3: Through 2027, media-only providers lose agency clients first

The split is already visible. Some white-label shops handle media and leave the agency to supply every ad. Others include testing in the fee. The first group looks cheaper on a quote. Then the agency pays a designer, waits on revisions and explains to the client why the same three images have run for 60 days. The missing creative work did not disappear; it moved onto the agency’s payroll.

Say you spend $20,000 a month across four clients. If your fulfillment partner tests two new concepts a month while you need ten, the quote may look tidy while CPA drifts and the churn call lands on your desk. That is the mechanism behind this prediction: the agency owns the client relationship, so it absorbs the cost of a partner’s thin creative output.

groas includes new ads tested and winners scaled in its $1,499 Meta fee, alongside ongoing audience and budget work. That bundling is the direction of travel. A cheap media-only fee is not a cheap fulfillment contract if the agency has to supply the part that keeps the ads fresh.

What would prove me wrong: Low-cost media-only providers keep agency clients longer than providers that include creative through 2027. Check client retention, not testimonials. I am betting the media-only shops lose agencies first.

Prediction 4: By the end of 2027, small agencies get served or get dropped

If you run a web design shop with three clients spending $1,500 a month each on Meta, you know the two answers. One provider offers month-to-month service with no client minimum and no annual contract. Another says it needs $2,000 to $3,000 a month in ad spend per client for meaningful data. To the shop with three $1,500 accounts, that sounds a lot like you are too small.

Typical white-label ad-spend minimums sit at $1,000 to $3,000 a month. Do not confuse two reasons for a floor. One is whether a budget gives the provider enough room to test. The other is whether its staffing model can make money on the account. A person spending three hours on every account has a different cost problem from a system that handles routine budget and placement decisions continuously.

Automation reduces that labor problem; it does not manufacture testing data. I expect providers built around always-on execution to take more small books on a flat fee, while providers built around weekly human touches keep their minimums and lose those agencies. Ask what the floor protects: the test or the staffing model. If a provider courting your three accounts cannot explain the difference, I would not sign.

What would prove me wrong: Minimums rise through 2027 even as automation spreads, because the cost to serve small accounts does not fall the way I expect. I would have misread the cost curve.

Prediction 5: By the end of 2027, reports lead with actions, not metrics

Most white-label weekly reports still lead with impressions, click-through rate and ROAS. Those numbers matter. But when Advantage+ controls delivery, they do not, on their own, show what the fulfillment partner contributed. Your client can see them in Ads Manager. The useful record sits underneath: what changed, why it changed and what happens next. A contract that locks reporting to metrics rather than decisions leaves an agency struggling to show its partner’s work.

By the end of 2027, I expect the standard white-label report to read more like an action record:

  • New concepts launched: 12.
  • Losers stopped: 7 after $250 in spend without an add to cart.
  • Winners scaled: 2 from $50 to $150 a day after cost per result held for three days.
  • Next tests: the hooks and concepts ready to run.

Those are examples of what a report should explain, not results I am claiming for an account. Budget, audience and placement changes should take three lines, not three pages. Agencies keep 30 to 50% gross on resale, falling to 27 to 37% net after 3 to 6 hours of their own time. Every hour spent turning a metrics export into a client update eats that margin. Demand the change log now.

What would prove me wrong: Clients keep renewing on metrics-only PDFs through 2027. If you are switching because results have been poor, you already know a polished report did not save the account.

Switching providers? Put these five terms in the contract

If poor results are forcing a switch, do not sign the same scope with a new logo. Weak delivery hides in vague scopes: media-only work billed as full management, creative treated as extra, and a metrics PDF standing in for proof of work. Comparison pages can rank providers while leaving out wholesale rates, creative responsibilities, account ownership and exit terms. Put these five lines in writing before you move a client:

  1. A flat fee, not a percentage. Name the wholesale fee and hold it flat as spend grows. Providers already bill flat per account, as a percentage of spend, by spend tier, by hour or task, or on performance. Pick flat. If a provider insists on a percentage, ask what extra work a $10,000 month gets over a $3,000 month.
  2. A creative count with owners attached. Specify how many new statics and videos arrive each month, who writes the hooks, who edits, and who stops losing ads. If the client supplies creative or it bills separately, put that cost beside the fulfillment quote before comparing prices.
  3. Small-book terms. Get the client-count minimum, contract length and ad-spend floor in writing. If a provider wants $2,000-plus per client to start, ask whether that threshold is about testing room or its cost to serve the account.
  4. An action log, not just a metrics export. Every weekly report should list what changed, why and what comes next. Impressions and ROAS can stay; they just should not pretend to be the whole story.
  5. Ownership and exit. Your agency owns the ad account, Page access, pixel data and creatives. On exit, the history stays where it ran and the files move with you. If that sentence is missing, you are renting results you cannot keep.

The December 31, 2027 scorecard

Pull five provider pages on December 31, 2027 and grade me. Look beyond the headline price to the scope and a sample report:

  1. Pricing: If most still lead with percentage of spend rather than a flat fee, I called the pricing shift wrong.
  2. Creative scopes: If campaigns launched and hours worked still outrank new ads tested each month, I called the deliverable wrong.
  3. Small books: If a three-client agency spending $1,500 per client still hits minimums and setup fees at most shops, I called the cost curve wrong.
  4. Reports: If the first page still leads with impressions and click-through rate instead of decisions and live tests, I called the reporting shift wrong.
  5. Retention: If media-only shops keep agencies longer than creative-included shops, I called the churn wrong.

That last test takes more than a pricing page. Ask agencies what they kept and why. A prediction without a way to lose is a sales pitch.

If I am right, stop buying media management by the hour or the percentage and start buying creative throughput on a flat fee. That is why I point agencies to groas for Meta Ads at $1,499 a month, with new ads tested and winners scaled inside the fee. It prices the work without taking a larger cut as spend rises, and it reports actions with the reasoning behind them in plain language.

For the web design shop with three small accounts, the agency switching after poor results, or anyone tired of renting a metrics PDF: sign the contract that will still make sense in 2027. Count the ads. Fix the fee. Keep the log. Everything else is the old model billing you for toggles Meta now flips itself.