Why this matters
Small marketing budgets die by being spread evenly. After your paid test, you have numbers; the discipline that turns them into growth is a monthly review that cuts the losers and concentrates money and hours on the winner.
What "done" looks like
- A 30-minute monthly review of every channel, with cost per customer for each
- The worst performer cut or paused each month
- Freed budget and hours moved to the best performer
- 10–20% reserved for one new experiment
- Most of your spend concentrated in one or two proven channels
How to do it
- List every channel with dollars spent, hours spent, and customers produced.
- Rank by cost per customer, counting your time at an honest hourly rate.
- Cut the bottom — anything 2–3x worse than the leader gets paused.
- Reinvest in the winner until it shows diminishing returns; scale budget about 2x at a time, not 10x.
- Keep an experiment slot (10–20%) so a new channel can enter the ranking.
- Write the numbers down monthly so trends, not moods, drive decisions.
Common mistakes
- Sunk-cost loyalty: keeping a dead channel because you already built the account and audience
- Doubling down on vanity metrics — followers and impressions instead of customers
- Scaling a winner 10x overnight; most channels get more expensive as spend rises
Real-world examples
- Dropbox saw in its data that roughly a third of signups came from word of mouth, doubled down with a two-sided referral program, and grew from 100K to 4M users in 15 months.
- Google popularized 70/20/10 resource allocation — most on the proven core, a slice on adjacent bets, a sliver on experiments — a useful default for small marketing budgets too.
From a founder's point of view
Marketing looks like it rewards creativity, but at small scale it mostly rewards concentration. The winning channel is rarely the fun one. Cutting a channel you enjoy is the hardest call on this list — and usually the highest-return one.
Rule of thumb
Every month you should be able to name the channel that got more money and the channel that lost its budget. If nothing changed, you didn't review — you renewed.