THE CHANNEL SELECTION MATRIX
Channel selection is the discipline of finding where your customers actually are and meeting them there, then stacking channels so they amplify each other. Get it right and every pound works harder. Get it wrong and you’re shouting into an empty room. I have done both.
The Four Channel Families
1) Product-Led Channels
In-product merchandising, email, push notifications, SMS.
Use when
- You already have users/customers
- Cross-sell or upsell is meaningful
- Budget is tight
Avoid when
- You need net-new customer acquisition
- You have no active user base
2) Partner-Led Channels
Display, affiliate, and retail ad networks (Amazon, DoorDash, etc.).
Use when
- You want to test many approaches quickly
- Partners have your audience
- You can monitor fraud and quality
Avoid when
- You can’t monitor fraud
- You need highly precise measurement
- Fees crush thin margins
3) Search (SEO + SEM)
Organic search optimisation plus paid search ads.
Use when
- People search for your product/category
- You can sustain content + SEO investment
- Purchase intent is real
Avoid when
- No one searches for what you offer
- You can’t invest long-term
- You trust last-click blindly
4) Social (Organic + Paid)
Organic social content plus paid social advertising.
Use when
- You need versatility across funnel stages
- Precise targeting matters
- Measurement rigor matters
Not to Be Confused With
📺 Channel vs tactic
Social is a channel; a viral video is a tactic. Tactics expire. Dollar Shave Club’s famous launch video was a moment. Channels compound. Build channel capability, not campaign one-offs.
📱 A new channel vs a new surface
When technologies emerge they’re treated as distinct channels; as they mature they get subsumed. Mobile was once a channel. Now it’s a surface every channel runs on. Don’t build a “TikTok strategy”; build a social strategy that includes TikTok.
💰 Paid vs organic in one family
They serve different funnel stages and need different measurement. Don’t grade paid social on the same yardstick as your organic community. Don’t let last-click attribution tell you brand spend does nothing.
Channel Stacking (Use Them Together)
Launch Stack
Social (paid) for awareness + Search (SEM) to capture demand + Product-Led to convert and retain.
Demand Capture Stack
SEO for organic discovery + SEM for paid capture + Email to nurture and re-engage.
Growth Stack
Product-Led (merchandising + email) + Organic social for community + Partner-Led for incremental reach.
Full Funnel Stack
All four channel families coordinated so brand demand amplifies direct response performance.
The 5-Step Selection Process
-
Assess the market. I copied what worked for similar businesses. But I knew I was seeing their winners, not their failed experiments. Some of the urge to copy is just FOMO, and fear of missing out is not a channel strategy.
-
Try affiliates. I let affiliate partners explore channels for me. They found pockets of demand I would never have reached directly.
-
Systematically explore. I tested other channels myself, from easiest to hardest. The order was onsite ads, then Google paid search, then in-product promotion: not because that was the grand plan, but because the first two needed almost no engineering support. Onsite linking needed engineers to change the core product, and nobody hands you that on day one. I had to earn credibility on the first two before anyone would let me near the product.
-
Double down. When I found what worked, I exploited it: concentrated spend where marginal returns were highest.
-
Rinse and repeat. Channels evolve and saturate, so I revisit the whole process regularly. This is the step everyone skips, and it is the most important one.
The Two Golden Rules
-
Start at the bottom of the funnel. Channels closest to conversion usually give the fastest early ROI.
-
Measure incrementally. Don’t rely on attribution alone; test what actually changes because of your spend.
Budget Allocation Principle
Don’t optimise for highest average return per channel. Optimise for highest marginal return across all channels. Your next dollar goes where incremental return is highest.
Know Its Limits
My position: “copy what works” is a starting point, not a strategy. It has a survivorship bias baked in. You see competitors’ winning channels (the Super Bowl ads, the viral videos), not the graveyard of channels they tried and quietly abandoned. And a channel that works for a competitor at a different funnel stage, or with different unit economics, may not work for you at all.
Channels also saturate. Doubling down works until it doesn’t: rising costs and audience fatigue erode every channel eventually. The brands that keep growing are the ones still exploring while their current winners are working. Not the ones scrambling when they stop.
And as you scale across families, measurement gets genuinely harder. Don’t let attribution arguments stall the testing. A rough incrementality read beats a precise fiction.