Why very small budgets fail structurally
This is the part most first-time sellers do not know: ad platforms need a minimum number of conversions to optimise. Meta and Google both use machine learning to decide who sees your ad, and that learning requires conversion volume — roughly 50 conversions per week per ad set is the commonly cited threshold for Meta.
If your cost per acquisition is ₹250 and you spend ₹300 a day, you generate about one conversion daily. Seven a week. The algorithm never exits its learning phase, so it never optimises, so your cost per acquisition stays high, so you conclude ads do not work.
The budget was not too small to be profitable. It was too small to be measurable.
A realistic starting structure
₹1,000–₹2,000 per day is the practical floor for a new Indian D2C brand. That is ₹30,000–₹60,000 a month.
Split it roughly:
- 70% prospecting — cold audiences, broad targeting, letting the platform find buyers
- 30% retargeting — site visitors, cart abandoners, engaged social audiences
Retargeting will show a much better return. Resist the urge to move everything into it — retargeting only works on traffic that prospecting created. Shift the budget entirely and the pool dries up within days.
What the first month actually looks like
Week 1 — Learning. Costs look terrible. Cost per acquisition may be two or three times what you hoped. This is normal and not a signal to stop.
Week 2 — First signal. Some creative and audience combinations start pulling ahead. Kill the clear losers, keep everything else running.
Week 3 — Consolidation. Move budget toward what is working. Cost per acquisition should start descending toward something sustainable.
Week 4 — Assessment. Now you have enough data to judge honestly. If ROAS is approaching your break-even multiple, scale carefully. If it is nowhere close, the problem is more likely the product, the price, or the landing page than the ads.
Judging an ad account before week three is the most common expensive mistake we see.
Work out your break-even multiple first
Before spending anything, calculate the ROAS you need to break even:
Break-even ROAS = 1 ÷ gross margin
At 50% gross margin you need 2.0x. At 33% margin you need 3.0x. At 25% you need 4.0x.
Then adjust for RTO. If 25% of your orders come back, your effective revenue is 75% of what the ad platform reports, so a reported 4.0x is really closer to 3.0x. Indian D2C brands routinely miss this and think they are profitable when they are not.
What to have ready before spending
- At least three distinct creative concepts, not three crops of one image
- A landing page that loads fast on a cheap Android phone over 4G
- Conversion tracking verified — a test purchase that appears correctly in the platform
- Stock for the orders you are about to generate
- A month of budget you can afford to lose, because the first month is tuition
The honest caveat
There is no universal correct number. ₹1,000 a day is right for a ₹1,200 average order value in a normal category. A ₹15,000 average order value in a considered-purchase category needs a different budget and a much longer measurement window.
The principle that does generalise: spend enough to learn, for long enough to learn it. Anything less is not a cheap test — it is an expensive non-answer.
How this fits our model
We run ads inside your own ad accounts, on your card, with no markup on spend. We are paid on net sales after returns. That means wasted ad spend costs us as directly as it costs you — which is the entire reason we prune aggressively rather than letting an underperforming campaign run.
Sources
Learning-phase conversion thresholds are documented in Meta’s own advertiser guidance. Budget ranges and the first-month sequence reflect standard practice across the accounts we run. Break-even ROAS arithmetic is arithmetic — substitute your own margin.