How do you choose the right marketing channels for a campaign?
TL;DR
- Pick channels by three audience questions: where they notice, where they trust, and how often they need to hear it.
- Research complements attribution data. It helps most when you're reaching a new audience or launching something new.
- Plan for repetition. A classic view of advertising holds that about three meaningful exposures do most of the work.
- Neon Blue returns notice, trust and preferred-channel reads by segment, labeled as simulated. Run your first study free
Neon Blue builds simulations of human behavior. You choose channels by starting from the audience, not the media menu: where do the people you want to reach spend attention, which channels do they trust for this kind of decision, and how many times do they need to see the message before it sticks. Past campaign data answers some of this for audiences you already reach. For a new audience, a new product or a new message, there is no history to lean on. Neon Blue runs channel studies on a simulated audience of the buyers you want, before the media plan is locked.
What should you know about an audience before choosing channels?
| Question | Why it matters | How to ask it |
|---|---|---|
| Where do they spend attention for this category? | Reach only counts where people look | "Where did you last see or hear about [category]?" |
| Which sources do they trust here? | A seen message isn't a believed one | "Which of these would you trust for advice on [category]?" |
| What format do they prefer? | Video, audio, reviews and in-store work differently | "How do you prefer to learn about a new [product]?" |
| What triggers a purchase? | The last channel before purchase matters | "What made you buy the last time?" |
| Who influences them? | Partners, kids, friends or creators | "Who, if anyone, did you ask before buying?" |
These questions cut differently by segment, so read them by segment.
How many times does an audience need to see a message?
In 1972, Herbert Krugman published "Why Three Exposures May Be Enough" in the Journal of Advertising Research. He argued that the first exposure prompts "what is it?", the second prompts "what of it?" as people judge whether it's relevant, and the third works as a reminder when they're ready to act. Further exposures mostly repeat the third.
The point for channel planning: spread too thin across many channels, and most people see the message once and never get to "what of it?". It is often better to own two or three channels where your audience will see the message several times than to appear once everywhere. Research on where your audience spends attention tells you which two or three.
Which channel planning questions can a study answer?
- In the past month, where have you seen or heard about [category]? [Channel list]
- Which of these would you trust most for a recommendation on [category]?
- If [brand] wanted to tell you about something new, where would you most likely notice it?
- Which of these would feel intrusive coming from [brand]?
- After seeing this message in [channel], what would you do next?
Question 4 often gets skipped. Channels that feel intrusive can hurt a brand even when they reach people. The creative itself should be pretested separately.
What does a channel study look like on a simulated audience?
Hypothetical: a sportswear brand is launching a walking shoe aimed at adults over 50, an audience it has rarely targeted, and needs to choose three channels.
- Audience: US adults aged 50 to 70 who walk for exercise at least three times a week, split by those who buy athletic shoes online and in store, and by those who have and haven't bought the brand before.
- Spec: built the way we describe in how to build a synthetic audience, with age and region matched to census targets for the group.
- Design: attention, trust and intrusiveness questions across eight candidate channels, then a message read in the top three.
- Readout: channel ranking by notice and trust for each segment, what feels intrusive, and the next step people would take.
Example output layout (illustrative, not a real result)
| Channel | Would notice (top 2) | Would trust | Feels intrusive | Next step after seeing it (coded) |
|---|---|---|---|---|
| Channel 1 | __% | __% | __% | Coded theme |
| Channel 2 | __% | __% | __% | Coded theme |
| Channel 3 | __% | __% | __% | Coded theme |
If a channel scores high on notice but low on trust, it can carry awareness but not the persuasion work. That split shapes which message goes where, using claims you have already tested.
How does channel research fit with attribution and media mix data?
- Attribution shows what worked for audiences you already reached. Research shows what might work for ones you haven't.
- Media mix models need history. For a new product or audience, a channel study fills the gap until real data builds up.
- Test before scaling. Use the study to choose channels, run a small live flight, then scale what works.
For how channel choices connect to brand-building versus short-term activation, see what research a brand manager needs before a launch. For a big campaign, add live interviews with real buyers on the final plan and keep them separate from simulated reads. The launch decision itself is covered in launch go or no-go.
Questions about choosing marketing channels
How do you decide which marketing channels to use?
Start with your audience: where they spend attention for your category, which sources they trust, and what triggers a purchase. Then pick the two or three channels that score well on all three.
How many marketing channels should a campaign use?
Fewer than most plans include. Enough frequency in two or three channels usually beats a single exposure across many.
Can you research channels for an audience you have never marketed to?
Yes. That is where research helps most, because there is no campaign history. A simulated audience of the new buyers gives a first read before spending.
How does Neon Blue help with channel planning?
Neon Blue builds a simulated audience of the buyers you want to reach and reads notice, trust and intrusiveness by channel and segment, labeled as simulated. Start a free first study.