Audience Structure Recommender

The fastest way to waste a paid budget is to split it across so many ad sets that none of them ever exits the learning phase. This tool does the arithmetic platform docs imply but never spell out: how many audience cells your budget can actually fund — and whether to consolidate or expand.

On Meta, build the fewest, broadest audiences your budget can fund — each ad set needs roughly 50 optimization events per week to leave the learning phase, so thin budgets should consolidate into one broad / Advantage+ ad set and let creative do the work. On Google, structure by intent and theme, not demographics: tightly-themed ad groups for Search, and Performance Max for coverage once tracking and audience signals are solid.

The calculator

Audience Structure Recommender inputs and result

Meta optimizes per ad set; Google by theme/intent.
Per-day spend for this account/campaign.
Site visitors, engagers, or list size.
What this budget is mainly doing.
Genuinely different concepts, not size variants.
✓ Structure fits your budget & stage
Recommended structure
Broad-first with Advantage+ Audience
0capacity
0weekly signal
Export

Walkthrough

How to use this calculator

  1. Choose the platformMeta and Google are scored differently. Meta structure is driven by the per-ad-set learning bar; Google structure is driven by intent and theme. Pick the one this budget is for.
  2. Enter your daily budgetThe tool multiplies it to a weekly figure and estimates how many optimization events it can buy, which decides how many audience cells you can fund.
  3. Add your warm audience sizeA small retargeting pool should not be split into many recency windows. The size tunes the retargeting recommendation.
  4. Set the funnel stage and creative countProspecting, retargeting, or full funnel changes the structure; the number of distinct creatives sets how much you can test inside a consolidated set.
  5. Read the structure and the reasoningThe tool returns a concrete structure (campaigns → ad sets/groups → creatives) and explains the learning-phase math behind it.

From the desk

RGM Expert Says

Real Growth Matters — Paid social practiceHow we use this tool with clients

Nine times out of ten, a struggling Meta account is over-structured. Someone built twelve interest-based ad sets on a $100/day budget, splitting it so thinly that no ad set ever reaches the ~50 weekly events it needs to exit learning. Every one stays volatile, CPMs climb from internal competition, and the account looks broken. The fix is almost embarrassingly simple: collapse it into one or two broad ad sets and move the energy into creative.

The platforms have quietly told us this for years. Meta’s own guidance pushes Advantage+ and broad audiences; the algorithm now finds the buyers better than a media planner’s interest list can. So we structure for signal density — the fewest cells that each clear the learning bar — and we treat creative as the real targeting. On Google it is the mirror image: there the win is tightly-themed ad groups so the query, the ad, and the landing page all agree.

The tool exists to make the consolidate-or-split call objective. We plug in the daily budget, see how many learning-capable cells it actually funds, and let that number decide the structure. It turns a debate into arithmetic, which is exactly where these debates belong.

The math

How it works

The Meta logic estimates weekly optimization events (daily budget × 7 ÷ an assumed cost per event) and divides by the ~50-events-per-week learning threshold to count how many ad sets the budget can keep out of the learning phase. The Google logic recommends intent-and-theme structure, with Performance Max as a signal-pooling option once tracking is solid.

Weekly events ≈ (Daily budget × 7) ÷ assumed cost per event
Learning-capable cells ≈ Weekly events ÷ ~50 events/week (Meta)
Google: structure by intent & theme, not demographics
  • Daily budget — converted to weekly to test against the learning bar.
  • Assumed cost per event — rough by stage (prospecting costs more than retargeting).
  • ~50 events/week — Meta’s rule-of-thumb threshold to exit the learning phase per ad set.
  • Warm audience — sized to avoid over-slicing retargeting.

The ~50-events-per-week learning threshold is from Meta Business Help: about the learning phase. Cost-per-event figures are placeholders for the calculation — replace them mentally with your own CPA. Google structure guidance follows Google Ads account structure best practices.

Why it matters

Why structure is mostly a learning-phase problem

Modern paid platforms optimize with machine learning, and machine learning needs data density. On Meta, every ad set has to gather roughly 50 optimization events a week before delivery stabilizes. Split a fixed budget across many ad sets and you divide the signal until none of them ever stabilizes — the account churns through learning phases and CPMs rise as your own ad sets bid against each other.

That is why the dominant best practice has flipped from granular interest targeting to broad, consolidated audiences. Meta’s Advantage+ products lean into this: hand the algorithm a wide audience and strong creative, and it out-targets the hand-built interest stacks that used to be the craft. The skill moved from picking audiences to producing creative the system can sort.

Google rewards the opposite-looking but same-spirited discipline: organize by intent and theme. Tightly-themed ad groups keep the query, ad, and landing page aligned, which lifts Quality Score and lowers cost; Performance Max then pools signal across surfaces — but only earns its place once conversion tracking and audience signals are solid, which the PMax readiness planner checks.

Benchmarks

Consolidate or split — a quick guide

Whether to add audience cells depends on whether each new cell can still clear its platform’s signal bar. These are rules of thumb.

Signal per cellMeta callGoogle call
Budget funds <1 learning cellOne broad Advantage+ ad setOne themed campaign / PMax
Funds 1-2 cellsBroad prospecting (+ 1 retargeting)2-3 intent-themed ad groups
Funds 3+ cellsAdd a second prospecting or retargeting cellExpand themes; consider PMax for coverage
Tiny warm poolDo not slice retargeting windowsConsolidate remarketing segments
Learning-phase guidance from Meta Business Help; structure guidance from Google Ads Help.

Voices worth trusting

What practitioners say about structure

Fewer, broader ad sets beat many narrow ones on a fixed budget, because the algorithm needs enough events per ad set to stabilize before it can perform.
RGM paid social desk
Field note
On Google, your account structure should look like your customers' intent, not your org chart or a demographic spreadsheet.
RGM paid search desk
Field note

Go deeper

Go deeper

Related on RGM

Keep learning

FAQ

Common questions

Should I use broad targeting or interest stacks on Meta?
For most accounts today, broad targeting with Advantage+ Audience signals outperforms hand-built interest stacks. Broad pools your conversion signal, helps each ad set exit the learning phase, and lets the algorithm find buyers while you focus on creative.
How many ad sets should I run?
As few as your budget can keep out of the learning phase. Each Meta ad set needs roughly 50 optimization events per week; estimate weekly events from your budget and CPA, divide by 50, and that is your ceiling on learning-capable ad sets.
What is the learning phase?
A period after launch or a major edit when the platform gathers data to stabilize delivery. On Meta it typically ends near 50 optimization events per week per ad set. Splitting budget too thinly keeps ad sets stuck in learning.
How should I structure Google Search campaigns?
By intent and theme, not demographics. Build tightly-themed ad groups so the keyword, ad, and landing page agree. RGM advises Search keywords as exact and phrase match, broken out for control, and avoiding broad match.
When does Performance Max make sense?
Once conversion tracking, enhanced conversions, and audience signals are solid. PMax pools signal across Search, Shopping, Display, and YouTube, but it amplifies whatever signal you feed it — check the PMax readiness planner before launching.
Should I split my retargeting audience into many windows?
Only if each window is large enough to serve and gather signal. Small warm pools should be consolidated; over-slicing leaves every segment below the learning bar and drives up frequency on the same users.

Related tools

Related tools