Almost every paid-media proposal we replace looks the same: 70-80% into Google Search, a token Meta retargeting campaign, and zero room for creative production. It performs fine for a quarter, then plateaus, because bottom-of-funnel keywords have a ceiling — once you've captured everyone actively searching, growth stops.
Below is the exact allocation we run for a typical $10,000-per-month paid media budget on a Sydney service business with a 30-90 day sales cycle. Every line has a job, every line is measured, and the split changes deliberately based on stage and seasonality.
The default split: where every dollar goes
Our standard starting allocation: 45% Google Search non-brand ($4,500), 15% Google Search brand defence ($1,500), 25% Meta prospecting with video creative ($2,500), 10% retargeting across Meta + Google Display ($1,000), and 5% reserved for testing new channels each quarter ($500).
That 5% test budget is non-negotiable. It's where we trial YouTube shorts, Reddit ads, TikTok Spark Ads, or LinkedIn for B2B. Most tests fail. The ones that win get promoted into the core mix the following quarter and the split rebalances.
Why we cap brand defence at 15%
Brand-keyword campaigns have the highest ROAS on the spreadsheet — and they're the most over-funded line item in the industry. The dirty secret is most of those clicks would have arrived organically as the #1 organic result. You're paying Google to intercept traffic you already earned.
We cap brand at 15% purely as a defensive moat against competitor bidding. If a competitor isn't bidding on your brand name, drop the cap further. Reallocating $500-$1,000/month from brand to prospecting is one of the fastest ways to unlock net-new pipeline without raising total spend.
Google Search non-brand: the workhorse
The 45% allocated to non-brand Search funds 15-25 tightly-themed ad groups, each mapped to a specific service or intent. We split by match type discipline — exact match for the proven money keywords, phrase match for discovery, broad match only when paired with a smart bidding strategy and a clean conversion signal.
Conversion tracking is the foundation. If you're optimising to form fills, you'll get form fills. If you're optimising to qualified leads (booked calls, qualified opportunities, closed revenue uploaded via offline conversion imports), you'll get qualified pipeline. Most accounts we audit are still optimising to form fills and wonder why the lead quality is terrible.
Meta prospecting: 25% and underpriced
Meta is still the most underpriced attention on the internet for service businesses. The 25% allocation funds broad-targeted prospecting campaigns with video-first creative — short, vertical, authentic, made for a phone, not a boardroom.
We don't run static images in 2026 except as retargeting reinforcement. Every prospecting dollar goes behind 15-30 second video variants, refreshed every two weeks. The creative is the targeting — Meta's algorithm finds the right person if the creative resonates with the right person.
Retargeting: 10% is enough
Most agencies massively over-spend on retargeting because it looks like the highest-ROAS line. It's not. You're showing ads to people who were already going to buy. The lift is real but small — usually 5-15% incremental conversions on top of organic return visits.
We cap retargeting at 10% split across Meta carousel + Google Display, with frequency capped at 3-4 impressions per user per week. Beyond that, you're annoying warm leads, not converting them. Tight audience windows too — 7, 14, and 30 day windows segmented by page depth, not a single 180-day catch-all.
The 5% test budget: where the next 12 months come from
Channels mature and get expensive. The brands still winning in five years are the ones running quarterly tests on whatever's currently undervalued. In 2026 our active test queue includes YouTube Shorts as a Meta-style prospecting channel, Reddit promoted posts for B2B, and TikTok Spark Ads for any brand with strong native creative.
Set the test budget at $500-$1,000/month, give each test 60-90 days to produce a clean read, and have a pre-agreed metric for promote/kill decisions. Without that discipline you'll keep two dozen mediocre channels alive and starve the winners.
The single biggest lever isn't the split — it's creative volume
We ship 8-12 new ad variants every fortnight per client. Channels are commodities; ideas are not. The agency that wins isn't the one with a clever bid strategy — every account that hits scale is using broad match + smart bidding now. The agency that wins is the one shipping ten times the creative volume of its competitors.
Budget for production from day one. Allocate 15-25% of total spend off the top to creative — ideally a fixed retainer with a video editor or a content studio. The accounts that produce one ad a month and wonder why CPL is rising are competing against accounts producing one ad a day.
When to scale and when to hold
Scale rules: never more than 20% week-on-week on an individual ad set, only on ad sets currently below CPL target, and only after the campaign has cleared the learning phase (50+ conversions in the last 7 days). Hold rules: if blended CPL drifts more than 25% above the rolling 30-day average, pause and diagnose creative first, audience second, landing page third — never raise bids as the first response.
If you want us to audit your current split or build this allocation for you, get in touch. More tactics over on the WebRise Learn blog.