Growth Channels: Choosing and Testing a Channel on a Limited Budget

The rule: test sequentially, scale with focus.

1. The channel map

• Content and search: slow to start, but a compounding asset whose effect persists.
• Paid advertising: fast and measurable, but it stops when spending stops.
• Direct sales: suited to higher prices, and scales only through hiring.
• Partnerships and distribution: access to a ready audience via a third party; requires time to negotiate.
• Referral: highest quality and lowest cost, but it is a consequence of a good product rather than a substitute for one.
• Communities and events: effective in specialised sectors and enterprise selling.
• Public relations: builds credibility more than it builds direct demand.

2. Choosing the first channel

Three criteria:

1. Where is your customer actually? Not where you'd prefer to work.
2. Does the channel fit your price? Low prices can't carry high-cost channels, and the reverse also holds.
3. Do you have an edge in it? Prior experience, a network, or the ability to produce distinctive content.

Rank your best three, and start with one.

3. Testing methodology

A sound test is designed before any spending:

• Hypothesis: "A customer can be acquired through this channel for under X."
• Minimum spend: enough to produce a statistically readable signal. Trivial spend produces results you can't judge.
• Duration: long enough for the decision cycle in your sector to complete.
• The measure: not clicks, but paying customers and their acquisition cost.
• A decision criterion written beforehand: what does success mean, and what does failure mean?

Without a criterion set in advance, you will interpret the results in whatever way suits your preference.

4. Read the whole funnel

A channel can look like a failure for reasons unrelated to it. Analyse the full path: impression → click → signup → qualified lead → paying customer → retained customer.

A drop at a late stage means a problem in the product, pricing, or qualification — not the channel. Stopping the channel in that case conceals the real problem rather than solving it.

5. When to double down and when to stop

Double down when acquisition cost settles within an acceptable range, customers from the channel are retained, and the economics remain healthy at increased volume — that last point is decisive: many channels work at small scale and break economically when scaled.

Stop when cost doesn't improve across several optimisation cycles, when customers from the channel churn at a higher rate, or when it consumes management effort disproportionate to its return.

6. When to add a second channel

Once the first is operable almost routinely — a documented process, predictable results, and no daily founder attention required. Adding a channel before that yields two mediocre channels instead of one good one.

7. Owned versus rented assets

Distinguish between what you own and what you rent:

• Owned: an email list, content on your domain, a user base, reputation. Its effect persists and depends less on a third party.
• Rented: access through an advertising platform or an algorithm. Fast, but it changes when the platform's policy or pricing changes.

Total dependence on a rented channel is a strategic risk. Allocate part of your effort to building what you own.

Common mistakes

• Spreading a small budget across many channels.
• Judging a channel on spend below the necessary threshold.
• Measuring clicks rather than paying customers.
• No decision criterion set in advance.
• Stopping a channel because of a product or pricing problem.
• Scaling a channel whose economics break at higher volume.
• Total dependence on a rented channel.

Checklist

• Top three channels ranked against clear criteria
• One channel under test at a time
• Defined hypothesis, budget, duration, and measure
• A decision criterion written before spending
• Measurement by paying customers rather than clicks
• Full funnel analysis before judgement
• Verification that economics hold at scale
• Parallel investment in an owned asset

FAQ

What is the cheapest channel to start with?
Referral and content carry low cash cost, but both require time and a good product.

How long does a channel test take?
At least twice your sector's decision cycle, so the trend can be read.

Should I hire a specialist or use an agency?
At the start, test yourself or with limited support. Delegating a channel you don't yet understand produces spending without learning.