How Much Should You Actually Spend?
“What budget do I need?” is the wrong first question. The right one is: how much does Meta need to learn before it can perform? Answer that, and your budget almost sets itself.
By the end of this lesson you’ll know
- How to work out a realistic minimum budget
- Why spreading a small budget thin quietly kills results
- A sensible timeline before you judge performance
- The three things that waste small budgets fastest
Start from the learning phase, not a guess
From the last lesson: an ad set needs roughly 50 results within about a week to exit the learning phase and stabilize. That single fact drives your minimum budget. Work backwards from your cost per result:
Your rough weekly minimum = (cost per result) × 50. If a lead costs you about ₹100, an ad set optimizing for leads wants roughly ₹5,000/week to learn cleanly. Cheaper actions (a click, a video view) need far less; a purchase usually needs more.
You won’t know your exact cost per result on day one — that’s fine. Estimate from your product price and margins, start there, and adjust once real data arrives.
Match the budget to the objective
Softer goals are cheaper to learn on because the “result” is easy. Harder goals cost more because the action is rarer. Rough starting points for a small business testing the waters:
| Objective | Result being counted | Learns on a smaller budget? |
|---|---|---|
| Traffic | Link clicks | Yes — cheapest to learn |
| Engagement | Interactions / messages | Yes |
| Leads | Form submissions | Moderate |
| Sales | Purchases | Needs the most — rarest action |
The biggest small-budget mistake
Spreading ₹10,000 across five ad sets at ₹2,000 each feels productive. It’s the opposite. Now none of them get enough data to exit learning, so all five stay expensive and unstable. One well-fed ad set beats five starving ones. Concentrate first; expand only once something works.
Would your total budget give at least one ad set ~50 results a week? If not, you have too many ad sets, or a goal that’s too expensive for your budget. Simplify until the math works.
Set an honest timeline
Meta ads are not a slot machine. A fair test looks like this:
- Days 1–3: Learning. Costs look scary. Do not touch it, do not judge it.
- Week 1: Ad set works to exit learning. Delivery starts to settle.
- Weeks 2–4: Now you have real data worth reading and optimizing.
Judging a campaign after two days is like weighing yourself mid-meal. Give it the window it needs.
The three things that waste small budgets
- Too many ad sets — data gets diluted, nothing learns.
- Editing too often — every change restarts the learning phase.
- Weak creative — from the auction lesson, poor creative raises your cost on every single impression.
Your first campaign is tuition, not a jackpot. Treat the early budget as the cost of learning what works for your business — which audience, which creative, which offer. Once you find that, then you scale. Expecting a 5× return in week one is how good campaigns get switched off too early.
Key takeaways
- Set budget from the learning phase: roughly cost-per-result × 50 per week.
- Cheaper objectives (traffic, engagement) learn on less; sales needs more.
- One well-fed ad set beats several starved ones — don’t spread thin.
- Give a fair test 2–4 weeks, and treat early spend as learning, not a jackpot.