Marketing budgets look bulletproof in January. By June, half your channels are underperforming, two platforms changed their algorithms, and your broker is asking why lead cost jumped 40%. This is the moment most agents freeze spending or slash everything equally—both mistakes. A mid-year audit separates the 30% of your budget generating 70% of your deals from the dead weight you're funding out of habit.

The exercise isn't about cutting for cutting's sake. It's about reallocating dollars from low-yield activities to the channels already working, then testing one or two new tactics with the freed-up capital. Here's the framework Ontario, BC, and Alberta agents are using to stress-test their marketing spend before Q3.

Pull Your Actual Cost-Per-Acquisition by Channel

Start with a spreadsheet. Column A: every marketing channel you've spent money on since January 1. Column B: total spend per channel. Column C: closed deals attributed to that channel. Column D: cost per acquisition (CPA). Most agents skip Column C because attribution feels messy—someone saw your Instagram ad, then Googled you, then called off a pillar page. Assign the deal to the last touch for now. Perfectionism kills this audit.

Your targets depend on average commission, but rough benchmarks for 2026: Facebook/Instagram lead gen should land under $800 CPA in the GTA, under $600 in Calgary, under $1,000 in Metro Vancouver. Google Local Services Ads run $400–$700 per qualified lead in most markets, but conversion rates vary—track to close, not just lead. Sphere marketing (past-client emails, event mailers, birthday cards) should cost under $200 per transaction if you're working a database over 150 names. Zillow Premier Agent and similar portals: $1,200–$2,500 CPA is standard, but if you're above $3,000, you're subsidizing someone else's deal flow.

Anything with a CPA more than double your channel average gets flagged. If you've spent $4,000 on a channel with zero closed transactions by June, it's a cut unless you have a documented 90–120 day sales cycle and confirmed pipeline.

Audit Platform Changes That Broke Your Funnel

Meta's 2025 shift to Advantage+ campaigns reduced granular targeting. Google's AI-driven Performance Max absorbed Smart Shopping and Local campaigns. If your cost-per-lead spiked in Q2 and you haven't adjusted creative or audience signals, you're fighting last year's algorithm with last year's playbook.

Run a 30-day test on updated formats. For Meta, that means feeding the pixel 50+ conversion events (lead form submissions, landing page visits, calls) to exit the learning phase faster. For Google LSA, it means tightening your service area if you're getting leads 40 minutes outside your farm area—LSA charges per lead, not per click, so geography creep kills ROI. For YouTube pre-roll (increasingly viable in markets with older buyer demos), test 15-second non-skippable ads with a direct CTA instead of 60-second brand spots.

If a platform changed and you didn't, assume the channel is underperforming due to operator error, not platform failure. Retest before you cut.

Identify Your One Unfair Advantage and Fund It Properly

Every top-producing agent has one channel that outperforms market averages by 2–3x. It's usually tied to a specific skill, relationship, or content engine they've built over 18+ months. A Mississauga agent might close 60% of deals from a Punjabi-language YouTube channel. A Vancouver agent might generate 40% of listings from a hyper-local newsletter covering strata council decisions in False Creek. An Edmonton investor-focused agent might get 70% of referrals from a private Facebook group for accredited investors.

This channel rarely gets the budget it deserves because it doesn't look like "marketing" in the traditional sense. It's not a paid ad. It's not a CRM drip sequence. It's the thing you do that nobody else in your market can replicate easily. If your unfair advantage generated 8 deals in H1, model what happens if you double the effort in H2—hire a VA to handle production, increase posting frequency, sponsor a related event. Allocate 30–40% of your total marketing budget here, even if it means pulling from "proven" channels like Zillow or Facebook.

The math: if your average commission is $12,000 and your unfair advantage channel has a $400 CPA, every $1,000 invested returns $30,000 in gross commission. You're not capped by market size—you're capped by production capacity.

Cut Anything You Can't Measure or Defend in 60 Seconds

Sponsoring a local hockey team: can you name three clients who came from that sponsorship? Bench ads on bus shelters: do you track the vanity URL or unique phone number printed on the ad? The neighbourhood newsletter that costs $600/month: have you received a single inbound call with "I saw your ad in the newsletter" as the source?

Brand awareness has value, but not if you're spending 20% of your budget on it while your cost-per-lead in measurable channels is climbing. The test: if your broker or accountability partner asks, "Why are we spending $X on this?" and you need more than 60 seconds to justify it with data, it's a cut candidate. Redirect that budget to channels where you can draw a straight line from spend to signed BRA.

Eliminate any "because I've always done it" line items—market conditions in 2026 don't reward inertia
Kill channels where your cost-per-lead exceeds your average commission (yes, this happens with poorly managed PPC)
Pause brand-building tactics until your cost-per-acquisition in performance channels is under control
Stop paying for leads you don't follow up on within 4 hours—if your conversion rate is under 2%, the channel isn't the problem

Reallocate 15% to One New Test Before Q3

Once you've cut the bottom 20% and doubled down on your unfair advantage, take 10–15% of your revised budget and test something new. Not a random experiment—a calculated bet based on where your ideal client is underserved. If you're in a condo-heavy market and nobody's running TikTok ads targeting first-time buyers aged 26–34, test $1,000 over 30 days. If you're in a luxury segment and your competitors aren't sending dimensional mail (actual printed magazines, market reports in custom folders), test a 90-day campaign to 200 high-net-worth households.

The goal isn't to find the next Meta. It's to avoid the trap where your top two channels mature, CPMs rise, and you have no alternative ready. Set a hard stop-loss: if the test doesn't generate at least one qualified appointment per $500 spent within 60 days, kill it and reallocate.

Run your audit this week. Pull the numbers, flag the cuts, move the money. The agents who treat their marketing budget like a portfolio—rebalancing quarterly, cutting losers fast, compounding winners—are the ones who'll hit their annual GCI target even if the market stays flat. Everyone else is hoping their January plan magically works in a June reality.