One of the most practical frameworks for allocating a marketing budget at early to mid stage is the 70/30 split: 70% of available budget goes into channels that are proven and producing results, 30% goes into testing new channels, new messages, or new approaches.
The logic is straightforward but the discipline to maintain it is harder than it sounds.
Why you need both buckets
A company that puts 100% of budget into proven channels will hit the ceiling of those channels without having developed any alternatives. The channels that work today will eventually become more expensive, more competitive, or simply exhausted in terms of audience reach. When that happens, a company with no experimental pipeline has no fallback.
A company that puts too much into experimentation sacrifices the near-term revenue and growth it needs to stay solvent and hit targets. Experiments take time to produce results. Some will fail. If the failure rate is high and the experimental budget is large, the company loses momentum without generating the learning fast enough to compensate.
What goes in the 70%
The proven bucket contains channels where you have enough data to be confident about the CAC, the quality of customers produced, and the scalability of the channel within its current parameters. You understand the creative that converts, the audience targeting that works, and the landing experience that follows the click.
This budget is not set and forget. It requires ongoing management and optimisation. But the decision-making is incremental and the risk of any single decision is low. You are operating within a system you understand.
What goes in the 30%
The experimental bucket is for hypotheses you have not yet validated. A new channel you have reason to believe reaches your ICP. A new message angle you want to test against your current positioning. A new format, a new audience segment, a new partnership model.
The discipline of keeping this to 30% prevents experimentation from becoming a distraction. Each experiment should have a clear hypothesis, a defined success metric, a minimum budget to generate meaningful signal, and a predetermined point at which you will read the result and decide whether to scale or stop.
When to adjust the ratio
The 70/30 split is a starting point, not a fixed rule. If a proven channel is approaching its ceiling, the experimental budget may need to increase temporarily as you search for its replacement. If near-term targets are urgent and the experimental pipeline is not yet producing results, the proven budget may need to increase at the expense of experimentation.
What you should never do is collapse the experimental budget to zero. The companies that find themselves with no alternative to a deteriorating channel are the ones that cut experimentation during periods of pressure, repeatedly, until there is nothing left in the pipeline when the proven channel stops working.
70% scales what works today. 30% finds what will work tomorrow. Cutting the 30% solves a short-term problem and creates a much larger one six months later.