Choosing a Growth Channel

A growth channel is useful only when it can repeatedly put the business in front of the right people, create a credible path to purchase, and produce evidence about the actual bottleneck.

The mistake is treating channels as a list of tactics to collect.

Before choosing one, ask:

• Who is the customer?
• What triggers them to look for a solution?
• Where do they already learn, compare or buy?
• Who controls the decision?
• What proof do they need?
• What does a successful customer cost to acquire and serve?

More acquisition is not always the problem.

A business can have:

• weak awareness;
• enough traffic but poor conversion;
• good conversion but poor retention;
• strong demand but inadequate delivery capacity;
• a positioning problem;
• a pricing problem;
• a customer-fit problem.

If the bottleneck is retention or offer clarity, adding another acquisition channel can increase waste rather than solve the underlying problem.

List the channels that could realistically reach the target customer.

Depending on the business, these may include:

• founder-led outbound;
• referrals;
• partnerships;
• search;
• communities;
• events;
• content;
• email/newsletters;
• paid acquisition;
• marketplaces;
• direct sales.

For each, record:

Reach: Can we reach enough relevant people?

Access: Can we actually get their attention?

Trust: What evidence do they need before acting?

Conversion path: What happens between first contact and purchase?

Economics: What does the channel require in time, money and operational capacity?

Learning value: What will we learn even if the experiment does not produce immediate revenue?

Do not try to “do marketing” across every channel at once.

Choose one important uncertainty.

Examples:

Hypothesis: A specific customer segment will respond to a problem-specific outbound message.

Test: Run a small, relevant outreach experiment.

Measure: Qualified conversations and downstream buying behaviour, not replies alone.

Or:

Hypothesis: Existing customers know other people with the same problem.

Test: Ask for relevant introductions after a successful outcome.

Measure: Qualified introductions and resulting customer behaviour.

The test should match the uncertainty.

A channel can look successful while producing poor customers.

Track, where practical:

• qualified opportunities;
• conversion;
• time to decision;
• acquisition cost;
• gross margin or contribution;
• retention;
• expansion or repeat purchase;
• support burden;
• founder time;
• learning generated.

Do not optimize a single metric such as clicks, impressions or reply rate when the business outcome happens later.

Before starting the experiment, define:

Continue: What evidence would justify another test?

Change: What result would suggest the message, offer, segment or channel needs modification?

Stop: What evidence would make further investment unreasonable?

Do not set arbitrary universal thresholds. A sensible stopping point depends on the economics, sales cycle, customer volume and cost of learning in the specific business.

Every channel consumes scarce resources.

A channel may offer inexpensive reach but require heavy founder involvement. Another may be slower to start but produce better customer fit. Paid acquisition can provide speed while introducing cash risk. Referrals can be efficient while depending on an existing customer base. Outbound can create direct learning while demanding careful targeting and relevant messaging.

The right comparison is not “which channel is best?”

It is:

Which channel gives this business the best next learning and commercial test for the resources we can responsibly spend?

Before committing meaningful budget or founder time, write:

1. Customer: Who exactly are we trying to reach?
2. Problem: What situation makes them care now?
3. Channel hypothesis: Why should this channel reach them?
4. Test: What is the smallest credible experiment?
5. Metric: What downstream behaviour matters?
6. Cost: What will the experiment consume?
7. Stop rule: What result would make us stop or change?
8. Review date: When will we decide what the evidence means?

Y Combinator's startup guidance emphasizes direct customer learning, building something people want, and avoiding premature scaling. Those principles support testing channels through real customer behaviour rather than assuming a tactic will work because it is popular.

Evidence boundary: No single channel or universal channel cadence works for every business. Use the company's customer, buying process, economics and observed results to make the decision.

Current bottleneck: [Awareness / conversion / retention / fit / delivery / other]

Target customer: [Specific segment]

Channel hypothesis: [Why this channel should reach them]

Smallest test: [What will we do?]

Evidence: [What happened?]

Trade-off: [What did it cost in money, time or attention?]

Next decision: [Continue / change / stop]

Review date: [When will we decide?]