How do you decide where to spend a limited marketing budget?

Show a method rather than a channel preference.

• Start from the objective and the funnel stage. Building demand for a new category needs different channels from capturing existing demand. Search captures intent that already exists; social and content create it.
• Follow the data you already have. Which channels have historically produced customers — not leads, customers — at what cost, and with what retention. Channels differ enormously in lead quality, and cost per lead is a misleading metric on its own.
• Protect a test budget. A common split is roughly 70% on what works, 20% on scaling what shows promise, and 10% on genuine experiments. Without the last portion you never find the next channel.
• Consider payback period, not just cost. A channel with a higher acquisition cost but faster payback may be better for cash flow than a cheaper one that takes a year to recover.
• Account for the compounding channels. SEO and content cost more up front and cost less over time; paid stops the day you stop paying.

Note: Being willing to cut a channel that a stakeholder is attached to, on evidence, is what distinguishes a marketer from an executor.