Every agency has an opinion about this, and most of the opinions are really about which platform the agency prefers to manage. The useful answer starts somewhere else entirely — with a question about your category.
The question that decides it
Does anyone already search for what you sell?
If a plumber's customer wakes up to a flooded kitchen, they type "emergency plumber near me". The demand exists; it just needs capturing. That is Google's job, and it does it better than anything else ever built.
If you sell a product nobody knows they want — a new category, a better version of something people tolerate, a gift — nobody is searching for it. There is no demand to capture; there is demand to create. That is Meta's job, and Google search will sit there spending almost nothing because the queries do not exist.
| Signal in your business | Start here |
|---|---|
| Customers describe a problem and go looking for a solution | Google Search |
| Urgency: broken, blocked, deadline, emergency | Google Search |
| Competitors already bid on obvious keywords | Google Search |
| Visual product, impulse or discovery purchase | Meta |
| New category, or you are the cheaper alternative to a habit | Meta |
| Strong existing content or customer video | Meta |
| E-commerce with a clean product feed | Both — Shopping and Advantage+ |
The thresholds nobody mentions
Before splitting anything, check that your budget clears the minimum each platform needs to function. Below these, you are not testing a channel — you are paying for noise.
Meta needs conversion volume
Meta's delivery system optimises by learning from conversions. Its documented guidance is roughly 50 optimisation events per ad set per week before performance stabilises. If your cost per conversion is $40, that is $2,000 a week — $8,000 a month — for one ad set to exit the learning phase.
The practical consequence: with a modest budget, run one campaign and one or two ad sets. The instinct to split $3,000 across six carefully segmented audiences guarantees that none of them ever learns anything. If your true conversion is too expensive to hit that volume, optimise for an earlier event — add to cart, lead form, qualified page view — and watch the downstream number separately.
Google needs enough search volume
Google's constraint is the opposite: not your budget, but the market's. If your best keywords are searched 200 times a month in your area, you can spend $600 and then the inventory simply runs out. Check the volume before you plan the budget — a keyword tool will tell you in ten minutes whether the channel can absorb what you intend to spend.
A budget too small to clear the learning threshold does not produce a small result. It produces no result, and a wrong conclusion about the channel.
How we would split $10,000
Assuming a business with some existing search demand and a product people can see:
- $3,500 — Google Search, brand and high-intent terms only. Tight match types, an aggressive negative keyword list from day one, and every term reviewed weekly for the first month. This is the cheapest revenue in the account; it should never be starved.
- $4,000 — Meta, one prospecting campaign, broad targeting, three to five creatives. Broad beats narrow at this budget: the system finds the audience faster than you can define it, and you are not splitting volume across ad sets that will never learn.
- $1,500 — retargeting, split across both. The cheapest conversions you will find, and the easiest to over-spend on. Cap it deliberately; a retargeting campaign with no ceiling will happily buy back customers who were coming anyway.
- $1,000 — creative production. Not media. On Meta especially, creative is the targeting: three new concepts a month beats any amount of audience tinkering.
Hold that structure for eight weeks before drawing conclusions. Four weeks is the learning phase plus noise. The temptation to reallocate after ten days is the single most common way a first budget gets wasted.
What to fix before spending a dollar
Both platforms multiply what already happens on your site. If the site does not convert, paid media just buys you more evidence of that, faster.
- Conversion tracking that actually fires. Test it yourself, end to end, on a phone. A surprising share of accounts optimise toward an event that stopped working months ago.
- A landing page that matches the ad. The promise in the headline should be the first thing on the page. Sending search traffic to a homepage is the most expensive habit in small-business advertising.
- A known break-even. What can you afford to pay for a customer? Until that number exists, no campaign can be judged.
- Somewhere for leads to land. A form that emails an inbox nobody checks converts at zero, whatever the dashboard says.
The honest summary
Google captures demand that already exists. Meta creates demand that does not. Most businesses need both eventually, and almost none should start with both at full strength. Start where your customers already are, spend enough for the platform to learn something, and give it eight weeks before you decide.