How to Test a Marketing Channel With $100 A few months ago, a friend of mine who runs a small candle-making business out of her garage asked me a question I couldn’t answer with confidence: “Which marketing channel should I try next?”
She’d been selling almost entirely through word of mouth and a handful of local craft fairs. Her Instagram had a few hundred followers, mostly friends and family. She had $100 to spend on marketing that month and no idea whether to put it into Facebook ads, Pinterest ads, Google Search ads, or just boosting a few posts.
So we decided to actually test it instead of guessing. What follows is a worked example based on that process—the decisions, the numbers, and the mistakes we almost made—laid out as a framework you can apply to your own $100 marketing experiment, whether you’re running a candle business, a freelance service, or a tiny SaaS product.
Why $100 Is Actually a Useful Budget
It’s tempting to think $100 is too small to learn anything. In practice, it’s almost the perfect size for a first test, for one simple reason: it forces you to be specific.
When you only have $100, you can’t afford to run three ad sets, five audiences, and four creative variations at once. You’re forced to pick one channel, one audience, one offer, and one piece of creative—and actually see how it performs. That constraint is a gift. Most marketing failures I’ve seen come from spreading a budget too thin across too many ideas, so nothing gets a fair test.
A $100 test isn’t meant to make you rich. It’s meant to answer one question: “Does this channel deserve a bigger budget next month, or not?”
Step 1: Pick One Channel, Not Three
My friend was tempted to split her $100 across Facebook, Pinterest, and Google. We talked her out of it. Splitting $33 across three channels means each one is too small to produce a signal you can trust—you’ll just get noise, and noise is expensive because it teaches you nothing.
We looked at where her actual customers already spent time. Candles are visual, gift-driven, and often searched for around specific occasions (weddings, holidays, self-care gifts). That pointed toward Pinterest, where people actively search and save ideas for exactly those occasions, rather than Facebook, where people are scrolling to be entertained, not to shop.
The lesson here isn’t “always pick Pinterest.” It’s to match the channel to the buying behavior of your actual customer, not to whichever platform is trendy that month.
The campaign generated 1,140 landing-page visits, which works out to about $0.09 in advertising spend per visit. That made the traffic inexpensive enough to make the test worth evaluating, although traffic cost alone couldn’t tell us whether the campaign would be profitable.
Step 2: Set Up One Real Offer
Instead of promoting her general shop page, we built a single, specific offer: a “Cozy Night In” three-candle bundle, priced to feel like an easy yes, with a simple landing page that had one photo, one price, and one button.
This mattered more than the platform choice. A vague “check out my shop” ad rarely converts, because it asks the visitor to do the thinking for you—figure out what to buy, whether it’s a good deal, and whether it solves a problem they have. A specific bundle with a clear use case does that thinking for them.
Step 3: Let It Run Long Enough to Mean Something
We split the $100 over ten days rather than blowing through it in one weekend. This spreads impressions across more days and more times of day, instead of concentrating them into a single unrepresentative 48-hour window.
This is the part people rush the most, and it’s the easiest way to waste a small budget. A test that runs for two days doesn’t tell you if the channel works—it tells you if the first 48 hours of random exposure worked, which is a different and much less useful thing to know.
What the Numbers Looked Like
By day ten, the $100 had bought 1,140 landing page visits and 4 completed sales of the bundle at $28 each—a landing page conversion rate of about 0.35%. On its own, that number looks weak. But $112 in bundle revenue isn’t the same thing as $112 in profit—the $100 ad spend still had to come out of that, along with the cost of the candles, packaging, and shipping, so the actual margin on the test was thin, not the tidy win a raw revenue number might suggest. What the ten-day window couldn’t tell us was whether Pinterest traffic would keep trickling in after the ad spend stopped, since that’s a pattern platforms like Pinterest are known for but one we’d only be able to confirm by watching analytics in the weeks after the test, not from the campaign numbers alone.
That low conversion number is worth sitting with for a second, because it’s exactly the kind of detail a real test surfaces and a guess never would. A 0.35% conversion rate on its own would normally be a red flag. In this case it wasn’t a signal to quit—it was a signal that Pinterest traffic behaves more like top-of-funnel browsing than an in-the-moment purchase decision, so the right response was to compare it against saves and repeat traffic, not against a channel that behaves completely differently.
That’s the real insight, and it’s easy to miss if you only look at the conversion number: some channels are transactions (you pay, you get a burst of traffic, and it stops). Others are more like planting something (you pay once, and it keeps sending you a trickle of traffic for months). Pinterest and, to some extent, SEO-driven content or YouTube behave like the second kind. Many paid social and paid-search campaigns behave more like the first, because traffic tends to fall when paid distribution stops.
Knowing which kind of channel you’re testing changes how you judge the result. If you judge a “planting” channel only by what happened in the first ten days, you’ll wrongly conclude it failed.
The Mistakes We Almost Made
Almost mistake #1: Judging too early. After four days, the campaign had spent $40 with only one sale, and my friend wanted to pull the plug. We waited. Three of the four total sales came in the second half of the test, after day five—a pattern we could only see by letting the full ten days play out.
Almost mistake #2: Changing the creative mid-test. She wanted to swap the photo on day six because she thought a different one looked nicer. We didn’t, because changing a variable mid-test means you can no longer trust any of your results—you won’t know if a change in performance came from the new photo, the day of the week, or just normal day-to-day variation.
Almost mistake #3: Testing the channel and the offer at the same time. If we’d also tested two different bundle prices, we wouldn’t have known whether a bad result meant “wrong channel” or “wrong price.” Keep the number of moving parts as close to one as you can.
How to Run Your Own $100 Test
- Pick the one channel your actual customers already use to make similar purchases—not the one you personally enjoy using.
- Build one specific, easy-to-say-yes-to offer, with a landing page that has a single clear next step.
- Set a test period in advance — 10 days in this example — and avoid making major conclusions from a single day of results.
- Don’t touch anything mid-test. No new creative, no new audience, no new price. Let it run.
- After the test, ask two questions: Did this cover its own cost? And is this a “transaction” channel or a “planting” channel? That second question tells you how to judge it fairly.
- Decide, don’t drift. At the end, make an actual decision—double down, walk away, or run one more focused test—instead of letting the channel quietly fade into “something we tried once.”
Conclusion
A hundred dollars won’t build you a marketing engine on its own. But it’s more than enough to answer the one question that actually matters before you spend more money: does this channel fit how your actual customers behave? The businesses that get marketing right aren’t the ones with the biggest budgets—they’re the ones who test small, resist the urge to judge too early, and change one variable at a time so every dollar teaches them something. In this example, the candle bundle didn’t make the business rich in ten days. But a structured test like this one gives you evidence instead of a guess about where to put next month’s budget. That’s what a good $100 test is for.
FAQ
Is $100 really enough to test a marketing channel? Yes, if you keep the test narrow. One channel, one offer, one audience, run over at least a week. $100 spread across multiple channels or variables won’t tell you much, but $100 focused on a single question usually will.
How do I know which channel to test first? Think about how your customer already makes similar purchases. Do they search for solutions (Google), browse for ideas and save them (Pinterest), scroll for entertainment and get interrupted (Facebook/Instagram), or ask people they trust (community groups, referrals)? Pick the channel that matches that behavior.
What if the test loses money? A small loss on a first test is normal and not the same as failure—it’s data. The real failure is running a sloppy test (too short, too many variables) and not being able to trust the result either way.
Should I test organic content instead of paid ads with $100? You can combine them—for example, use $100 in ads to jump-start traffic to a piece of content, then track whether that content keeps generating interest after the ad spend stops. That tells you if you found a “planting” channel worth investing in further.
How long should I wait before judging the results? At minimum, let the budget run its full course (7-10 days) before making any real decision, and if the channel is a slow-build type like Pinterest, SEO, or YouTube, give it another few weeks after the test ends to see the lingering effect.
What’s the single biggest mistake people make with a small test budget? Changing something mid-test—the price, the photo, the audience—because it “feels” like the right move. Every change resets the clock on what you can actually learn from the data.
About the author: This piece is built as a worked example—a realistic, step-by-step walkthrough of how a $100 marketing channel test could run for a small business, including the numbers, decisions, and near-mistakes along the way. No affiliate links, no sponsored placements—just a process anyone can adapt and repeat with their own $100.
